Merge branch 'v4' into rename-islandic-and-multi-lang-code

.github/azure-steps.yml

@@ -52,12 +52,17 @@ steps:
       python -W error -c "import spacy"
     displayName: "Test import"
 
-  - script: |
+#  - script: |
-      python -m spacy download ca_core_news_sm
+#      python -m spacy download ca_core_news_sm
-      python -m spacy download ca_core_news_md
+#      python -m spacy download ca_core_news_md
-      python -c "import spacy; nlp=spacy.load('ca_core_news_sm'); doc=nlp('test')"
+#      python -c "import spacy; nlp=spacy.load('ca_core_news_sm'); doc=nlp('test')"
-    displayName: 'Test download CLI'
+#    displayName: 'Test download CLI'
-    condition: eq(variables['python_version'], '3.8')
+#    condition: eq(variables['python_version'], '3.8')
+#
+#  - script: |
+#      python -W error -c "import ca_core_news_sm; nlp = ca_core_news_sm.load(); doc=nlp('test')"
+#    displayName: 'Test no warnings on load (#11713)'
+#    condition: eq(variables['python_version'], '3.8')
 
   - script: |
       python -m spacy convert extra/example_data/ner_example_data/ner-token-per-line-conll2003.json .
@@ -81,17 +86,17 @@ steps:
     displayName: 'Test train CLI'
     condition: eq(variables['python_version'], '3.8')
 
-  - script: |
+#  - script: |
-      python -c "import spacy; config = spacy.util.load_config('ner.cfg'); config['components']['ner'] = {'source': 'ca_core_news_sm'}; config.to_disk('ner_source_sm.cfg')"
+#      python -c "import spacy; config = spacy.util.load_config('ner.cfg'); config['components']['ner'] = {'source': 'ca_core_news_sm'}; config.to_disk('ner_source_sm.cfg')"
-      PYTHONWARNINGS="error,ignore::DeprecationWarning" python -m spacy assemble ner_source_sm.cfg output_dir
+#      PYTHONWARNINGS="error,ignore::DeprecationWarning" python -m spacy assemble ner_source_sm.cfg output_dir
-    displayName: 'Test assemble CLI'
+#    displayName: 'Test assemble CLI'
-    condition: eq(variables['python_version'], '3.8')
+#    condition: eq(variables['python_version'], '3.8')
-
+#
-  - script: |
+#  - script: |
-      python -c "import spacy; config = spacy.util.load_config('ner.cfg'); config['components']['ner'] = {'source': 'ca_core_news_md'}; config.to_disk('ner_source_md.cfg')"
+#      python -c "import spacy; config = spacy.util.load_config('ner.cfg'); config['components']['ner'] = {'source': 'ca_core_news_md'}; config.to_disk('ner_source_md.cfg')"
-      python -m spacy assemble ner_source_md.cfg output_dir 2>&1 | grep -q W113
+#      python -m spacy assemble ner_source_md.cfg output_dir 2>&1 | grep -q W113
-    displayName: 'Test assemble CLI vectors warning'
+#    displayName: 'Test assemble CLI vectors warning'
-    condition: eq(variables['python_version'], '3.8')
+#    condition: eq(variables['python_version'], '3.8')
 
   - script: |
       python -m pip install -U -r requirements.txt
@@ -102,7 +107,7 @@ steps:
     displayName: "Run CPU tests"
 
   - script: |
-      python -m pip install --pre thinc-apple-ops
+      python -m pip install 'spacy[apple]'
       python -m pytest --pyargs spacy
     displayName: "Run CPU tests with thinc-apple-ops"
     condition: and(startsWith(variables['imageName'], 'macos'), eq(variables['python.version'], '3.11'))

.github/workflows/lock.yml

@@ -15,11 +15,11 @@ jobs:
   action:
     runs-on: ubuntu-latest
     steps:
-      - uses: dessant/lock-threads@v3
+      - uses: dessant/lock-threads@v4
         with:
           process-only: 'issues'
           issue-inactive-days: '30'
-          issue-comment: > 
+          issue-comment: >
-            This thread has been automatically locked since there 
+            This thread has been automatically locked since there
-            has not been any recent activity after it was closed. 
+            has not been any recent activity after it was closed.
             Please open a new issue for related bugs.

.github/workflows/spacy_universe_alert.yml

@@ -19,6 +19,8 @@ jobs:
 
       - uses: actions/checkout@v3
       - uses: actions/setup-python@v4
+        with:
+          python-version: '3.10'
       - name: Install Bernadette app dependency and send an alert
         env:
           SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}

.pre-commit-config.yaml

@@ -5,7 +5,7 @@ repos:
     - id: black
       language_version: python3.7
       additional_dependencies: ['click==8.0.4']
--   repo: https://gitlab.com/pycqa/flake8
+-   repo: https://github.com/pycqa/flake8
     rev: 5.0.4
     hooks:
     - id: flake8

README.md

@@ -14,7 +14,7 @@ parsing, **named entity recognition**, **text classification** and more,
 multi-task learning with pretrained **transformers** like BERT, as well as a
 production-ready [**training system**](https://spacy.io/usage/training) and easy
 model packaging, deployment and workflow management. spaCy is commercial
-open-source software, released under the MIT license.
+open-source software, released under the [MIT license](https://github.com/explosion/spaCy/blob/master/LICENSE).
 
 💫 **Version 3.4 out now!**
 [Check out the release notes here.](https://github.com/explosion/spaCy/releases)
@@ -46,6 +46,7 @@ open-source software, released under the MIT license.
 | 🛠 **[Changelog]**                                                                                                                                                                                                         | Changes and version history.                                                                                                                                                                                                                                                                                                 |
 | 💝 **[Contribute]**                                                                                                                                                                                                       | How to contribute to the spaCy project and code base.                                                                                                                                                                                                                                                                        |
 | <a href="https://explosion.ai/spacy-tailored-pipelines"><img src="https://user-images.githubusercontent.com/13643239/152853098-1c761611-ccb0-4ec6-9066-b234552831fe.png" width="125" alt="spaCy Tailored Pipelines"/></a> | Get a custom spaCy pipeline, tailor-made for your NLP problem by spaCy's core developers. Streamlined, production-ready, predictable and maintainable. Start by completing our 5-minute questionnaire to tell us what you need and we'll be in touch! **[Learn more &rarr;](https://explosion.ai/spacy-tailored-pipelines)** |
+| <a href="https://explosion.ai/spacy-tailored-analysis"><img src="https://user-images.githubusercontent.com/1019791/206151300-b00cd189-e503-4797-aa1e-1bb6344062c5.png" width="125" alt="spaCy Tailored Pipelines"/></a> | Bespoke advice for problem solving, strategy and analysis for applied NLP projects. Services include data strategy, code reviews, pipeline design and annotation coaching. Curious? Fill in our 5-minute questionnaire to tell us what you need and we'll be in touch! **[Learn more &rarr;](https://explosion.ai/spacy-tailored-analysis)** |
 
 [spacy 101]: https://spacy.io/usage/spacy-101
 [new in v3.0]: https://spacy.io/usage/v3
@@ -59,6 +60,7 @@ open-source software, released under the MIT license.
 [changelog]: https://spacy.io/usage#changelog
 [contribute]: https://github.com/explosion/spaCy/blob/master/CONTRIBUTING.md
 
+
 ## 💬 Where to ask questions
 
 The spaCy project is maintained by the [spaCy team](https://explosion.ai/about).

azure-pipelines.yml

@@ -41,7 +41,7 @@ jobs:
       matrix:
         # We're only running one platform per Python version to speed up builds
         Python36Linux:
-          imageName: "ubuntu-latest"
+          imageName: "ubuntu-20.04"
           python.version: "3.6"
         #        Python36Windows:
         #          imageName: "windows-latest"
@@ -50,7 +50,7 @@ jobs:
         #          imageName: "macos-latest"
         #          python.version: "3.6"
         #        Python37Linux:
-        #          imageName: "ubuntu-latest"
+        #          imageName: "ubuntu-20.04"
         #          python.version: "3.7"
         Python37Windows:
           imageName: "windows-latest"
@@ -87,13 +87,13 @@ jobs:
         #          python.version: "3.10"
         Python311Linux:
           imageName: 'ubuntu-latest'
-          python.version: '3.11.0'
+          python.version: '3.11'
         Python311Windows:
           imageName: 'windows-latest'
-          python.version: '3.11.0'
+          python.version: '3.11'
         Python311Mac:
           imageName: 'macos-latest'
-          python.version: '3.11.0'
+          python.version: '3.11'
       maxParallel: 4
     pool:
       vmImage: $(imageName)

build-constraints.txt

@@ -5,4 +5,5 @@ numpy==1.17.3; python_version=='3.8' and platform_machine!='aarch64'
 numpy==1.19.2; python_version=='3.8' and platform_machine=='aarch64'
 numpy==1.19.3; python_version=='3.9'
 numpy==1.21.3; python_version=='3.10'
-numpy; python_version>='3.11'
+numpy==1.23.2; python_version=='3.11'
+numpy; python_version>='3.12'

pyproject.toml

@@ -5,7 +5,7 @@ requires = [
     "cymem>=2.0.2,<2.1.0",
     "preshed>=3.0.2,<3.1.0",
     "murmurhash>=0.28.0,<1.1.0",
-    "thinc>=8.1.0,<8.2.0",
+    "thinc>=9.0.0.dev1,<9.1.0",
     "numpy>=1.15.0",
 ]
 build-backend = "setuptools.build_meta"

requirements.txt

@@ -3,14 +3,15 @@ spacy-legacy>=3.0.10,<3.1.0
 spacy-loggers>=1.0.0,<2.0.0
 cymem>=2.0.2,<2.1.0
 preshed>=3.0.2,<3.1.0
-thinc>=8.1.0,<8.2.0
+thinc>=9.0.0.dev1,<9.1.0
 ml_datasets>=0.2.0,<0.3.0
 murmurhash>=0.28.0,<1.1.0
-wasabi>=0.9.1,<1.1.0
+wasabi>=0.9.1,<1.2.0
 srsly>=2.4.3,<3.0.0
 catalogue>=2.0.6,<2.1.0
-typer>=0.3.0,<0.5.0
+typer>=0.3.0,<0.8.0
-pathy>=0.3.5
+pathy>=0.10.0
+smart-open>=5.2.1,<7.0.0
 # Third party dependencies
 numpy>=1.15.0
 requests>=2.13.0,<3.0.0
@@ -30,7 +31,7 @@ pytest-timeout>=1.3.0,<2.0.0
 mock>=2.0.0,<3.0.0
 flake8>=3.8.0,<6.0.0
 hypothesis>=3.27.0,<7.0.0
-mypy>=0.980,<0.990; platform_machine != "aarch64" and python_version >= "3.7"
+mypy>=0.990,<0.1000; platform_machine != "aarch64" and python_version >= "3.7"
 types-dataclasses>=0.1.3; python_version < "3.7"
 types-mock>=0.1.1
 types-setuptools>=57.0.0

setup.cfg

@@ -38,13 +38,14 @@ install_requires =
     murmurhash>=0.28.0,<1.1.0
     cymem>=2.0.2,<2.1.0
     preshed>=3.0.2,<3.1.0
-    thinc>=8.1.0,<8.2.0
+    thinc>=9.0.0.dev1,<9.1.0
-    wasabi>=0.9.1,<1.1.0
+    wasabi>=0.9.1,<1.2.0
     srsly>=2.4.3,<3.0.0
     catalogue>=2.0.6,<2.1.0
     # Third-party dependencies
-    typer>=0.3.0,<0.5.0
+    typer>=0.3.0,<0.8.0
-    pathy>=0.3.5
+    pathy>=0.10.0
+    smart-open>=5.2.1,<7.0.0
     tqdm>=4.38.0,<5.0.0
     numpy>=1.15.0
     requests>=2.13.0,<3.0.0

setup.py

@@ -38,7 +38,6 @@ MOD_NAMES = [
     "spacy.pipeline.dep_parser",
     "spacy.pipeline._edit_tree_internals.edit_trees",
     "spacy.pipeline.morphologizer",
-    "spacy.pipeline.multitask",
     "spacy.pipeline.ner",
     "spacy.pipeline.pipe",
     "spacy.pipeline.trainable_pipe",
@@ -49,6 +48,7 @@ MOD_NAMES = [
     "spacy.pipeline._parser_internals.arc_eager",
     "spacy.pipeline._parser_internals.ner",
     "spacy.pipeline._parser_internals.nonproj",
+    "spacy.pipeline._parser_internals.search",
     "spacy.pipeline._parser_internals._state",
     "spacy.pipeline._parser_internals.stateclass",
     "spacy.pipeline._parser_internals.transition_system",
@@ -68,6 +68,7 @@ MOD_NAMES = [
     "spacy.matcher.dependencymatcher",
     "spacy.symbols",
     "spacy.vectors",
+    "spacy.tests.parser._search",
 ]
 COMPILE_OPTIONS = {
     "msvc": ["/Ox", "/EHsc"],

spacy/about.py

@@ -1,6 +1,6 @@
 # fmt: off
 __title__ = "spacy"
-__version__ = "3.4.2"
+__version__ = "3.5.0"
 __download_url__ = "https://github.com/explosion/spacy-models/releases/download"
 __compatibility__ = "https://raw.githubusercontent.com/explosion/spacy-models/master/compatibility.json"
 __projects__ = "https://github.com/explosion/projects"

spacy/cli/__init__.py

@@ -16,6 +16,7 @@ from .debug_config import debug_config  # noqa: F401
 from .debug_model import debug_model  # noqa: F401
 from .debug_diff import debug_diff  # noqa: F401
 from .evaluate import evaluate  # noqa: F401
+from .apply import apply  # noqa: F401
 from .convert import convert  # noqa: F401
 from .init_pipeline import init_pipeline_cli  # noqa: F401
 from .init_config import init_config, fill_config  # noqa: F401
@@ -27,6 +28,7 @@ from .project.dvc import project_update_dvc  # noqa: F401
 from .project.push import project_push  # noqa: F401
 from .project.pull import project_pull  # noqa: F401
 from .project.document import project_document  # noqa: F401
+from .find_threshold import find_threshold  # noqa: F401
 
 
 @app.command("link", no_args_is_help=True, deprecated=True, hidden=True)

spacy/cli/_util.py

@@ -24,7 +24,7 @@ from ..errors import RENAMED_LANGUAGE_CODES
 from .. import about
 
 if TYPE_CHECKING:
-    from pathy import Pathy  # noqa: F401
+    from pathy import FluidPath  # noqa: F401
 
 
 SDIST_SUFFIX = ".tar.gz"
@@ -169,15 +169,15 @@ def load_project_config(
         sys.exit(1)
     validate_project_version(config)
     validate_project_commands(config)
+    if interpolate:
+        err = f"{PROJECT_FILE} validation error"
+        with show_validation_error(title=err, hint_fill=False):
+            config = substitute_project_variables(config, overrides)
     # Make sure directories defined in config exist
     for subdir in config.get("directories", []):
         dir_path = path / subdir
         if not dir_path.exists():
             dir_path.mkdir(parents=True)
-    if interpolate:
-        err = f"{PROJECT_FILE} validation error"
-        with show_validation_error(title=err, hint_fill=False):
-            config = substitute_project_variables(config, overrides)
     return config
 
 
@@ -342,7 +342,7 @@ def import_code(code_path: Optional[Union[Path, str]]) -> None:
             msg.fail(f"Couldn't load Python code: {code_path}", e, exits=1)
 
 
-def upload_file(src: Path, dest: Union[str, "Pathy"]) -> None:
+def upload_file(src: Path, dest: Union[str, "FluidPath"]) -> None:
     """Upload a file.
 
     src (Path): The source path.
@@ -350,13 +350,20 @@ def upload_file(src: Path, dest: Union[str, "Pathy"]) -> None:
     """
     import smart_open
 
+    # Create parent directories for local paths
+    if isinstance(dest, Path):
+        if not dest.parent.exists():
+            dest.parent.mkdir(parents=True)
+
     dest = str(dest)
     with smart_open.open(dest, mode="wb") as output_file:
         with src.open(mode="rb") as input_file:
             output_file.write(input_file.read())
 
 
-def download_file(src: Union[str, "Pathy"], dest: Path, *, force: bool = False) -> None:
+def download_file(
+    src: Union[str, "FluidPath"], dest: Path, *, force: bool = False
+) -> None:
     """Download a file using smart_open.
 
     url (str): The URL of the file.
@@ -369,7 +376,7 @@ def download_file(src: Union[str, "Pathy"], dest: Path, *, force: bool = False)
     if dest.exists() and not force:
         return None
     src = str(src)
-    with smart_open.open(src, mode="rb", ignore_ext=True) as input_file:
+    with smart_open.open(src, mode="rb", compression="disable") as input_file:
         with dest.open(mode="wb") as output_file:
             shutil.copyfileobj(input_file, output_file)
 
@@ -379,7 +386,7 @@ def ensure_pathy(path):
     slow and annoying Google Cloud warning)."""
     from pathy import Pathy  # noqa: F811
 
-    return Pathy(path)
+    return Pathy.fluid(path)
 
 
 def git_checkout(
@@ -586,6 +593,29 @@ def setup_gpu(use_gpu: int, silent=None) -> None:
             local_msg.info("To switch to GPU 0, use the option: --gpu-id 0")
 
 
+def walk_directory(path: Path, suffix: Optional[str] = None) -> List[Path]:
+    if not path.is_dir():
+        return [path]
+    paths = [path]
+    locs = []
+    seen = set()
+    for path in paths:
+        if str(path) in seen:
+            continue
+        seen.add(str(path))
+        if path.parts[-1].startswith("."):
+            continue
+        elif path.is_dir():
+            paths.extend(path.iterdir())
+        elif suffix is not None and not path.parts[-1].endswith(suffix):
+            continue
+        else:
+            locs.append(path)
+    # It's good to sort these, in case the ordering messes up cache.
+    locs.sort()
+    return locs
+
+
 def _format_number(number: Union[int, float], ndigits: int = 2) -> str:
     """Formats a number (float or int) rounding to `ndigits`, without truncating trailing 0s,
     as happens with `round(number, ndigits)`"""

spacy/cli/apply.py

@@ -0,0 +1,143 @@
+import tqdm
+import srsly
+
+from itertools import chain
+from pathlib import Path
+from typing import Optional, List, Iterable, cast, Union
+
+from wasabi import msg
+
+from ._util import app, Arg, Opt, setup_gpu, import_code, walk_directory
+
+from ..tokens import Doc, DocBin
+from ..vocab import Vocab
+from ..util import ensure_path, load_model
+
+
+path_help = """Location of the documents to predict on.
+Can be a single file in .spacy format or a .jsonl file.
+Files with other extensions are treated as single plain text documents.
+If a directory is provided it is traversed recursively to grab
+all files to be processed.
+The files can be a mixture of .spacy, .jsonl and text files.
+If .jsonl is provided the specified field is going
+to be grabbed ("text" by default)."""
+
+out_help = "Path to save the resulting .spacy file"
+code_help = (
+    "Path to Python file with additional " "code (registered functions) to be imported"
+)
+gold_help = "Use gold preprocessing provided in the .spacy files"
+force_msg = (
+    "The provided output file already exists. "
+    "To force overwriting the output file, set the --force or -F flag."
+)
+
+
+DocOrStrStream = Union[Iterable[str], Iterable[Doc]]
+
+
+def _stream_docbin(path: Path, vocab: Vocab) -> Iterable[Doc]:
+    """
+    Stream Doc objects from DocBin.
+    """
+    docbin = DocBin().from_disk(path)
+    for doc in docbin.get_docs(vocab):
+        yield doc
+
+
+def _stream_jsonl(path: Path, field: str) -> Iterable[str]:
+    """
+    Stream "text" field from JSONL. If the field "text" is
+    not found it raises error.
+    """
+    for entry in srsly.read_jsonl(path):
+        if field not in entry:
+            msg.fail(
+                f"{path} does not contain the required '{field}' field.", exits=1
+            )
+        else:
+            yield entry[field]
+
+
+def _stream_texts(paths: Iterable[Path]) -> Iterable[str]:
+    """
+    Yields strings from text files in paths.
+    """
+    for path in paths:
+        with open(path, "r") as fin:
+            text = fin.read()
+            yield text
+
+
+@app.command("apply")
+def apply_cli(
+    # fmt: off
+    model: str = Arg(..., help="Model name or path"),
+    data_path: Path = Arg(..., help=path_help, exists=True),
+    output_file: Path = Arg(..., help=out_help, dir_okay=False),
+    code_path: Optional[Path] = Opt(None, "--code", "-c", help=code_help),
+    text_key: str = Opt("text", "--text-key", "-tk", help="Key containing text string for JSONL"),
+    force_overwrite: bool = Opt(False, "--force", "-F", help="Force overwriting the output file"),
+    use_gpu: int = Opt(-1, "--gpu-id", "-g", help="GPU ID or -1 for CPU."),
+    batch_size: int = Opt(1, "--batch-size", "-b", help="Batch size."),
+    n_process: int = Opt(1, "--n-process", "-n", help="number of processors to use.")
+):
+    """
+    Apply a trained pipeline to documents to get predictions.
+    Expects a loadable spaCy pipeline and path to the data, which
+    can be a directory or a file.
+    The data files can be provided in multiple formats:
+        1. .spacy files
+        2. .jsonl files with a specified "field" to read the text from.
+        3. Files with any other extension are assumed to be containing
+           a single document.
+    DOCS: https://spacy.io/api/cli#apply
+    """
+    data_path = ensure_path(data_path)
+    output_file = ensure_path(output_file)
+    code_path = ensure_path(code_path)
+    if output_file.exists() and not force_overwrite:
+        msg.fail(force_msg, exits=1)
+    if not data_path.exists():
+        msg.fail(f"Couldn't find data path: {data_path}", exits=1)
+    import_code(code_path)
+    setup_gpu(use_gpu)
+    apply(data_path, output_file, model, text_key, batch_size, n_process)
+
+
+def apply(
+    data_path: Path,
+    output_file: Path,
+    model: str,
+    json_field: str,
+    batch_size: int,
+    n_process: int,
+):
+    docbin = DocBin(store_user_data=True)
+    paths = walk_directory(data_path)
+    if len(paths) == 0:
+        docbin.to_disk(output_file)
+        msg.warn("Did not find data to process,"
+                 f" {data_path} seems to be an empty directory.")
+        return
+    nlp = load_model(model)
+    msg.good(f"Loaded model {model}")
+    vocab = nlp.vocab
+    streams: List[DocOrStrStream] = []
+    text_files = []
+    for path in paths:
+        if path.suffix == ".spacy":
+            streams.append(_stream_docbin(path, vocab))
+        elif path.suffix == ".jsonl":
+            streams.append(_stream_jsonl(path, json_field))
+        else:
+            text_files.append(path)
+    if len(text_files) > 0:
+        streams.append(_stream_texts(text_files))
+    datagen = cast(DocOrStrStream, chain(*streams))
+    for doc in tqdm.tqdm(nlp.pipe(datagen, batch_size=batch_size, n_process=n_process)):
+        docbin.add(doc)
+    if output_file.suffix == "":
+        output_file = output_file.with_suffix(".spacy")
+    docbin.to_disk(output_file)

spacy/cli/convert.py

@@ -1,4 +1,4 @@
-from typing import Callable, Iterable, Mapping, Optional, Any, List, Union
+from typing import Callable, Iterable, Mapping, Optional, Any, Union
 from enum import Enum
 from pathlib import Path
 from wasabi import Printer
@@ -193,33 +193,6 @@ def autodetect_ner_format(input_data: str) -> Optional[str]:
     return None
 
 
-def walk_directory(path: Path, converter: str) -> List[Path]:
-    if not path.is_dir():
-        return [path]
-    paths = [path]
-    locs = []
-    seen = set()
-    for path in paths:
-        if str(path) in seen:
-            continue
-        seen.add(str(path))
-        if path.parts[-1].startswith("."):
-            continue
-        elif path.is_dir():
-            paths.extend(path.iterdir())
-        elif converter == "json" and not path.parts[-1].endswith("json"):
-            continue
-        elif converter == "conll" and not path.parts[-1].endswith("conll"):
-            continue
-        elif converter == "iob" and not path.parts[-1].endswith("iob"):
-            continue
-        else:
-            locs.append(path)
-    # It's good to sort these, in case the ordering messes up cache.
-    locs.sort()
-    return locs
-
-
 def verify_cli_args(
     msg: Printer,
     input_path: Path,

spacy/cli/debug_data.py

@@ -13,6 +13,7 @@ from ._util import import_code, debug_cli, _format_number
 from ..training import Example, remove_bilu_prefix
 from ..training.initialize import get_sourced_components
 from ..schemas import ConfigSchemaTraining
+from ..pipeline import TrainablePipe
 from ..pipeline._parser_internals import nonproj
 from ..pipeline._parser_internals.nonproj import DELIMITER
 from ..pipeline import Morphologizer, SpanCategorizer
@@ -934,6 +935,7 @@ def _get_labels_from_model(nlp: Language, factory_name: str) -> Set[str]:
     labels: Set[str] = set()
     for pipe_name in pipe_names:
         pipe = nlp.get_pipe(pipe_name)
+        assert isinstance(pipe, TrainablePipe)
         labels.update(pipe.labels)
     return labels
 

spacy/cli/download.py

@@ -8,7 +8,6 @@ from ._util import app, Arg, Opt, WHEEL_SUFFIX, SDIST_SUFFIX
 from .. import about
 from ..util import is_package, get_minor_version, run_command
 from ..util import is_prerelease_version
-from ..errors import OLD_MODEL_SHORTCUTS
 
 
 @app.command(
@@ -61,12 +60,6 @@ def download(
         version = components[-1]
     else:
         model_name = model
-        if model in OLD_MODEL_SHORTCUTS:
-            msg.warn(
-                f"As of spaCy v3.0, shortcuts like '{model}' are deprecated. Please "
-                f"use the full pipeline package name '{OLD_MODEL_SHORTCUTS[model]}' instead."
-            )
-            model_name = OLD_MODEL_SHORTCUTS[model]
         compatibility = get_compatibility()
         version = get_version(model_name, compatibility)
 

spacy/cli/find_threshold.py

@@ -0,0 +1,233 @@
+import functools
+import operator
+from pathlib import Path
+import logging
+from typing import Optional, Tuple, Any, Dict, List
+
+import numpy
+import wasabi.tables
+
+from ..pipeline import TextCategorizer, MultiLabel_TextCategorizer
+from ..errors import Errors
+from ..training import Corpus
+from ._util import app, Arg, Opt, import_code, setup_gpu
+from .. import util
+
+_DEFAULTS = {
+    "n_trials": 11,
+    "use_gpu": -1,
+    "gold_preproc": False,
+}
+
+
+@app.command(
+    "find-threshold",
+    context_settings={"allow_extra_args": False, "ignore_unknown_options": True},
+)
+def find_threshold_cli(
+    # fmt: off
+    model: str = Arg(..., help="Model name or path"),
+    data_path: Path = Arg(..., help="Location of binary evaluation data in .spacy format", exists=True),
+    pipe_name: str = Arg(..., help="Name of pipe to examine thresholds for"),
+    threshold_key: str = Arg(..., help="Key of threshold attribute in component's configuration"),
+    scores_key: str = Arg(..., help="Metric to optimize"),
+    n_trials: int = Opt(_DEFAULTS["n_trials"], "--n_trials", "-n", help="Number of trials to determine optimal thresholds"),
+    code_path: Optional[Path] = Opt(None, "--code", "-c", help="Path to Python file with additional code (registered functions) to be imported"),
+    use_gpu: int = Opt(_DEFAULTS["use_gpu"], "--gpu-id", "-g", help="GPU ID or -1 for CPU"),
+    gold_preproc: bool = Opt(_DEFAULTS["gold_preproc"], "--gold-preproc", "-G", help="Use gold preprocessing"),
+    verbose: bool = Opt(False, "--silent", "-V", "-VV", help="Display more information for debugging purposes"),
+    # fmt: on
+):
+    """
+    Runs prediction trials for a trained model with varying tresholds to maximize
+    the specified metric. The search space for the threshold is traversed linearly
+    from 0 to 1 in `n_trials` steps. Results are displayed in a table on `stdout`
+    (the corresponding API call to `spacy.cli.find_threshold.find_threshold()`
+    returns all results).
+
+    This is applicable only for components whose predictions are influenced by
+    thresholds - e.g. `textcat_multilabel` and `spancat`, but not `textcat`. Note
+    that the full path to the corresponding threshold attribute in the config has to
+    be provided.
+
+    DOCS: https://spacy.io/api/cli#find-threshold
+    """
+
+    util.logger.setLevel(logging.DEBUG if verbose else logging.INFO)
+    import_code(code_path)
+    find_threshold(
+        model=model,
+        data_path=data_path,
+        pipe_name=pipe_name,
+        threshold_key=threshold_key,
+        scores_key=scores_key,
+        n_trials=n_trials,
+        use_gpu=use_gpu,
+        gold_preproc=gold_preproc,
+        silent=False,
+    )
+
+
+def find_threshold(
+    model: str,
+    data_path: Path,
+    pipe_name: str,
+    threshold_key: str,
+    scores_key: str,
+    *,
+    n_trials: int = _DEFAULTS["n_trials"],  # type: ignore
+    use_gpu: int = _DEFAULTS["use_gpu"],  # type: ignore
+    gold_preproc: bool = _DEFAULTS["gold_preproc"],  # type: ignore
+    silent: bool = True,
+) -> Tuple[float, float, Dict[float, float]]:
+    """
+    Runs prediction trials for models with varying tresholds to maximize the specified metric.
+    model (Union[str, Path]): Pipeline to evaluate. Can be a package or a path to a data directory.
+    data_path (Path): Path to file with DocBin with docs to use for threshold search.
+    pipe_name (str): Name of pipe to examine thresholds for.
+    threshold_key (str): Key of threshold attribute in component's configuration.
+    scores_key (str): Name of score to metric to optimize.
+    n_trials (int): Number of trials to determine optimal thresholds.
+    use_gpu (int): GPU ID or -1 for CPU.
+    gold_preproc (bool): Whether to use gold preprocessing. Gold preprocessing helps the annotations align to the
+        tokenization, and may result in sequences of more consistent length. However, it may reduce runtime accuracy due
+        to train/test skew.
+    silent (bool): Whether to print non-error-related output to stdout.
+    RETURNS (Tuple[float, float, Dict[float, float]]): Best found threshold, the corresponding score, scores for all
+        evaluated thresholds.
+    """
+
+    setup_gpu(use_gpu, silent=silent)
+    data_path = util.ensure_path(data_path)
+    if not data_path.exists():
+        wasabi.msg.fail("Evaluation data not found", data_path, exits=1)
+    nlp = util.load_model(model)
+
+    if pipe_name not in nlp.component_names:
+        raise AttributeError(
+            Errors.E001.format(name=pipe_name, opts=nlp.component_names)
+        )
+    pipe = nlp.get_pipe(pipe_name)
+    if not hasattr(pipe, "scorer"):
+        raise AttributeError(Errors.E1045)
+
+    if type(pipe) == TextCategorizer:
+        wasabi.msg.warn(
+            "The `textcat` component doesn't use a threshold as it's not applicable to the concept of "
+            "exclusive classes. All thresholds will yield the same results."
+        )
+
+    if not silent:
+        wasabi.msg.info(
+            title=f"Optimizing for {scores_key} for component '{pipe_name}' with {n_trials} "
+            f"trials."
+        )
+
+    # Load evaluation corpus.
+    corpus = Corpus(data_path, gold_preproc=gold_preproc)
+    dev_dataset = list(corpus(nlp))
+    config_keys = threshold_key.split(".")
+
+    def set_nested_item(
+        config: Dict[str, Any], keys: List[str], value: float
+    ) -> Dict[str, Any]:
+        """Set item in nested dictionary. Adapted from https://stackoverflow.com/a/54138200.
+        config (Dict[str, Any]): Configuration dictionary.
+        keys (List[Any]): Path to value to set.
+        value (float): Value to set.
+        RETURNS (Dict[str, Any]): Updated dictionary.
+        """
+        functools.reduce(operator.getitem, keys[:-1], config)[keys[-1]] = value
+        return config
+
+    def filter_config(
+        config: Dict[str, Any], keys: List[str], full_key: str
+    ) -> Dict[str, Any]:
+        """Filters provided config dictionary so that only the specified keys path remains.
+        config (Dict[str, Any]): Configuration dictionary.
+        keys (List[Any]): Path to value to set.
+        full_key (str): Full user-specified key.
+        RETURNS (Dict[str, Any]): Filtered dictionary.
+        """
+        if keys[0] not in config:
+            wasabi.msg.fail(
+                title=f"Failed to look up `{full_key}` in config: sub-key {[keys[0]]} not found.",
+                text=f"Make sure you specified {[keys[0]]} correctly. The following sub-keys are available instead: "
+                f"{list(config.keys())}",
+                exits=1,
+            )
+        return {
+            keys[0]: filter_config(config[keys[0]], keys[1:], full_key)
+            if len(keys) > 1
+            else config[keys[0]]
+        }
+
+    # Evaluate with varying threshold values.
+    scores: Dict[float, float] = {}
+    config_keys_full = ["components", pipe_name, *config_keys]
+    table_col_widths = (10, 10)
+    thresholds = numpy.linspace(0, 1, n_trials)
+    print(wasabi.tables.row(["Threshold", f"{scores_key}"], widths=table_col_widths))
+    for threshold in thresholds:
+        # Reload pipeline with overrides specifying the new threshold.
+        nlp = util.load_model(
+            model,
+            config=set_nested_item(
+                filter_config(
+                    nlp.config, config_keys_full, ".".join(config_keys_full)
+                ).copy(),
+                config_keys_full,
+                threshold,
+            ),
+        )
+        if hasattr(pipe, "cfg"):
+            setattr(
+                nlp.get_pipe(pipe_name),
+                "cfg",
+                set_nested_item(getattr(pipe, "cfg"), config_keys, threshold),
+            )
+
+        eval_scores = nlp.evaluate(dev_dataset)
+        if scores_key not in eval_scores:
+            wasabi.msg.fail(
+                title=f"Failed to look up score `{scores_key}` in evaluation results.",
+                text=f"Make sure you specified the correct value for `scores_key`. The following scores are "
+                f"available: {list(eval_scores.keys())}",
+                exits=1,
+            )
+        scores[threshold] = eval_scores[scores_key]
+
+        if not isinstance(scores[threshold], (float, int)):
+            wasabi.msg.fail(
+                f"Returned score for key '{scores_key}' is not numeric. Threshold optimization only works for numeric "
+                f"scores.",
+                exits=1,
+            )
+        print(
+            wasabi.row(
+                [round(threshold, 3), round(scores[threshold], 3)],
+                widths=table_col_widths,
+            )
+        )
+
+    best_threshold = max(scores.keys(), key=(lambda key: scores[key]))
+
+    # If all scores are identical, emit warning.
+    if len(set(scores.values())) == 1:
+        wasabi.msg.warn(
+            title="All scores are identical. Verify that all settings are correct.",
+            text=""
+            if (
+                not isinstance(pipe, MultiLabel_TextCategorizer)
+                or scores_key in ("cats_macro_f", "cats_micro_f")
+            )
+            else "Use `cats_macro_f` or `cats_micro_f` when optimizing the threshold for `textcat_multilabel`.",
+        )
+
+    else:
+        if not silent:
+            print(
+                f"\nBest threshold: {round(best_threshold, ndigits=4)} with {scores_key} value of {scores[best_threshold]}."
+            )
+
+    return best_threshold, scores[best_threshold], scores

spacy/cli/project/assets.py

@@ -189,7 +189,11 @@ def convert_asset_url(url: str) -> str:
     RETURNS (str): The converted URL.
     """
     # If the asset URL is a regular GitHub URL it's likely a mistake
-    if re.match(r"(http(s?)):\/\/github.com", url) and "releases/download" not in url:
+    if (
+        re.match(r"(http(s?)):\/\/github.com", url)
+        and "releases/download" not in url
+        and "/raw/" not in url
+    ):
         converted = url.replace("github.com", "raw.githubusercontent.com")
         converted = re.sub(r"/(tree|blob)/", "/", converted)
         msg.warn(

spacy/cli/project/remote_storage.py

@@ -5,14 +5,17 @@ import hashlib
 import urllib.parse
 import tarfile
 from pathlib import Path
+from wasabi import msg
 
-from .._util import get_hash, get_checksum, download_file, ensure_pathy
+from .._util import get_hash, get_checksum, upload_file, download_file
-from ...util import make_tempdir, get_minor_version, ENV_VARS, check_bool_env_var
+from .._util import ensure_pathy, make_tempdir
+from ...util import get_minor_version, ENV_VARS, check_bool_env_var
 from ...git_info import GIT_VERSION
 from ... import about
+from ...errors import Errors
 
 if TYPE_CHECKING:
-    from pathy import Pathy  # noqa: F401
+    from pathy import FluidPath  # noqa: F401
 
 
 class RemoteStorage:
@@ -27,7 +30,7 @@ class RemoteStorage:
         self.url = ensure_pathy(url)
         self.compression = compression
 
-    def push(self, path: Path, command_hash: str, content_hash: str) -> "Pathy":
+    def push(self, path: Path, command_hash: str, content_hash: str) -> "FluidPath":
         """Compress a file or directory within a project and upload it to a remote
         storage. If an object exists at the full URL, nothing is done.
 
@@ -48,9 +51,7 @@ class RemoteStorage:
             mode_string = f"w:{self.compression}" if self.compression else "w"
             with tarfile.open(tar_loc, mode=mode_string) as tar_file:
                 tar_file.add(str(loc), arcname=str(path))
-            with tar_loc.open(mode="rb") as input_file:
+            upload_file(tar_loc, url)
-                with url.open(mode="wb") as output_file:
-                    output_file.write(input_file.read())
         return url
 
     def pull(
@@ -59,7 +60,7 @@ class RemoteStorage:
         *,
         command_hash: Optional[str] = None,
         content_hash: Optional[str] = None,
-    ) -> Optional["Pathy"]:
+    ) -> Optional["FluidPath"]:
         """Retrieve a file from the remote cache. If the file already exists,
         nothing is done.
 
@@ -84,7 +85,23 @@ class RemoteStorage:
                 with tarfile.open(tar_loc, mode=mode_string) as tar_file:
                     # This requires that the path is added correctly, relative
                     # to root. This is how we set things up in push()
-                    tar_file.extractall(self.root)
+
+                    # Disallow paths outside the current directory for the tar
+                    # file (CVE-2007-4559, directory traversal vulnerability)
+                    def is_within_directory(directory, target):
+                        abs_directory = os.path.abspath(directory)
+                        abs_target = os.path.abspath(target)
+                        prefix = os.path.commonprefix([abs_directory, abs_target])
+                        return prefix == abs_directory
+
+                    def safe_extract(tar, path):
+                        for member in tar.getmembers():
+                            member_path = os.path.join(path, member.name)
+                            if not is_within_directory(path, member_path):
+                                raise ValueError(Errors.E852)
+                        tar.extractall(path)
+
+                    safe_extract(tar_file, self.root)
         return url
 
     def find(
@@ -93,25 +110,37 @@ class RemoteStorage:
         *,
         command_hash: Optional[str] = None,
         content_hash: Optional[str] = None,
-    ) -> Optional["Pathy"]:
+    ) -> Optional["FluidPath"]:
         """Find the best matching version of a file within the storage,
         or `None` if no match can be found. If both the creation and content hash
         are specified, only exact matches will be returned. Otherwise, the most
         recent matching file is preferred.
         """
         name = self.encode_name(str(path))
+        urls = []
         if command_hash is not None and content_hash is not None:
-            url = self.make_url(path, command_hash, content_hash)
+            url = self.url / name / command_hash / content_hash
             urls = [url] if url.exists() else []
         elif command_hash is not None:
-            urls = list((self.url / name / command_hash).iterdir())
+            if (self.url / name / command_hash).exists():
+                urls = list((self.url / name / command_hash).iterdir())
         else:
-            urls = list((self.url / name).iterdir())
+            if (self.url / name).exists():
-            if content_hash is not None:
+                for sub_dir in (self.url / name).iterdir():
-                urls = [url for url in urls if url.parts[-1] == content_hash]
+                    urls.extend(sub_dir.iterdir())
+                if content_hash is not None:
+                    urls = [url for url in urls if url.parts[-1] == content_hash]
+        if len(urls) >= 2:
+            try:
+                urls.sort(key=lambda x: x.stat().last_modified)  # type: ignore
+            except Exception:
+                msg.warn(
+                    "Unable to sort remote files by last modified. The file(s) "
+                    "pulled from the cache may not be the most recent."
+                )
         return urls[-1] if urls else None
 
-    def make_url(self, path: Path, command_hash: str, content_hash: str) -> "Pathy":
+    def make_url(self, path: Path, command_hash: str, content_hash: str) -> "FluidPath":
         """Construct a URL from a subpath, a creation hash and a content hash."""
         return self.url / self.encode_name(str(path)) / command_hash / content_hash
 

spacy/cli/project/run.py

@@ -53,6 +53,7 @@ def project_run(
     force: bool = False,
     dry: bool = False,
     capture: bool = False,
+    skip_requirements_check: bool = False,
 ) -> None:
     """Run a named script defined in the project.yml. If the script is part
     of the default pipeline (defined in the "run" section), DVC is used to
@@ -69,6 +70,7 @@ def project_run(
         sys.exit will be called with the return code. You should use capture=False
         when you want to turn over execution to the command, and capture=True
         when you want to run the command more like a function.
+    skip_requirements_check (bool): Whether to skip the requirements check.
     """
     config = load_project_config(project_dir, overrides=overrides)
     commands = {cmd["name"]: cmd for cmd in config.get("commands", [])}
@@ -76,9 +78,10 @@ def project_run(
     validate_subcommand(list(commands.keys()), list(workflows.keys()), subcommand)
 
     req_path = project_dir / "requirements.txt"
-    if config.get("check_requirements", True) and os.path.exists(req_path):
+    if not skip_requirements_check:
-        with req_path.open() as requirements_file:
+        if config.get("check_requirements", True) and os.path.exists(req_path):
-            _check_requirements([req.replace("\n", "") for req in requirements_file])
+            with req_path.open() as requirements_file:
+                _check_requirements([req.strip() for req in requirements_file])
 
     if subcommand in workflows:
         msg.info(f"Running workflow '{subcommand}'")
@@ -90,6 +93,7 @@ def project_run(
                 force=force,
                 dry=dry,
                 capture=capture,
+                skip_requirements_check=True,
             )
     else:
         cmd = commands[subcommand]
@@ -97,8 +101,8 @@ def project_run(
             if not (project_dir / dep).exists():
                 err = f"Missing dependency specified by command '{subcommand}': {dep}"
                 err_help = "Maybe you forgot to run the 'project assets' command or a previous step?"
-                err_kwargs = {"exits": 1} if not dry else {}
+                err_exits = 1 if not dry else None
-                msg.fail(err, err_help, **err_kwargs)
+                msg.fail(err, err_help, exits=err_exits)
         check_spacy_commit = check_bool_env_var(ENV_VARS.PROJECT_USE_GIT_VERSION)
         with working_dir(project_dir) as current_dir:
             msg.divider(subcommand)
@@ -338,6 +342,12 @@ def _check_requirements(requirements: List[str]) -> Tuple[bool, bool]:
             failed_pkgs_msgs.append(dnf.report())
         except pkg_resources.VersionConflict as vc:
             conflicting_pkgs_msgs.append(vc.report())
+        except Exception:
+            msg.warn(
+                f"Unable to check requirement: {req} "
+                "Checks are currently limited to requirement specifiers "
+                "(PEP 508)"
+            )
 
     if len(failed_pkgs_msgs) or len(conflicting_pkgs_msgs):
         msg.warn(

spacy/cli/templates/quickstart_training.jinja

@@ -1,7 +1,7 @@
 {# This is a template for training configs used for the quickstart widget in
 the docs and the init config command. It encodes various best practices and
 can help generate the best possible configuration, given a user's requirements. #}
-{%- set use_transformer = hardware != "cpu" -%}
+{%- set use_transformer = hardware != "cpu" and transformer_data -%}
 {%- set transformer = transformer_data[optimize] if use_transformer else {} -%}
 {%- set listener_components = ["tagger", "morphologizer", "parser", "ner", "textcat", "textcat_multilabel", "entity_linker", "spancat", "trainable_lemmatizer"] -%}
 [paths]

spacy/cli/templates/quickstart_training_recommendations.yml

@@ -37,6 +37,15 @@ bn:
     accuracy:
       name: sagorsarker/bangla-bert-base
       size_factor: 3
+ca:
+  word_vectors: null
+  transformer:
+    efficiency:
+      name: projecte-aina/roberta-base-ca-v2
+      size_factor: 3
+    accuracy:
+      name: projecte-aina/roberta-base-ca-v2
+      size_factor: 3
 da:
   word_vectors: da_core_news_lg
   transformer:

spacy/default_config.cfg

@@ -90,6 +90,8 @@ dev_corpus = "corpora.dev"
 train_corpus = "corpora.train"
 # Optional callback before nlp object is saved to disk after training
 before_to_disk = null
+# Optional callback that is invoked at the start of each training step
+before_update = null
 
 [training.logger]
 @loggers = "spacy.ConsoleLogger.v1"

spacy/displacy/__init__.py

@@ -228,12 +228,13 @@ def parse_spans(doc: Doc, options: Dict[str, Any] = {}) -> Dict[str, Any]:
             "kb_id": span.kb_id_ if span.kb_id_ else "",
             "kb_url": kb_url_template.format(span.kb_id_) if kb_url_template else "#",
         }
-        for span in doc.spans[spans_key]
+        for span in doc.spans.get(spans_key, [])
     ]
     tokens = [token.text for token in doc]
 
     if not spans:
-        warnings.warn(Warnings.W117.format(spans_key=spans_key))
+        keys = list(doc.spans.keys())
+        warnings.warn(Warnings.W117.format(spans_key=spans_key, keys=keys))
     title = doc.user_data.get("title", None) if hasattr(doc, "user_data") else None
     settings = get_doc_settings(doc)
     return {

spacy/errors.py

@@ -131,13 +131,6 @@ class Warnings(metaclass=ErrorsWithCodes):
             "and make it independent. For example, `replace_listeners = "
             "[\"model.tok2vec\"]` See the documentation for details: "
             "https://spacy.io/usage/training#config-components-listeners")
-    W088 = ("The pipeline component {name} implements a `begin_training` "
-            "method, which won't be called by spaCy. As of v3.0, `begin_training` "
-            "has been renamed to `initialize`, so you likely want to rename the "
-            "component method. See the documentation for details: "
-            "https://spacy.io/api/language#initialize")
-    W089 = ("As of spaCy v3.0, the `nlp.begin_training` method has been renamed "
-            "to `nlp.initialize`.")
     W090 = ("Could not locate any {format} files in path '{path}'.")
     W091 = ("Could not clean/remove the temp directory at {dir}: {msg}.")
     W092 = ("Ignoring annotations for sentence starts, as dependency heads are set.")
@@ -199,7 +192,7 @@ class Warnings(metaclass=ErrorsWithCodes):
     W117 = ("No spans to visualize found in Doc object with spans_key: '{spans_key}'. If this is "
             "surprising to you, make sure the Doc was processed using a model "
             "that supports span categorization, and check the `doc.spans[spans_key]` "
-            "property manually if necessary.")
+            "property manually if necessary.\n\nAvailable keys: {keys}")
     W118 = ("Term '{term}' not found in glossary. It may however be explained in documentation "
             "for the corpora used to train the language. Please check "
             "`nlp.meta[\"sources\"]` for any relevant links.")
@@ -212,8 +205,8 @@ class Warnings(metaclass=ErrorsWithCodes):
     W121 = ("Attempting to trace non-existent method '{method}' in pipe '{pipe}'")
     W122 = ("Couldn't trace method '{method}' in pipe '{pipe}'. This can happen if the pipe class "
             "is a Cython extension type.")
-    W123 = ("Argument {arg} with value {arg_value} is used instead of {config_value} as specified in the config. Be "
+    W123 = ("Argument `enable` with value {enable} does not contain all values specified in the config option "
-            "aware that this might affect other components in your pipeline.")
+            "`enabled` ({enabled}). Be aware that this might affect other components in your pipeline.")
 
 
 class Errors(metaclass=ErrorsWithCodes):
@@ -250,8 +243,6 @@ class Errors(metaclass=ErrorsWithCodes):
             "https://spacy.io/usage/models")
     E011 = ("Unknown operator: '{op}'. Options: {opts}")
     E012 = ("Cannot add pattern for zero tokens to matcher.\nKey: {key}")
-    E016 = ("MultitaskObjective target should be function or one of: dep, "
-            "tag, ent, dep_tag_offset, ent_tag.")
     E017 = ("Can only add 'str' inputs to StringStore. Got type: {value_type}")
     E018 = ("Can't retrieve string for hash '{hash_value}'. This usually "
             "refers to an issue with the `Vocab` or `StringStore`.")
@@ -345,6 +336,11 @@ class Errors(metaclass=ErrorsWithCodes):
             "clear the existing vectors and resize the table.")
     E074 = ("Error interpreting compiled match pattern: patterns are expected "
             "to end with the attribute {attr}. Got: {bad_attr}.")
+    E079 = ("Error computing states in beam: number of predicted beams "
+            "({pbeams}) does not equal number of gold beams ({gbeams}).")
+    E080 = ("Duplicate state found in beam: {key}.")
+    E081 = ("Error getting gradient in beam: number of histories ({n_hist}) "
+            "does not equal number of losses ({losses}).")
     E082 = ("Error deprojectivizing parse: number of heads ({n_heads}), "
             "projective heads ({n_proj_heads}) and labels ({n_labels}) do not "
             "match.")
@@ -544,6 +540,10 @@ class Errors(metaclass=ErrorsWithCodes):
             "during training, make sure to include it in 'annotating components'")
 
     # New errors added in v3.x
+    E851 = ("The 'textcat' component labels should only have values of 0 or 1, "
+            "but found value of '{val}'.")
+    E852 = ("The tar file pulled from the remote attempted an unsafe path "
+            "traversal.")
     E853 = ("Unsupported component factory name '{name}'. The character '.' is "
             "not permitted in factory names.")
     E854 = ("Unable to set doc.ents. Check that the 'ents_filter' does not "
@@ -723,13 +723,6 @@ class Errors(metaclass=ErrorsWithCodes):
             "method in component '{name}'. If you want to use this "
             "method, make sure it's overwritten on the subclass.")
     E940 = ("Found NaN values in scores.")
-    E941 = ("Can't find model '{name}'. It looks like you're trying to load a "
-            "model from a shortcut, which is obsolete as of spaCy v3.0. To "
-            "load the model, use its full name instead:\n\n"
-            "nlp = spacy.load(\"{full}\")\n\nFor more details on the available "
-            "models, see the models directory: https://spacy.io/models. If you "
-            "want to create a blank model, use spacy.blank: "
-            "nlp = spacy.blank(\"{name}\")")
     E942 = ("Executing `after_{name}` callback failed. Expected the function to "
             "return an initialized nlp object but got: {value}. Maybe "
             "you forgot to return the modified object in your function?")
@@ -950,6 +943,7 @@ class Errors(metaclass=ErrorsWithCodes):
              "sure it's overwritten on the subclass.")
     E1046 = ("{cls_name} is an abstract class and cannot be instantiated. If you are looking for spaCy's default "
              "knowledge base, use `InMemoryLookupKB`.")
+    E1047 = ("`find_threshold()` only supports components with a `scorer` attribute.")
 
     # v4 error strings
     E4000 = ("Expected a Doc as input, but got: '{type}'")
@@ -969,7 +963,6 @@ OLD_MODEL_SHORTCUTS = {
 RENAMED_LANGUAGE_CODES = {"xx": "mul", "is": "isl"}
 
 
-
 class MatchPatternError(ValueError):
     def __init__(self, key, errors):
         """Custom error for validating match patterns.

spacy/lang/nl/stop_words.py

@@ -15,7 +15,7 @@
 
 STOP_WORDS = set(
     """
-aan af al alle alles allebei alleen allen als altijd ander anders andere anderen aangaangde aangezien achter achterna
+aan af al alle alles allebei alleen allen als altijd ander anders andere anderen aangaande aangezien achter achterna
 afgelopen aldus alhoewel anderzijds
 
 ben bij bijna bijvoorbeeld behalve beide beiden beneden bent bepaald beter betere betreffende binnen binnenin boven

spacy/lang/ru/lemmatizer.py

@@ -28,34 +28,39 @@ class RussianLemmatizer(Lemmatizer):
                 from pymorphy2 import MorphAnalyzer
             except ImportError:
                 raise ImportError(
-                    "The Russian lemmatizer mode 'pymorphy2' requires the "
+                    "The lemmatizer mode 'pymorphy2' requires the "
-                    "pymorphy2 library. Install it with: pip install pymorphy2"
+                    "pymorphy2 library and dictionaries. Install them with: "
+                    "pip install pymorphy2"
+                    "# for Ukrainian dictionaries:"
+                    "pip install pymorphy2-dicts-uk"
                 ) from None
             if getattr(self, "_morph", None) is None:
-                self._morph = MorphAnalyzer()
+                self._morph = MorphAnalyzer(lang="ru")
-        elif mode == "pymorphy3":
+        elif mode in {"pymorphy3", "pymorphy3_lookup"}:
             try:
                 from pymorphy3 import MorphAnalyzer
             except ImportError:
                 raise ImportError(
-                    "The Russian lemmatizer mode 'pymorphy3' requires the "
+                    "The lemmatizer mode 'pymorphy3' requires the "
-                    "pymorphy3 library. Install it with: pip install pymorphy3"
+                    "pymorphy3 library and dictionaries. Install them with: "
+                    "pip install pymorphy3"
+                    "# for Ukrainian dictionaries:"
+                    "pip install pymorphy3-dicts-uk"
                 ) from None
             if getattr(self, "_morph", None) is None:
-                self._morph = MorphAnalyzer()
+                self._morph = MorphAnalyzer(lang="ru")
         super().__init__(
             vocab, model, name, mode=mode, overwrite=overwrite, scorer=scorer
         )
 
-    def pymorphy2_lemmatize(self, token: Token) -> List[str]:
+    def _pymorphy_lemmatize(self, token: Token) -> List[str]:
         string = token.text
         univ_pos = token.pos_
         morphology = token.morph.to_dict()
         if univ_pos == "PUNCT":
             return [PUNCT_RULES.get(string, string)]
         if univ_pos not in ("ADJ", "DET", "NOUN", "NUM", "PRON", "PROPN", "VERB"):
-            # Skip unchangeable pos
+            return self._pymorphy_lookup_lemmatize(token)
-            return [string.lower()]
         analyses = self._morph.parse(string)
         filtered_analyses = []
         for analysis in analyses:
@@ -63,8 +68,10 @@ class RussianLemmatizer(Lemmatizer):
                 # Skip suggested parse variant for unknown word for pymorphy
                 continue
             analysis_pos, _ = oc2ud(str(analysis.tag))
-            if analysis_pos == univ_pos or (
+            if (
-                analysis_pos in ("NOUN", "PROPN") and univ_pos in ("NOUN", "PROPN")
+                analysis_pos == univ_pos
+                or (analysis_pos in ("NOUN", "PROPN") and univ_pos in ("NOUN", "PROPN"))
+                or ((analysis_pos == "PRON") and (univ_pos == "DET"))
             ):
                 filtered_analyses.append(analysis)
         if not len(filtered_analyses):
@@ -107,15 +114,27 @@ class RussianLemmatizer(Lemmatizer):
             dict.fromkeys([analysis.normal_form for analysis in filtered_analyses])
         )
 
-    def pymorphy2_lookup_lemmatize(self, token: Token) -> List[str]:
+    def _pymorphy_lookup_lemmatize(self, token: Token) -> List[str]:
         string = token.text
         analyses = self._morph.parse(string)
-        if len(analyses) == 1:
+        # often multiple forms would derive from the same normal form
-            return [analyses[0].normal_form]
+        # thus check _unique_ normal forms
+        normal_forms = set([an.normal_form for an in analyses])
+        if len(normal_forms) == 1:
+            return [next(iter(normal_forms))]
         return [string]
 
+    def pymorphy2_lemmatize(self, token: Token) -> List[str]:
+        return self._pymorphy_lemmatize(token)
+
+    def pymorphy2_lookup_lemmatize(self, token: Token) -> List[str]:
+        return self._pymorphy_lookup_lemmatize(token)
+
     def pymorphy3_lemmatize(self, token: Token) -> List[str]:
-        return self.pymorphy2_lemmatize(token)
+        return self._pymorphy_lemmatize(token)
+
+    def pymorphy3_lookup_lemmatize(self, token: Token) -> List[str]:
+        return self._pymorphy_lookup_lemmatize(token)
 
 
 def oc2ud(oc_tag: str) -> Tuple[str, Dict[str, str]]:

spacy/lang/ru/tokenizer_exceptions.py

@@ -61,6 +61,11 @@ for abbr in [
     {ORTH: "2к23", NORM: "2023"},
     {ORTH: "2к24", NORM: "2024"},
     {ORTH: "2к25", NORM: "2025"},
+    {ORTH: "2к26", NORM: "2026"},
+    {ORTH: "2к27", NORM: "2027"},
+    {ORTH: "2к28", NORM: "2028"},
+    {ORTH: "2к29", NORM: "2029"},
+    {ORTH: "2к30", NORM: "2030"},
 ]:
     _exc[abbr[ORTH]] = [abbr]
 
@@ -268,8 +273,8 @@ for abbr in [
     {ORTH: "з-ка", NORM: "заимка"},
     {ORTH: "п-к", NORM: "починок"},
     {ORTH: "киш.", NORM: "кишлак"},
-    {ORTH: "п. ст. ", NORM: "поселок станция"},
+    {ORTH: "п. ст.", NORM: "поселок станция"},
-    {ORTH: "п. ж/д ст. ", NORM: "поселок при железнодорожной станции"},
+    {ORTH: "п. ж/д ст.", NORM: "поселок при железнодорожной станции"},
     {ORTH: "ж/д бл-ст", NORM: "железнодорожный блокпост"},
     {ORTH: "ж/д б-ка", NORM: "железнодорожная будка"},
     {ORTH: "ж/д в-ка", NORM: "железнодорожная ветка"},
@@ -280,12 +285,12 @@ for abbr in [
     {ORTH: "ж/д п.п.", NORM: "железнодорожный путевой пост"},
     {ORTH: "ж/д о.п.", NORM: "железнодорожный остановочный пункт"},
     {ORTH: "ж/д рзд.", NORM: "железнодорожный разъезд"},
-    {ORTH: "ж/д ст. ", NORM: "железнодорожная станция"},
+    {ORTH: "ж/д ст.", NORM: "железнодорожная станция"},
     {ORTH: "м-ко", NORM: "местечко"},
     {ORTH: "д.", NORM: "деревня"},
     {ORTH: "с.", NORM: "село"},
     {ORTH: "сл.", NORM: "слобода"},
-    {ORTH: "ст. ", NORM: "станция"},
+    {ORTH: "ст.", NORM: "станция"},
     {ORTH: "ст-ца", NORM: "станица"},
     {ORTH: "у.", NORM: "улус"},
     {ORTH: "х.", NORM: "хутор"},
@@ -388,8 +393,9 @@ for abbr in [
     {ORTH: "прим.", NORM: "примечание"},
     {ORTH: "прим.ред.", NORM: "примечание редакции"},
     {ORTH: "см. также", NORM: "смотри также"},
-    {ORTH: "кв.м.", NORM: "квадрантный метр"},
+    {ORTH: "см.", NORM: "смотри"},
-    {ORTH: "м2", NORM: "квадрантный метр"},
+    {ORTH: "кв.м.", NORM: "квадратный метр"},
+    {ORTH: "м2", NORM: "квадратный метр"},
     {ORTH: "б/у", NORM: "бывший в употреблении"},
     {ORTH: "сокр.", NORM: "сокращение"},
     {ORTH: "чел.", NORM: "человек"},

spacy/lang/uk/lemmatizer.py

@@ -29,7 +29,7 @@ class UkrainianLemmatizer(RussianLemmatizer):
                 ) from None
             if getattr(self, "_morph", None) is None:
                 self._morph = MorphAnalyzer(lang="uk")
-        elif mode == "pymorphy3":
+        elif mode in {"pymorphy3", "pymorphy3_lookup"}:
             try:
                 from pymorphy3 import MorphAnalyzer
             except ImportError:

spacy/language.py

@@ -43,8 +43,7 @@ from .lookups import load_lookups
 from .compat import Literal
 
 
-if TYPE_CHECKING:
+PipeCallable = Callable[[Doc], Doc]
-    from .pipeline import Pipe  # noqa: F401
 
 
 # This is the base config will all settings (training etc.)
@@ -181,7 +180,7 @@ class Language:
         self.vocab: Vocab = vocab
         if self.lang is None:
             self.lang = self.vocab.lang
-        self._components: List[Tuple[str, "Pipe"]] = []
+        self._components: List[Tuple[str, PipeCallable]] = []
         self._disabled: Set[str] = set()
         self.max_length = max_length
         # Create the default tokenizer from the default config
@@ -303,7 +302,7 @@ class Language:
         return SimpleFrozenList(names)
 
     @property
-    def components(self) -> List[Tuple[str, "Pipe"]]:
+    def components(self) -> List[Tuple[str, PipeCallable]]:
         """Get all (name, component) tuples in the pipeline, including the
         currently disabled components.
         """
@@ -322,12 +321,12 @@ class Language:
         return SimpleFrozenList(names, error=Errors.E926.format(attr="component_names"))
 
     @property
-    def pipeline(self) -> List[Tuple[str, "Pipe"]]:
+    def pipeline(self) -> List[Tuple[str, PipeCallable]]:
         """The processing pipeline consisting of (name, component) tuples. The
         components are called on the Doc in order as it passes through the
         pipeline.
 
-        RETURNS (List[Tuple[str, Pipe]]): The pipeline.
+        RETURNS (List[Tuple[str, Callable[[Doc], Doc]]]): The pipeline.
         """
         pipes = [(n, p) for n, p in self._components if n not in self._disabled]
         return SimpleFrozenList(pipes, error=Errors.E926.format(attr="pipeline"))
@@ -527,7 +526,7 @@ class Language:
         assigns: Iterable[str] = SimpleFrozenList(),
         requires: Iterable[str] = SimpleFrozenList(),
         retokenizes: bool = False,
-        func: Optional["Pipe"] = None,
+        func: Optional[PipeCallable] = None,
     ) -> Callable[..., Any]:
         """Register a new pipeline component. Can be used for stateless function
         components that don't require a separate factory. Can be used as a
@@ -542,7 +541,7 @@ class Language:
             e.g. "token.ent_id". Used for pipeline analysis.
         retokenizes (bool): Whether the component changes the tokenization.
             Used for pipeline analysis.
-        func (Optional[Callable]): Factory function if not used as a decorator.
+        func (Optional[Callable[[Doc], Doc]): Factory function if not used as a decorator.
 
         DOCS: https://spacy.io/api/language#component
         """
@@ -553,11 +552,11 @@ class Language:
                 raise ValueError(Errors.E853.format(name=name))
         component_name = name if name is not None else util.get_object_name(func)
 
-        def add_component(component_func: "Pipe") -> Callable:
+        def add_component(component_func: PipeCallable) -> Callable:
             if isinstance(func, type):  # function is a class
                 raise ValueError(Errors.E965.format(name=component_name))
 
-            def factory_func(nlp, name: str) -> "Pipe":
+            def factory_func(nlp, name: str) -> PipeCallable:
                 return component_func
 
             internal_name = cls.get_factory_name(name)
@@ -607,7 +606,7 @@ class Language:
             print_pipe_analysis(analysis, keys=keys)
         return analysis
 
-    def get_pipe(self, name: str) -> "Pipe":
+    def get_pipe(self, name: str) -> PipeCallable:
         """Get a pipeline component for a given component name.
 
         name (str): Name of pipeline component to get.
@@ -628,7 +627,7 @@ class Language:
         config: Dict[str, Any] = SimpleFrozenDict(),
         raw_config: Optional[Config] = None,
         validate: bool = True,
-    ) -> "Pipe":
+    ) -> PipeCallable:
         """Create a pipeline component. Mostly used internally. To create and
         add a component to the pipeline, you can use nlp.add_pipe.
 
@@ -640,7 +639,7 @@ class Language:
         raw_config (Optional[Config]): Internals: the non-interpolated config.
         validate (bool): Whether to validate the component config against the
             arguments and types expected by the factory.
-        RETURNS (Pipe): The pipeline component.
+        RETURNS (Callable[[Doc], Doc]): The pipeline component.
 
         DOCS: https://spacy.io/api/language#create_pipe
         """
@@ -695,24 +694,18 @@ class Language:
 
     def create_pipe_from_source(
         self, source_name: str, source: "Language", *, name: str
-    ) -> Tuple["Pipe", str]:
+    ) -> Tuple[PipeCallable, str]:
         """Create a pipeline component by copying it from an existing model.
 
         source_name (str): Name of the component in the source pipeline.
         source (Language): The source nlp object to copy from.
         name (str): Optional alternative name to use in current pipeline.
-        RETURNS (Tuple[Callable, str]): The component and its factory name.
+        RETURNS (Tuple[Callable[[Doc], Doc], str]): The component and its factory name.
         """
         # Check source type
         if not isinstance(source, Language):
             raise ValueError(Errors.E945.format(name=source_name, source=type(source)))
-        # Check vectors, with faster checks first
+        if self.vocab.vectors != source.vocab.vectors:
-        if (
-            self.vocab.vectors.shape != source.vocab.vectors.shape
-            or self.vocab.vectors.key2row != source.vocab.vectors.key2row
-            or self.vocab.vectors.to_bytes(exclude=["strings"])
-            != source.vocab.vectors.to_bytes(exclude=["strings"])
-        ):
             warnings.warn(Warnings.W113.format(name=source_name))
         if source_name not in source.component_names:
             raise KeyError(
@@ -746,7 +739,7 @@ class Language:
         config: Dict[str, Any] = SimpleFrozenDict(),
         raw_config: Optional[Config] = None,
         validate: bool = True,
-    ) -> "Pipe":
+    ) -> PipeCallable:
         """Add a component to the processing pipeline. Valid components are
         callables that take a `Doc` object, modify it and return it. Only one
         of before/after/first/last can be set. Default behaviour is "last".
@@ -769,7 +762,7 @@ class Language:
         raw_config (Optional[Config]): Internals: the non-interpolated config.
         validate (bool): Whether to validate the component config against the
             arguments and types expected by the factory.
-        RETURNS (Pipe): The pipeline component.
+        RETURNS (Callable[[Doc], Doc]): The pipeline component.
 
         DOCS: https://spacy.io/api/language#add_pipe
         """
@@ -790,14 +783,6 @@ class Language:
                 factory_name, source, name=name
             )
         else:
-            if not self.has_factory(factory_name):
-                err = Errors.E002.format(
-                    name=factory_name,
-                    opts=", ".join(self.factory_names),
-                    method="add_pipe",
-                    lang=util.get_object_name(self),
-                    lang_code=self.lang,
-                )
             pipe_component = self.create_pipe(
                 factory_name,
                 name=name,
@@ -883,7 +868,7 @@ class Language:
         *,
         config: Dict[str, Any] = SimpleFrozenDict(),
         validate: bool = True,
-    ) -> "Pipe":
+    ) -> PipeCallable:
         """Replace a component in the pipeline.
 
         name (str): Name of the component to replace.
@@ -892,7 +877,7 @@ class Language:
             component. Will be merged with default config, if available.
         validate (bool): Whether to validate the component config against the
             arguments and types expected by the factory.
-        RETURNS (Pipe): The new pipeline component.
+        RETURNS (Callable[[Doc], Doc]): The new pipeline component.
 
         DOCS: https://spacy.io/api/language#replace_pipe
         """
@@ -944,11 +929,11 @@ class Language:
             init_cfg = self._config["initialize"]["components"].pop(old_name)
             self._config["initialize"]["components"][new_name] = init_cfg
 
-    def remove_pipe(self, name: str) -> Tuple[str, "Pipe"]:
+    def remove_pipe(self, name: str) -> Tuple[str, PipeCallable]:
         """Remove a component from the pipeline.
 
         name (str): Name of the component to remove.
-        RETURNS (tuple): A `(name, component)` tuple of the removed component.
+        RETURNS (Tuple[str, Callable[[Doc], Doc]]): A `(name, component)` tuple of the removed component.
 
         DOCS: https://spacy.io/api/language#remove_pipe
         """
@@ -1254,15 +1239,6 @@ class Language:
             sgd(key, W, dW)  # type: ignore[call-arg, misc]
         return losses
 
-    def begin_training(
-        self,
-        get_examples: Optional[Callable[[], Iterable[Example]]] = None,
-        *,
-        sgd: Optional[Optimizer] = None,
-    ) -> Optimizer:
-        warnings.warn(Warnings.W089, DeprecationWarning)
-        return self.initialize(get_examples, sgd=sgd)
-
     def initialize(
         self,
         get_examples: Optional[Callable[[], Iterable[Example]]] = None,
@@ -1363,15 +1339,15 @@ class Language:
 
     def set_error_handler(
         self,
-        error_handler: Callable[[str, "Pipe", List[Doc], Exception], NoReturn],
+        error_handler: Callable[[str, PipeCallable, List[Doc], Exception], NoReturn],
     ):
-        """Set an error handler object for all the components in the pipeline that implement
+        """Set an error handler object for all the components in the pipeline
-        a set_error_handler function.
+        that implement a set_error_handler function.
 
-        error_handler (Callable[[str, Pipe, List[Doc], Exception], NoReturn]):
+        error_handler (Callable[[str, Callable[[Doc], Doc], List[Doc], Exception], NoReturn]):
-            Function that deals with a failing batch of documents. This callable function should take in
+            Function that deals with a failing batch of documents. This callable
-            the component's name, the component itself, the offending batch of documents, and the exception
+            function should take in the component's name, the component itself,
-            that was thrown.
+            the offending batch of documents, and the exception that was thrown.
         DOCS: https://spacy.io/api/language#set_error_handler
         """
         self.default_error_handler = error_handler
@@ -1879,31 +1855,22 @@ class Language:
         if isinstance(exclude, str):
             exclude = [exclude]
 
-        def fetch_pipes_status(value: Iterable[str], key: str) -> Iterable[str]:
+        # `enable` should not be merged with `enabled` (the opposite is true for `disable`/`disabled`). If the config
-            """Fetch value for `enable` or `disable` w.r.t. the specified config and passed arguments passed to
+        # specifies values for `enabled` not included in `enable`, emit warning.
-            .load(). If both arguments and config specified values for this field, the passed arguments take precedence
+        if id(enable) != id(_DEFAULT_EMPTY_PIPES):
-            and a warning is printed.
+            enabled = config["nlp"].get("enabled", [])
-            value (Iterable[str]): Passed value for `enable` or `disable`.
+            if len(enabled) and not set(enabled).issubset(enable):
-            key (str): Key for field in config (either "enabled" or "disabled").
+                warnings.warn(
-            RETURN (Iterable[str]):
+                    Warnings.W123.format(
-            """
+                        enable=enable,
-            # We assume that no argument was passed if the value is the specified default value.
+                        enabled=enabled,
-            if id(value) == id(_DEFAULT_EMPTY_PIPES):
-                return config["nlp"].get(key, [])
-            else:
-                if len(config["nlp"].get(key, [])):
-                    warnings.warn(
-                        Warnings.W123.format(
-                            arg=key[:-1],
-                            arg_value=value,
-                            config_value=config["nlp"][key],
-                        )
                     )
-                return value
+                )
 
+        # Ensure sets of disabled/enabled pipe names are not contradictory.
         disabled_pipes = cls._resolve_component_status(
-            fetch_pipes_status(disable, "disabled"),
+            list({*disable, *config["nlp"].get("disabled", [])}),
-            fetch_pipes_status(enable, "enabled"),
+            enable,
             config["nlp"]["pipeline"],
         )
         nlp._disabled = set(p for p in disabled_pipes if p not in exclude)
@@ -2084,10 +2051,12 @@ class Language:
         if enable:
             if isinstance(enable, str):
                 enable = [enable]
-            to_disable = [
+            to_disable = {
-                pipe_name for pipe_name in pipe_names if pipe_name not in enable
+                *[pipe_name for pipe_name in pipe_names if pipe_name not in enable],
-            ]
+                *disable,
-            if disable and disable != to_disable:
+            }
+            # If any pipe to be enabled is in to_disable, the specification is inconsistent.
+            if len(set(enable) & to_disable):
                 raise ValueError(Errors.E1042.format(enable=enable, disable=disable))
 
         return tuple(to_disable)

spacy/lexeme.pyi

@@ -20,7 +20,6 @@ class Lexeme:
     def vector_norm(self) -> float: ...
     vector: Floats1d
     rank: int
-    sentiment: float
     @property
     def orth_(self) -> str: ...
     @property

spacy/lexeme.pyx

@@ -173,19 +173,6 @@ cdef class Lexeme:
         def __set__(self, value):
             self.c.id = value
 
-    property sentiment:
-        """RETURNS (float): A scalar value indicating the positivity or
-            negativity of the lexeme."""
-        def __get__(self):
-            sentiment_table = self.vocab.lookups.get_table("lexeme_sentiment", {})
-            return sentiment_table.get(self.c.orth, 0.0)
-
-        def __set__(self, float x):
-            if "lexeme_sentiment" not in self.vocab.lookups:
-                self.vocab.lookups.add_table("lexeme_sentiment")
-            sentiment_table = self.vocab.lookups.get_table("lexeme_sentiment")
-            sentiment_table[self.c.orth] = x
-
     @property
     def orth_(self):
         """RETURNS (str): The original verbatim text of the lexeme

spacy/matcher/matcher.pyx

@@ -1,4 +1,4 @@
-# cython: infer_types=True, cython: profile=True
+# cython: infer_types=True, profile=True
 from typing import List, Iterable
 
 from libcpp.vector cimport vector

spacy/ml/parser_model.pyx

@@ -3,7 +3,6 @@ cimport numpy as np
 from libc.math cimport exp
 from libc.string cimport memset, memcpy
 from libc.stdlib cimport calloc, free, realloc
-from thinc.backends.linalg cimport Vec, VecVec
 from thinc.backends.cblas cimport saxpy, sgemm
 
 import numpy
@@ -102,11 +101,10 @@ cdef void predict_states(CBlas cblas, ActivationsC* A, StateC** states,
     sum_state_features(cblas, A.unmaxed,
         W.feat_weights, A.token_ids, n.states, n.feats, n.hiddens * n.pieces)
     for i in range(n.states):
-        VecVec.add_i(&A.unmaxed[i*n.hiddens*n.pieces],
+        saxpy(cblas)(n.hiddens * n.pieces, 1., W.feat_bias, 1, &A.unmaxed[i*n.hiddens*n.pieces], 1)
-            W.feat_bias, 1., n.hiddens * n.pieces)
         for j in range(n.hiddens):
             index = i * n.hiddens * n.pieces + j * n.pieces
-            which = Vec.arg_max(&A.unmaxed[index], n.pieces)
+            which = _arg_max(&A.unmaxed[index], n.pieces)
             A.hiddens[i*n.hiddens + j] = A.unmaxed[index + which]
     memset(A.scores, 0, n.states * n.classes * sizeof(float))
     if W.hidden_weights == NULL:
@@ -119,8 +117,7 @@ cdef void predict_states(CBlas cblas, ActivationsC* A, StateC** states,
             0.0, A.scores, n.classes)
         # Add bias
         for i in range(n.states):
-            VecVec.add_i(&A.scores[i*n.classes],
+            saxpy(cblas)(n.classes, 1., W.hidden_bias, 1, &A.scores[i*n.classes], 1)
-                W.hidden_bias, 1., n.classes)
     # Set unseen classes to minimum value
     i = 0
     min_ = A.scores[0]
@@ -158,7 +155,8 @@ cdef void cpu_log_loss(float* d_scores,
     """Do multi-label log loss"""
     cdef double max_, gmax, Z, gZ
     best = arg_max_if_gold(scores, costs, is_valid, O)
-    guess = Vec.arg_max(scores, O)
+    guess = _arg_max(scores, O)
+
     if best == -1 or guess == -1:
         # These shouldn't happen, but if they do, we want to make sure we don't
         # cause an OOB access.
@@ -488,3 +486,15 @@ cdef class precompute_hiddens:
             return d_best.reshape((d_best.shape + (1,)))
  
         return state_vector, backprop_relu
+
+cdef inline int _arg_max(const float* scores, const int n_classes) nogil:
+    if n_classes == 2:
+        return 0 if scores[0] > scores[1] else 1
+    cdef int i
+    cdef int best = 0
+    cdef float mode = scores[0]
+    for i in range(1, n_classes):
+        if scores[i] > mode:
+            mode = scores[i]
+            best = i
+    return best

spacy/pipeline/_parser_internals/_beam_utils.pxd

@@ -1,6 +1,6 @@
 from ...typedefs cimport class_t, hash_t
 
-# These are passed as callbacks to thinc.search.Beam
+# These are passed as callbacks to .search.Beam
 cdef int transition_state(void* _dest, void* _src, class_t clas, void* _moves) except -1
 
 cdef int check_final_state(void* _state, void* extra_args) except -1

spacy/pipeline/_parser_internals/_beam_utils.pyx

@@ -3,17 +3,16 @@
 cimport numpy as np
 import numpy
 from cpython.ref cimport PyObject, Py_XDECREF
-from thinc.extra.search cimport Beam
-from thinc.extra.search import MaxViolation
-from thinc.extra.search cimport MaxViolation
 
 from ...typedefs cimport hash_t, class_t
 from .transition_system cimport TransitionSystem, Transition
 from ...errors import Errors
+from .search cimport Beam, MaxViolation
+from .search import MaxViolation
 from .stateclass cimport StateC, StateClass
 
 
-# These are passed as callbacks to thinc.search.Beam
+# These are passed as callbacks to .search.Beam
 cdef int transition_state(void* _dest, void* _src, class_t clas, void* _moves) except -1:
     dest = <StateC*>_dest
     src = <StateC*>_src

spacy/pipeline/_parser_internals/arc_eager.pyx

@@ -15,7 +15,7 @@ from ...training.example cimport Example
 from .stateclass cimport StateClass
 from ._state cimport StateC, ArcC
 from ...errors import Errors
-from thinc.extra.search cimport Beam
+from .search cimport Beam
 
 cdef weight_t MIN_SCORE = -90000
 cdef attr_t SUBTOK_LABEL = hash_string('subtok')

spacy/pipeline/_parser_internals/ner.pyx

@@ -6,7 +6,6 @@ from libcpp.vector cimport vector
 from cymem.cymem cimport Pool
 
 from collections import Counter
-from thinc.extra.search cimport Beam
 
 from ...tokens.doc cimport Doc
 from ...tokens.span import Span
@@ -17,6 +16,7 @@ from ...attrs cimport IS_SPACE
 from ...structs cimport TokenC, SpanC
 from ...training import split_bilu_label
 from ...training.example cimport Example
+from .search cimport Beam
 from .stateclass cimport StateClass
 from ._state cimport StateC
 from .transition_system cimport Transition, do_func_t

spacy/pipeline/_parser_internals/search.pxd

@@ -0,0 +1,89 @@
+from cymem.cymem cimport Pool
+
+from libc.stdint cimport uint32_t
+from libc.stdint cimport uint64_t
+from libcpp.pair cimport pair
+from libcpp.queue cimport priority_queue
+from libcpp.vector cimport vector
+
+from ...typedefs cimport class_t, weight_t, hash_t
+
+ctypedef pair[weight_t, size_t] Entry
+ctypedef priority_queue[Entry] Queue
+
+
+ctypedef int (*trans_func_t)(void* dest, void* src, class_t clas, void* x) except -1
+
+ctypedef void* (*init_func_t)(Pool mem, int n, void* extra_args) except NULL
+
+ctypedef int (*del_func_t)(Pool mem, void* state, void* extra_args) except -1
+
+ctypedef int (*finish_func_t)(void* state, void* extra_args) except -1
+
+ctypedef hash_t (*hash_func_t)(void* state, void* x) except 0
+
+
+cdef struct _State:
+    void* content
+    class_t* hist
+    weight_t score
+    weight_t loss
+    int i
+    int t
+    bint is_done
+
+
+cdef class Beam:
+    cdef Pool mem
+    cdef class_t nr_class
+    cdef class_t width
+    cdef class_t size
+    cdef public weight_t min_density
+    cdef int t
+    cdef readonly bint is_done
+    cdef list histories
+    cdef list _parent_histories
+    cdef weight_t** scores
+    cdef int** is_valid
+    cdef weight_t** costs
+    cdef _State* _parents
+    cdef _State* _states
+    cdef del_func_t del_func
+
+    cdef int _fill(self, Queue* q, weight_t** scores, int** is_valid) except -1
+
+    cdef inline void* at(self, int i) nogil:
+        return self._states[i].content
+
+    cdef int initialize(self, init_func_t init_func, del_func_t del_func, int n, void* extra_args) except -1
+    cdef int advance(self, trans_func_t transition_func, hash_func_t hash_func,
+                     void* extra_args) except -1
+    cdef int check_done(self, finish_func_t finish_func, void* extra_args) except -1
+ 
+
+    cdef inline void set_cell(self, int i, int j, weight_t score, int is_valid, weight_t cost) nogil:
+        self.scores[i][j] = score
+        self.is_valid[i][j] = is_valid
+        self.costs[i][j] = cost
+
+    cdef int set_row(self, int i, const weight_t* scores, const int* is_valid,
+                     const weight_t* costs) except -1
+    cdef int set_table(self, weight_t** scores, int** is_valid, weight_t** costs) except -1
+
+
+cdef class MaxViolation:
+    cdef Pool mem
+    cdef weight_t cost
+    cdef weight_t delta
+    cdef readonly weight_t p_score
+    cdef readonly weight_t g_score
+    cdef readonly double Z
+    cdef readonly double gZ
+    cdef class_t n
+    cdef readonly list p_hist
+    cdef readonly list g_hist
+    cdef readonly list p_probs
+    cdef readonly list g_probs
+
+    cpdef int check(self, Beam pred, Beam gold) except -1
+    cpdef int check_crf(self, Beam pred, Beam gold) except -1

spacy/pipeline/_parser_internals/search.pyx

@@ -0,0 +1,306 @@
+# cython: profile=True, experimental_cpp_class_def=True, cdivision=True, infer_types=True
+cimport cython
+from libc.string cimport memset, memcpy
+from libc.math cimport log, exp
+import math
+
+from cymem.cymem cimport Pool
+from preshed.maps cimport PreshMap
+
+
+cdef class Beam:
+    def __init__(self, class_t nr_class, class_t width, weight_t min_density=0.0):
+        assert nr_class != 0
+        assert width != 0
+        self.nr_class = nr_class
+        self.width = width
+        self.min_density = min_density
+        self.size = 1
+        self.t = 0
+        self.mem = Pool()
+        self.del_func = NULL
+        self._parents = <_State*>self.mem.alloc(self.width, sizeof(_State))
+        self._states = <_State*>self.mem.alloc(self.width, sizeof(_State))
+        cdef int i
+        self.histories = [[] for i in range(self.width)]
+        self._parent_histories = [[] for i in range(self.width)]
+
+        self.scores = <weight_t**>self.mem.alloc(self.width, sizeof(weight_t*))
+        self.is_valid = <int**>self.mem.alloc(self.width, sizeof(weight_t*))
+        self.costs = <weight_t**>self.mem.alloc(self.width, sizeof(weight_t*))
+        for i in range(self.width):
+            self.scores[i] = <weight_t*>self.mem.alloc(self.nr_class, sizeof(weight_t))
+            self.is_valid[i] = <int*>self.mem.alloc(self.nr_class, sizeof(int))
+            self.costs[i] = <weight_t*>self.mem.alloc(self.nr_class, sizeof(weight_t))
+
+    def __len__(self):
+        return self.size
+
+    property score:
+        def __get__(self):
+            return self._states[0].score
+
+    property min_score:
+        def __get__(self):
+            return self._states[self.size-1].score
+
+    property loss:
+        def __get__(self):
+            return self._states[0].loss
+
+    property probs:
+        def __get__(self):
+            return _softmax([self._states[i].score for i in range(self.size)])
+
+    property scores:
+        def __get__(self):
+            return [self._states[i].score for i in range(self.size)]
+
+    property histories:
+        def __get__(self):
+            return self.histories
+
+    cdef int set_row(self, int i, const weight_t* scores, const int* is_valid,
+                     const weight_t* costs) except -1:
+        cdef int j
+        for j in range(self.nr_class):
+            self.scores[i][j] = scores[j]
+            self.is_valid[i][j] = is_valid[j]
+            self.costs[i][j] = costs[j]
+
+    cdef int set_table(self, weight_t** scores, int** is_valid, weight_t** costs) except -1:
+        cdef int i, j
+        for i in range(self.width):
+            memcpy(self.scores[i], scores[i], sizeof(weight_t) * self.nr_class)
+            memcpy(self.is_valid[i], is_valid[i], sizeof(bint) * self.nr_class)
+            memcpy(self.costs[i], costs[i], sizeof(int) * self.nr_class)
+
+    cdef int initialize(self, init_func_t init_func, del_func_t del_func, int n, void* extra_args) except -1:
+        for i in range(self.width):
+            self._states[i].content = init_func(self.mem, n, extra_args)
+            self._parents[i].content = init_func(self.mem, n, extra_args)
+        self.del_func = del_func
+
+    def __dealloc__(self):
+        if self.del_func == NULL:
+            return
+
+        for i in range(self.width):
+            self.del_func(self.mem, self._states[i].content, NULL)
+            self.del_func(self.mem, self._parents[i].content, NULL)
+
+    @cython.cdivision(True)
+    cdef int advance(self, trans_func_t transition_func, hash_func_t hash_func,
+                     void* extra_args) except -1:
+        cdef weight_t** scores = self.scores
+        cdef int** is_valid = self.is_valid
+        cdef weight_t** costs = self.costs
+
+        cdef Queue* q = new Queue()
+        self._fill(q, scores, is_valid)
+        # For a beam of width k, we only ever need 2k state objects. How?
+        # Each transition takes a parent and a class and produces a new state.
+        # So, we don't need the whole history --- just the parent. So at
+        # each step, we take a parent, and apply one or more extensions to
+        # it.
+        self._parents, self._states = self._states, self._parents
+        self._parent_histories, self.histories = self.histories, self._parent_histories
+        cdef weight_t score
+        cdef int p_i
+        cdef int i = 0
+        cdef class_t clas
+        cdef _State* parent
+        cdef _State* state
+        cdef hash_t key
+        cdef PreshMap seen_states = PreshMap(self.width)
+        cdef uint64_t is_seen
+        cdef uint64_t one = 1
+        while i < self.width and not q.empty():
+            data = q.top()
+            p_i = data.second / self.nr_class
+            clas = data.second % self.nr_class
+            score = data.first
+            q.pop()
+            parent = &self._parents[p_i]
+            # Indicates terminal state reached; i.e. state is done
+            if parent.is_done:
+                # Now parent will not be changed, so we don't have to copy.
+                # Once finished, should also be unbranching.
+                self._states[i], parent[0] = parent[0], self._states[i]
+                parent.i = self._states[i].i
+                parent.t = self._states[i].t
+                parent.is_done = self._states[i].t
+                self._states[i].score = score
+                self.histories[i] = list(self._parent_histories[p_i])
+                i += 1
+            else:
+                state = &self._states[i]
+                # The supplied transition function should adjust the destination
+                # state to be the result of applying the class to the source state
+                transition_func(state.content, parent.content, clas, extra_args)
+                key = hash_func(state.content, extra_args) if hash_func is not NULL else 0
+                is_seen = <uint64_t>seen_states.get(key)
+                if key == 0 or key == 1 or not is_seen:
+                    if key != 0 and key != 1:
+                        seen_states.set(key, <void*>one)
+                    state.score = score
+                    state.loss = parent.loss + costs[p_i][clas]
+                    self.histories[i] = list(self._parent_histories[p_i])
+                    self.histories[i].append(clas)
+                    i += 1
+        del q
+        self.size = i
+        assert self.size >= 1
+        for i in range(self.width):
+            memset(self.scores[i], 0, sizeof(weight_t) * self.nr_class)
+            memset(self.costs[i], 0, sizeof(weight_t) * self.nr_class)
+            memset(self.is_valid[i], 0, sizeof(int) * self.nr_class)
+        self.t += 1
+
+    cdef int check_done(self, finish_func_t finish_func, void* extra_args) except -1:
+        cdef int i
+        for i in range(self.size):
+            if not self._states[i].is_done:
+                self._states[i].is_done = finish_func(self._states[i].content, extra_args)
+        for i in range(self.size):
+            if not self._states[i].is_done:
+                self.is_done = False
+                break
+        else:
+            self.is_done = True
+
+    @cython.cdivision(True)
+    cdef int _fill(self, Queue* q, weight_t** scores, int** is_valid) except -1:
+        """Populate the queue from a k * n matrix of scores, where k is the
+        beam-width, and n is the number of classes.
+        """
+        cdef Entry entry
+        cdef weight_t score
+        cdef _State* s
+        cdef int i, j, move_id
+        assert self.size >= 1
+        cdef vector[Entry] entries
+        for i in range(self.size):
+            s = &self._states[i]
+            move_id = i * self.nr_class
+            if s.is_done:
+                # Update score by path average, following TACL '13 paper.
+                if self.histories[i]:
+                    entry.first = s.score + (s.score / self.t)
+                else:
+                    entry.first = s.score
+                entry.second = move_id
+                entries.push_back(entry)
+            else:
+                for j in range(self.nr_class):
+                    if is_valid[i][j]:
+                        entry.first = s.score + scores[i][j]
+                        entry.second = move_id + j
+                        entries.push_back(entry)
+        cdef double max_, Z, cutoff
+        if self.min_density == 0.0:
+            for i in range(entries.size()):
+                q.push(entries[i])
+        elif not entries.empty():
+            max_ = entries[0].first
+            Z = 0.
+            cutoff = 0.
+            # Softmax into probabilities, so we can prune
+            for i in range(entries.size()):
+                if entries[i].first > max_:
+                    max_ = entries[i].first
+            for i in range(entries.size()):
+                Z += exp(entries[i].first-max_)
+            cutoff = (1. / Z) * self.min_density
+            for i in range(entries.size()):
+                prob = exp(entries[i].first-max_) / Z
+                if prob >= cutoff:
+                    q.push(entries[i])
+
+
+cdef class MaxViolation:
+    def __init__(self):
+        self.p_score = 0.0
+        self.g_score = 0.0
+        self.Z = 0.0
+        self.gZ = 0.0
+        self.delta = -1
+        self.cost = 0
+        self.p_hist = []
+        self.g_hist = []
+        self.p_probs = []
+        self.g_probs = []
+
+    cpdef int check(self, Beam pred, Beam gold) except -1:
+        cdef _State* p = &pred._states[0]
+        cdef _State* g = &gold._states[0]
+        cdef weight_t d = p.score - g.score
+        if p.loss >= 1 and (self.cost == 0 or d > self.delta):
+            self.cost = p.loss
+            self.delta = d
+            self.p_hist = list(pred.histories[0])
+            self.g_hist = list(gold.histories[0])
+            self.p_score = p.score
+            self.g_score = g.score
+            self.Z = 1e-10
+            self.gZ = 1e-10
+            for i in range(pred.size):
+                if pred._states[i].loss > 0:
+                    self.Z += exp(pred._states[i].score)
+            for i in range(gold.size):
+                if gold._states[i].loss == 0:
+                    prob = exp(gold._states[i].score)
+                    self.Z += prob
+                    self.gZ += prob
+
+    cpdef int check_crf(self, Beam pred, Beam gold) except -1:
+        d = pred.score - gold.score
+        seen_golds = set([tuple(gold.histories[i]) for i in range(gold.size)])
+        if pred.loss > 0 and (self.cost == 0 or d > self.delta):
+            p_hist = []
+            p_scores = []
+            g_hist = []
+            g_scores = []
+            for i in range(pred.size):
+                if pred._states[i].loss > 0:
+                    p_scores.append(pred._states[i].score)
+                    p_hist.append(list(pred.histories[i]))
+                # This can happen from non-monotonic actions
+                # If we find a better gold analysis this way, be sure to keep it.
+                elif pred._states[i].loss <= 0 \
+                and tuple(pred.histories[i]) not in seen_golds:
+                    g_scores.append(pred._states[i].score)
+                    g_hist.append(list(pred.histories[i]))
+            for i in range(gold.size):
+                if gold._states[i].loss == 0:
+                    g_scores.append(gold._states[i].score)
+                    g_hist.append(list(gold.histories[i]))
+
+            all_probs = _softmax(p_scores + g_scores)
+            p_probs = all_probs[:len(p_scores)]
+            g_probs_all = all_probs[len(p_scores):]
+            g_probs = _softmax(g_scores)
+
+            self.cost = pred.loss
+            self.delta = d
+            self.p_hist = p_hist
+            self.g_hist = g_hist
+            # TODO: These variables are misnamed! These are the gradients of the loss.
+            self.p_probs = p_probs
+            # Intuition here:
+            # The gradient of the loss is:
+            # P(model) - P(truth)
+            # Normally, P(truth) is 1 for the gold
+            # But, if we want to do the "partial credit" scheme, we want
+            # to create a distribution over the gold, proportional to the scores
+            # awarded.
+            self.g_probs = [x-y for x, y in zip(g_probs_all, g_probs)]
+
+
+def _softmax(nums):
+    if not nums:
+        return []
+    max_ = max(nums)
+    nums = [(exp(n-max_) if n is not None else None) for n in nums]
+    Z = sum(n for n in nums if n is not None)
+    return [(n/Z if n is not None else None) for n in nums]

spacy/pipeline/edit_tree_lemmatizer.py

@@ -5,8 +5,9 @@ from itertools import islice
 import numpy as np
 
 import srsly
-from thinc.api import Config, Model, SequenceCategoricalCrossentropy
+from thinc.api import Config, Model
 from thinc.types import ArrayXd, Floats2d, Ints1d
+from thinc.legacy import LegacySequenceCategoricalCrossentropy
 
 from ._edit_tree_internals.edit_trees import EditTrees
 from ._edit_tree_internals.schemas import validate_edit_tree
@@ -129,7 +130,9 @@ class EditTreeLemmatizer(TrainablePipe):
         self, examples: Iterable[Example], scores: List[Floats2d]
     ) -> Tuple[float, List[Floats2d]]:
         validate_examples(examples, "EditTreeLemmatizer.get_loss")
-        loss_func = SequenceCategoricalCrossentropy(normalize=False, missing_value=-1)
+        loss_func = LegacySequenceCategoricalCrossentropy(
+            normalize=False, missing_value=-1
+        )
 
         truths = []
         for eg in examples:
@@ -347,9 +350,9 @@ class EditTreeLemmatizer(TrainablePipe):
 
             tree = dict(tree)
             if "orig" in tree:
-                tree["orig"] = self.vocab.strings[tree["orig"]]
+                tree["orig"] = self.vocab.strings.add(tree["orig"])
             if "orig" in tree:
-                tree["subst"] = self.vocab.strings[tree["subst"]]
+                tree["subst"] = self.vocab.strings.add(tree["subst"])
 
             trees.append(tree)
 

spacy/pipeline/morphologizer.pyx

@@ -1,7 +1,8 @@
 # cython: infer_types=True, profile=True, binding=True
 from typing import Callable, Dict, Iterable, List, Optional, Union
 import srsly
-from thinc.api import SequenceCategoricalCrossentropy, Model, Config
+from thinc.api import Model, Config
+from thinc.legacy import LegacySequenceCategoricalCrossentropy
 from thinc.types import Floats2d, Ints1d
 from itertools import islice
 
@@ -290,7 +291,7 @@ class Morphologizer(Tagger):
         DOCS: https://spacy.io/api/morphologizer#get_loss
         """
         validate_examples(examples, "Morphologizer.get_loss")
-        loss_func = SequenceCategoricalCrossentropy(names=tuple(self.labels), normalize=False)
+        loss_func = LegacySequenceCategoricalCrossentropy(names=tuple(self.labels), normalize=False)
         truths = []
         for eg in examples:
             eg_truths = []

spacy/pipeline/multitask.pyx

@@ -1,221 +0,0 @@
-# cython: infer_types=True, profile=True, binding=True
-from typing import Optional
-import numpy
-from thinc.api import CosineDistance, to_categorical, Model, Config
-from thinc.api import set_dropout_rate
-
-from ..tokens.doc cimport Doc
-
-from .trainable_pipe import TrainablePipe
-from .tagger import Tagger
-from ..training import validate_examples
-from ..language import Language
-from ._parser_internals import nonproj
-from ..attrs import POS, ID
-from ..errors import Errors
-
-
-default_model_config = """
-[model]
-@architectures = "spacy.MultiTask.v1"
-maxout_pieces = 3
-token_vector_width = 96
-
-[model.tok2vec]
-@architectures = "spacy.HashEmbedCNN.v2"
-pretrained_vectors = null
-width = 96
-depth = 4
-embed_size = 2000
-window_size = 1
-maxout_pieces = 2
-subword_features = true
-"""
-DEFAULT_MT_MODEL = Config().from_str(default_model_config)["model"]
-
-
-@Language.factory(
-    "nn_labeller",
-    default_config={"labels": None, "target": "dep_tag_offset", "model": DEFAULT_MT_MODEL}
-)
-def make_nn_labeller(nlp: Language, name: str, model: Model, labels: Optional[dict], target: str):
-    return MultitaskObjective(nlp.vocab, model, name)
-
-
-class MultitaskObjective(Tagger):
-    """Experimental: Assist training of a parser or tagger, by training a
-    side-objective.
-    """
-
-    def __init__(self, vocab, model, name="nn_labeller", *, target):
-        self.vocab = vocab
-        self.model = model
-        self.name = name
-        if target == "dep":
-            self.make_label = self.make_dep
-        elif target == "tag":
-            self.make_label = self.make_tag
-        elif target == "ent":
-            self.make_label = self.make_ent
-        elif target == "dep_tag_offset":
-            self.make_label = self.make_dep_tag_offset
-        elif target == "ent_tag":
-            self.make_label = self.make_ent_tag
-        elif target == "sent_start":
-            self.make_label = self.make_sent_start
-        elif hasattr(target, "__call__"):
-            self.make_label = target
-        else:
-            raise ValueError(Errors.E016)
-        cfg = {"labels": {}, "target": target}
-        self.cfg = dict(cfg)
-
-    @property
-    def labels(self):
-        return self.cfg.setdefault("labels", {})
-
-    @labels.setter
-    def labels(self, value):
-        self.cfg["labels"] = value
-
-    def set_annotations(self, docs, dep_ids):
-        pass
-
-    def initialize(self, get_examples, nlp=None, labels=None):
-        if not hasattr(get_examples, "__call__"):
-            err = Errors.E930.format(name="MultitaskObjective", obj=type(get_examples))
-            raise ValueError(err)
-        if labels is not None:
-            self.labels = labels
-        else:
-            for example in get_examples():
-                for token in example.y:
-                    label = self.make_label(token)
-                    if label is not None and label not in self.labels:
-                        self.labels[label] = len(self.labels)
-        self.model.initialize()   # TODO: fix initialization by defining X and Y
-
-    def predict(self, docs):
-        tokvecs = self.model.get_ref("tok2vec")(docs)
-        scores = self.model.get_ref("softmax")(tokvecs)
-        return tokvecs, scores
-
-    def get_loss(self, examples, scores):
-        cdef int idx = 0
-        correct = numpy.zeros((scores.shape[0],), dtype="i")
-        guesses = scores.argmax(axis=1)
-        docs = [eg.predicted for eg in examples]
-        for i, eg in enumerate(examples):
-            # Handles alignment for tokenization differences
-            doc_annots = eg.get_aligned()  # TODO
-            for j in range(len(eg.predicted)):
-                tok_annots = {key: values[j] for key, values in tok_annots.items()}
-                label = self.make_label(j, tok_annots)
-                if label is None or label not in self.labels:
-                    correct[idx] = guesses[idx]
-                else:
-                    correct[idx] = self.labels[label]
-                idx += 1
-        correct = self.model.ops.xp.array(correct, dtype="i")
-        d_scores = scores - to_categorical(correct, n_classes=scores.shape[1])
-        loss = (d_scores**2).sum()
-        return float(loss), d_scores
-
-    @staticmethod
-    def make_dep(token):
-        return token.dep_
-
-    @staticmethod
-    def make_tag(token):
-        return token.tag_
-
-    @staticmethod
-    def make_ent(token):
-        if token.ent_iob_ == "O":
-            return "O"
-        else:
-            return token.ent_iob_ + "-" + token.ent_type_
-
-    @staticmethod
-    def make_dep_tag_offset(token):
-        dep = token.dep_
-        tag = token.tag_
-        offset = token.head.i - token.i
-        offset = min(offset, 2)
-        offset = max(offset, -2)
-        return f"{dep}-{tag}:{offset}"
-
-    @staticmethod
-    def make_ent_tag(token):
-        if token.ent_iob_ == "O":
-            ent = "O"
-        else:
-            ent = token.ent_iob_ + "-" + token.ent_type_
-        tag = token.tag_
-        return f"{tag}-{ent}"
-
-    @staticmethod
-    def make_sent_start(token):
-        """A multi-task objective for representing sentence boundaries,
-        using BILU scheme. (O is impossible)
-        """
-        if token.is_sent_start and token.is_sent_end:
-            return "U-SENT"
-        elif token.is_sent_start:
-            return "B-SENT"
-        else:
-            return "I-SENT"
-
-
-class ClozeMultitask(TrainablePipe):
-    def __init__(self, vocab, model, **cfg):
-        self.vocab = vocab
-        self.model = model
-        self.cfg = cfg
-        self.distance = CosineDistance(ignore_zeros=True, normalize=False)  # TODO: in config
-
-    def set_annotations(self, docs, dep_ids):
-        pass
-
-    def initialize(self, get_examples, nlp=None):
-        self.model.initialize()  # TODO: fix initialization by defining X and Y
-        X = self.model.ops.alloc((5, self.model.get_ref("tok2vec").get_dim("nO")))
-        self.model.output_layer.initialize(X)
-
-    def predict(self, docs):
-        tokvecs = self.model.get_ref("tok2vec")(docs)
-        vectors = self.model.get_ref("output_layer")(tokvecs)
-        return tokvecs, vectors
-
-    def get_loss(self, examples, vectors, prediction):
-        validate_examples(examples, "ClozeMultitask.get_loss")
-        # The simplest way to implement this would be to vstack the
-        # token.vector values, but that's a bit inefficient, especially on GPU.
-        # Instead we fetch the index into the vectors table for each of our tokens,
-        # and look them up all at once. This prevents data copying.
-        ids = self.model.ops.flatten([eg.predicted.to_array(ID).ravel() for eg in examples])
-        target = vectors[ids]
-        gradient = self.distance.get_grad(prediction, target)
-        loss = self.distance.get_loss(prediction, target)
-        return float(loss), gradient
-
-    def update(self, examples, *, drop=0., sgd=None, losses=None):
-        pass
-
-    def rehearse(self, examples, drop=0., sgd=None, losses=None):
-        if losses is not None and self.name not in losses:
-            losses[self.name] = 0.
-        set_dropout_rate(self.model, drop)
-        validate_examples(examples, "ClozeMultitask.rehearse")
-        docs = [eg.predicted for eg in examples]
-        predictions, bp_predictions = self.model.begin_update()
-        loss, d_predictions = self.get_loss(examples, self.vocab.vectors.data, predictions)
-        bp_predictions(d_predictions)
-        if sgd is not None:
-            self.finish_update(sgd)
-        if losses is not None:
-            losses[self.name] += loss
-        return losses
-
-    def add_label(self, label):
-        raise NotImplementedError

spacy/pipeline/pipe.pyx

@@ -19,13 +19,6 @@ cdef class Pipe:
     DOCS: https://spacy.io/api/pipe
     """
 
-    @classmethod
-    def __init_subclass__(cls, **kwargs):
-        """Raise a warning if an inheriting class implements 'begin_training'
-         (from v2) instead of the new 'initialize' method (from v3)"""
-        if hasattr(cls, "begin_training"):
-            warnings.warn(Warnings.W088.format(name=cls.__name__))
-
     def __call__(self, Doc doc) -> Doc:
         """Apply the pipe to one document. The document is modified in place,
         and returned. This usually happens under the hood when the nlp object

spacy/pipeline/senter.pyx

@@ -3,7 +3,9 @@ from typing import Dict, Iterable, Optional, Callable, List, Union
 from itertools import islice
 
 import srsly
-from thinc.api import Model, SequenceCategoricalCrossentropy, Config
+from thinc.api import Model, Config
+from thinc.legacy import LegacySequenceCategoricalCrossentropy
+
 from thinc.types import Floats2d, Ints1d
 
 from ..tokens.doc cimport Doc
@@ -161,7 +163,7 @@ class SentenceRecognizer(Tagger):
         """
         validate_examples(examples, "SentenceRecognizer.get_loss")
         labels = self.labels
-        loss_func = SequenceCategoricalCrossentropy(names=labels, normalize=False)
+        loss_func = LegacySequenceCategoricalCrossentropy(names=labels, normalize=False)
         truths = []
         for eg in examples:
             eg_truth = []

spacy/pipeline/spancat.py

@@ -2,7 +2,7 @@ from typing import List, Dict, Callable, Tuple, Optional, Iterable, Any, cast
 from typing import Union
 from thinc.api import Config, Model, get_current_ops, set_dropout_rate, Ops
 from thinc.api import Optimizer
-from thinc.types import Ragged, Ints2d, Floats2d, Ints1d
+from thinc.types import Ragged, Ints2d, Floats2d
 
 import numpy
 
@@ -282,7 +282,10 @@ class SpanCategorizer(TrainablePipe):
         DOCS: https://spacy.io/api/spancategorizer#predict
         """
         indices = self.suggester(docs, ops=self.model.ops)
-        scores = self.model.predict((docs, indices))  # type: ignore
+        if indices.lengths.sum() == 0:
+            scores = self.model.ops.alloc2f(0, 0)
+        else:
+            scores = self.model.predict((docs, indices))  # type: ignore
         return {"indices": indices, "scores": scores}
 
     def set_candidates(

spacy/pipeline/tagger.pyx

@@ -2,7 +2,8 @@
 from typing import Callable, Dict, Iterable, List, Optional, Union
 import numpy
 import srsly
-from thinc.api import Model, set_dropout_rate, SequenceCategoricalCrossentropy, Config
+from thinc.api import Model, set_dropout_rate, Config
+from thinc.legacy import LegacySequenceCategoricalCrossentropy
 from thinc.types import Floats2d, Ints1d
 import warnings
 from itertools import islice
@@ -244,7 +245,7 @@ class Tagger(TrainablePipe):
 
         DOCS: https://spacy.io/api/tagger#rehearse
         """
-        loss_func = SequenceCategoricalCrossentropy()
+        loss_func = LegacySequenceCategoricalCrossentropy()
         if losses is None:
             losses = {}
         losses.setdefault(self.name, 0.0)
@@ -275,7 +276,7 @@ class Tagger(TrainablePipe):
         DOCS: https://spacy.io/api/tagger#get_loss
         """
         validate_examples(examples, "Tagger.get_loss")
-        loss_func = SequenceCategoricalCrossentropy(names=self.labels, normalize=False, neg_prefix=self.cfg["neg_prefix"])
+        loss_func = LegacySequenceCategoricalCrossentropy(names=self.labels, normalize=False, neg_prefix=self.cfg["neg_prefix"])
         # Convert empty tag "" to missing value None so that both misaligned
         # tokens and tokens with missing annotation have the default missing
         # value None.

spacy/pipeline/textcat.py

@@ -91,7 +91,6 @@ subword_features = true
         "cats_macro_f": None,
         "cats_macro_auc": None,
         "cats_f_per_type": None,
-        "cats_macro_auc_per_type": None,
     },
 )
 def make_textcat(
@@ -169,7 +168,11 @@ class TextCategorizer(TrainablePipe):
         self.model = model
         self.name = name
         self._rehearsal_model = None
-        cfg: Dict[str, Any] = {"labels": [], "threshold": threshold, "positive_label": None}
+        cfg: Dict[str, Any] = {
+            "labels": [],
+            "threshold": threshold,
+            "positive_label": None,
+        }
         self.cfg = dict(cfg)
         self.scorer = scorer
         self.save_activations = save_activations
@@ -416,5 +419,9 @@ class TextCategorizer(TrainablePipe):
     def _validate_categories(self, examples: Iterable[Example]):
         """Check whether the provided examples all have single-label cats annotations."""
         for ex in examples:
-            if list(ex.reference.cats.values()).count(1.0) > 1:
+            vals = list(ex.reference.cats.values())
+            if vals.count(1.0) > 1:
                 raise ValueError(Errors.E895.format(value=ex.reference.cats))
+            for val in vals:
+                if not (val == 1.0 or val == 0.0):
+                    raise ValueError(Errors.E851.format(val=val))

spacy/pipeline/textcat_multilabel.py

@@ -88,7 +88,6 @@ subword_features = true
         "cats_macro_f": None,
         "cats_macro_auc": None,
         "cats_f_per_type": None,
-        "cats_macro_auc_per_type": None,
     },
 )
 def make_multilabel_textcat(
@@ -156,11 +155,8 @@ class MultiLabel_TextCategorizer(TextCategorizer):
         name (str): The component instance name, used to add entries to the
             losses during training.
         threshold (float): Cutoff to consider a prediction "positive".
-<<<<<<< HEAD
-        save_activations (bool): save model activations in Doc when annotating.
-=======
         scorer (Optional[Callable]): The scoring method.
->>>>>>> upstream/master
+        save_activations (bool): save model activations in Doc when annotating.
 
         DOCS: https://spacy.io/api/textcategorizer#init
         """
@@ -205,6 +201,8 @@ class MultiLabel_TextCategorizer(TextCategorizer):
             for label in labels:
                 self.add_label(label)
         subbatch = list(islice(get_examples(), 10))
+        self._validate_categories(subbatch)
+
         doc_sample = [eg.reference for eg in subbatch]
         label_sample, _ = self._examples_to_truth(subbatch)
         self._require_labels()
@@ -215,4 +213,8 @@ class MultiLabel_TextCategorizer(TextCategorizer):
     def _validate_categories(self, examples: Iterable[Example]):
         """This component allows any type of single- or multi-label annotations.
         This method overwrites the more strict one from 'textcat'."""
-        pass
+        # check that annotation values are valid
+        for ex in examples:
+            for val in ex.reference.cats.values():
+                if not (val == 1.0 or val == 0.0):
+                    raise ValueError(Errors.E851.format(val=val))

spacy/pipeline/transition_parser.pyx

@@ -10,12 +10,12 @@ import random
 
 import srsly
 from thinc.api import get_ops, set_dropout_rate, CupyOps, NumpyOps
-from thinc.extra.search cimport Beam
 import numpy.random
 import numpy
 import warnings
 
 from ._parser_internals.stateclass cimport StateClass
+from ._parser_internals.search cimport Beam
 from ..ml.parser_model cimport alloc_activations, free_activations
 from ..ml.parser_model cimport predict_states, arg_max_if_valid
 from ..ml.parser_model cimport WeightsC, ActivationsC, SizesC, cpu_log_loss

spacy/schemas.py

@@ -329,6 +329,7 @@ class ConfigSchemaTraining(BaseModel):
     frozen_components: List[str] = Field(..., title="Pipeline components that shouldn't be updated during training")
     annotating_components: List[str] = Field(..., title="Pipeline components that should set annotations during training")
     before_to_disk: Optional[Callable[["Language"], "Language"]] = Field(..., title="Optional callback to modify nlp object after training, before it's saved to disk")
+    before_update: Optional[Callable[["Language", Dict[str, Any]], None]] = Field(..., title="Optional callback that is invoked at the start of each training step")
     # fmt: on
 
     class Config:

spacy/tests/README.md

@@ -40,7 +40,7 @@ py.test spacy/tests/tokenizer/test_exceptions.py::test_tokenizer_handles_emoji #
 
 To keep the behavior of the tests consistent and predictable, we try to follow a few basic conventions:
 
-- **Test names** should follow a pattern of `test_[module]_[tested behaviour]`. For example: `test_tokenizer_keeps_email` or `test_spans_override_sentiment`.
+- **Test names** should follow a pattern of `test_[module]_[tested behaviour]`. For example: `test_tokenizer_keeps_email`.
 - If you're testing for a bug reported in a specific issue, always create a **regression test**. Regression tests should be named `test_issue[ISSUE NUMBER]` and live in the [`regression`](regression) directory.
 - Only use `@pytest.mark.xfail` for tests that **should pass, but currently fail**. To test for desired negative behavior, use `assert not` in your test.
 - Very **extensive tests** that take a long time to run should be marked with `@pytest.mark.slow`. If your slow test is testing important behavior, consider adding an additional simpler version.

spacy/tests/conftest.py

@@ -1,6 +1,10 @@
 import pytest
 from spacy.util import get_lang_class
+import functools
 from hypothesis import settings
+import inspect
+import importlib
+import sys
 
 # Functionally disable deadline settings for tests
 # to prevent spurious test failures in CI builds.
@@ -47,6 +51,33 @@ def pytest_runtest_setup(item):
             pytest.skip("not referencing any issues")
 
 
+# Decorator for Cython-built tests
+# https://shwina.github.io/cython-testing/
+def cytest(func):
+    """
+    Wraps `func` in a plain Python function.
+    """
+
+    @functools.wraps(func)
+    def wrapped(*args, **kwargs):
+        bound = inspect.signature(func).bind(*args, **kwargs)
+        return func(*bound.args, **bound.kwargs)
+
+    return wrapped
+
+
+def register_cython_tests(cython_mod_name: str, test_mod_name: str):
+    """
+    Registers all callables with name `test_*` in Cython module `cython_mod_name`
+    as attributes in module `test_mod_name`, making them discoverable by pytest.
+    """
+    cython_mod = importlib.import_module(cython_mod_name)
+    for name in dir(cython_mod):
+        item = getattr(cython_mod, name)
+        if callable(item) and name.startswith("test_"):
+            setattr(sys.modules[test_mod_name], name, item)
+
+
 # Fixtures for language tokenizers (languages sorted alphabetically)
 
 
@@ -351,17 +382,17 @@ def ru_tokenizer():
     return get_lang_class("ru")().tokenizer
 
 
-@pytest.fixture
+@pytest.fixture(scope="session")
 def ru_lemmatizer():
     pytest.importorskip("pymorphy3")
     return get_lang_class("ru")().add_pipe("lemmatizer")
 
 
-@pytest.fixture
+@pytest.fixture(scope="session")
 def ru_lookup_lemmatizer():
-    pytest.importorskip("pymorphy2")
+    pytest.importorskip("pymorphy3")
     return get_lang_class("ru")().add_pipe(
-        "lemmatizer", config={"mode": "pymorphy2_lookup"}
+        "lemmatizer", config={"mode": "pymorphy3_lookup"}
     )
 
 
@@ -437,19 +468,19 @@ def uk_tokenizer():
     return get_lang_class("uk")().tokenizer
 
 
-@pytest.fixture
+@pytest.fixture(scope="session")
 def uk_lemmatizer():
     pytest.importorskip("pymorphy3")
     pytest.importorskip("pymorphy3_dicts_uk")
     return get_lang_class("uk")().add_pipe("lemmatizer")
 
 
-@pytest.fixture
+@pytest.fixture(scope="session")
 def uk_lookup_lemmatizer():
-    pytest.importorskip("pymorphy2")
+    pytest.importorskip("pymorphy3")
-    pytest.importorskip("pymorphy2_dicts_uk")
+    pytest.importorskip("pymorphy3_dicts_uk")
     return get_lang_class("uk")().add_pipe(
-        "lemmatizer", config={"mode": "pymorphy2_lookup"}
+        "lemmatizer", config={"mode": "pymorphy3_lookup"}
     )
 
 

spacy/tests/doc/test_array.py

@@ -123,14 +123,14 @@ def test_doc_from_array_heads_in_bounds(en_vocab):
 
     # head before start
     arr = doc.to_array(["HEAD"])
-    arr[0] = -1
+    arr[0] = numpy.int32(-1).astype(numpy.uint64)
     doc_from_array = Doc(en_vocab, words=words)
     with pytest.raises(ValueError):
         doc_from_array.from_array(["HEAD"], arr)
 
     # head after end
     arr = doc.to_array(["HEAD"])
-    arr[0] = 5
+    arr[0] = numpy.int32(5).astype(numpy.uint64)
     doc_from_array = Doc(en_vocab, words=words)
     with pytest.raises(ValueError):
         doc_from_array.from_array(["HEAD"], arr)

spacy/tests/doc/test_doc_api.py

@@ -380,9 +380,7 @@ def test_doc_api_serialize(en_tokenizer, text):
     assert [t.text for t in tokens] == [t.text for t in new_tokens]
     assert [t.orth for t in tokens] == [t.orth for t in new_tokens]
 
-    new_tokens = Doc(tokens.vocab).from_bytes(
+    new_tokens = Doc(tokens.vocab).from_bytes(tokens.to_bytes())
-        tokens.to_bytes(exclude=["sentiment"]), exclude=["sentiment"]
-    )
     assert tokens.text == new_tokens.text
     assert [t.text for t in tokens] == [t.text for t in new_tokens]
     assert [t.orth for t in tokens] == [t.orth for t in new_tokens]
@@ -990,3 +988,12 @@ def test_doc_spans_setdefault(en_tokenizer):
     assert len(doc.spans["key2"]) == 1
     doc.spans.setdefault("key3", default=SpanGroup(doc, spans=[doc[0:1], doc[1:2]]))
     assert len(doc.spans["key3"]) == 2
+
+
+def test_doc_sentiment_from_bytes_v3_to_v4():
+    """Test if a doc with sentiment attribute created in v3.x works with '.from_bytes' in v4.x without throwing errors. The sentiment attribute was removed in v4"""
+    doc_bytes = b"\x89\xa4text\xa5happy\xaaarray_head\x9fGQACKOLMN\xcd\x01\xc4\xcd\x01\xc6I\xcd\x01\xc5JP\xaaarray_body\x85\xc4\x02nd\xc3\xc4\x04type\xa3<u8\xc4\x04kind\xc4\x00\xc4\x05shape\x92\x01\x0f\xc4\x04data\xc4x\x05\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\xa4\x9a\xd3\x17\xca\xf0b\x03\xa4\x9a\xd3\x17\xca\xf0b\x03\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01\x00\x00\x00\x00\x00\x00\x00\xa9sentiment\xcb?\xf0\x00\x00\x00\x00\x00\x00\xa6tensor\x85\xc4\x02nd\xc3\xc4\x04type\xa3<f4\xc4\x04kind\xc4\x00\xc4\x05shape\x91\x00\xc4\x04data\xc4\x00\xa4cats\x80\xa5spans\xc4\x01\x90\xa7strings\x92\xa0\xa5happy\xb2has_unknown_spaces\xc2"
+    doc = Doc(Vocab()).from_bytes(doc_bytes)
+    assert doc.text == "happy"
+    with pytest.raises(AttributeError):
+        doc.sentiment == 1.0

spacy/tests/doc/test_json_doc_conversion.py

@@ -370,3 +370,12 @@ def test_json_to_doc_validation_error(doc):
     doc_json.pop("tokens")
     with pytest.raises(ValueError):
         Doc(doc.vocab).from_json(doc_json, validate=True)
+
+
+def test_to_json_underscore_doc_getters(doc):
+    def get_text_length(doc):
+        return len(doc.text)
+
+    Doc.set_extension("text_length", getter=get_text_length)
+    doc_json = doc.to_json(underscore=["text_length"])
+    assert doc_json["_"]["text_length"] == get_text_length(doc)

spacy/tests/doc/test_span.py

@@ -305,31 +305,6 @@ def test_span_similarity_match():
         assert span1[:1].similarity(doc.vocab["a"]) == 1.0
 
 
-def test_spans_default_sentiment(en_tokenizer):
-    """Test span.sentiment property's default averaging behaviour"""
-    text = "good stuff bad stuff"
-    tokens = en_tokenizer(text)
-    tokens.vocab[tokens[0].text].sentiment = 3.0
-    tokens.vocab[tokens[2].text].sentiment = -2.0
-    doc = Doc(tokens.vocab, words=[t.text for t in tokens])
-    assert doc[:2].sentiment == 3.0 / 2
-    assert doc[-2:].sentiment == -2.0 / 2
-    assert doc[:-1].sentiment == (3.0 + -2) / 3.0
-
-
-def test_spans_override_sentiment(en_tokenizer):
-    """Test span.sentiment property's default averaging behaviour"""
-    text = "good stuff bad stuff"
-    tokens = en_tokenizer(text)
-    tokens.vocab[tokens[0].text].sentiment = 3.0
-    tokens.vocab[tokens[2].text].sentiment = -2.0
-    doc = Doc(tokens.vocab, words=[t.text for t in tokens])
-    doc.user_span_hooks["sentiment"] = lambda span: 10.0
-    assert doc[:2].sentiment == 10.0
-    assert doc[-2:].sentiment == 10.0
-    assert doc[:-1].sentiment == 10.0
-
-
 def test_spans_are_hashable(en_tokenizer):
     """Test spans can be hashed."""
     text = "good stuff bad stuff"

spacy/tests/doc/test_span_group.py

@@ -1,7 +1,10 @@
+from typing import List
+
 import pytest
 from random import Random
 from spacy.matcher import Matcher
-from spacy.tokens import Span, SpanGroup
+from spacy.tokens import Span, SpanGroup, Doc
+from spacy.util import filter_spans
 
 
 @pytest.fixture
@@ -242,3 +245,13 @@ def test_span_group_extend(doc):
 def test_span_group_dealloc(span_group):
     with pytest.raises(AttributeError):
         print(span_group.doc)
+
+
+@pytest.mark.issue(11975)
+def test_span_group_typing(doc: Doc):
+    """Tests whether typing of `SpanGroup` as `Iterable[Span]`-like object is accepted by mypy."""
+    span_group: SpanGroup = doc.spans["SPANS"]
+    spans: List[Span] = list(span_group)
+    for i, span in enumerate(span_group):
+        assert span == span_group[i] == spans[i]
+    filter_spans(span_group)

spacy/tests/doc/test_underscore.py

@@ -3,6 +3,10 @@ from mock import Mock
 from spacy.tokens import Doc, Span, Token
 from spacy.tokens.underscore import Underscore
 
+# Helper functions
+def _get_tuple(s: Span):
+    return "._.", "span_extension", s.start_char, s.end_char, s.label, s.kb_id, s.id
+
 
 @pytest.fixture(scope="function", autouse=True)
 def clean_underscore():
@@ -171,3 +175,118 @@ def test_underscore_docstring(en_vocab):
     doc = Doc(en_vocab, words=["hello", "world"])
     assert test_method.__doc__ == "I am a docstring"
     assert doc._.test_docstrings.__doc__.rsplit(". ")[-1] == "I am a docstring"
+
+
+def test_underscore_for_unique_span(en_tokenizer):
+    """Test that spans with the same boundaries but with different labels are uniquely identified (see #9706)."""
+    Doc.set_extension(name="doc_extension", default=None)
+    Span.set_extension(name="span_extension", default=None)
+    Token.set_extension(name="token_extension", default=None)
+
+    # Initialize doc
+    text = "Hello, world!"
+    doc = en_tokenizer(text)
+    span_1 = Span(doc, 0, 2, "SPAN_1")
+    span_2 = Span(doc, 0, 2, "SPAN_2")
+
+    # Set custom extensions
+    doc._.doc_extension = "doc extension"
+    doc[0]._.token_extension = "token extension"
+    span_1._.span_extension = "span_1 extension"
+    span_2._.span_extension = "span_2 extension"
+
+    # Assert extensions
+    assert doc.user_data[_get_tuple(span_1)] == "span_1 extension"
+    assert doc.user_data[_get_tuple(span_2)] == "span_2 extension"
+
+    # Change label of span and assert extensions
+    span_1.label_ = "NEW_LABEL"
+    assert doc.user_data[_get_tuple(span_1)] == "span_1 extension"
+    assert doc.user_data[_get_tuple(span_2)] == "span_2 extension"
+
+    # Change KB_ID and assert extensions
+    span_1.kb_id_ = "KB_ID"
+    assert doc.user_data[_get_tuple(span_1)] == "span_1 extension"
+    assert doc.user_data[_get_tuple(span_2)] == "span_2 extension"
+
+    # Change extensions and assert
+    span_2._.span_extension = "updated span_2 extension"
+    assert doc.user_data[_get_tuple(span_1)] == "span_1 extension"
+    assert doc.user_data[_get_tuple(span_2)] == "updated span_2 extension"
+
+    # Change span ID and assert extensions
+    span_2.id = 2
+    assert doc.user_data[_get_tuple(span_1)] == "span_1 extension"
+    assert doc.user_data[_get_tuple(span_2)] == "updated span_2 extension"
+
+    # Assert extensions with original key
+    assert doc.user_data[("._.", "doc_extension", None, None)] == "doc extension"
+    assert doc.user_data[("._.", "token_extension", 0, None)] == "token extension"
+
+
+def test_underscore_for_unique_span_from_docs(en_tokenizer):
+    """Test that spans in the user_data keep the same data structure when using Doc.from_docs"""
+    Span.set_extension(name="span_extension", default=None)
+    Token.set_extension(name="token_extension", default=None)
+
+    # Initialize doc
+    text_1 = "Hello, world!"
+    doc_1 = en_tokenizer(text_1)
+    span_1a = Span(doc_1, 0, 2, "SPAN_1a")
+    span_1b = Span(doc_1, 0, 2, "SPAN_1b")
+
+    text_2 = "This is a test."
+    doc_2 = en_tokenizer(text_2)
+    span_2a = Span(doc_2, 0, 3, "SPAN_2a")
+
+    # Set custom extensions
+    doc_1[0]._.token_extension = "token_1"
+    doc_2[1]._.token_extension = "token_2"
+    span_1a._.span_extension = "span_1a extension"
+    span_1b._.span_extension = "span_1b extension"
+    span_2a._.span_extension = "span_2a extension"
+
+    doc = Doc.from_docs([doc_1, doc_2])
+    # Assert extensions
+    assert doc_1.user_data[_get_tuple(span_1a)] == "span_1a extension"
+    assert doc_1.user_data[_get_tuple(span_1b)] == "span_1b extension"
+    assert doc_2.user_data[_get_tuple(span_2a)] == "span_2a extension"
+
+    # Check extensions on merged doc
+    assert doc.user_data[_get_tuple(span_1a)] == "span_1a extension"
+    assert doc.user_data[_get_tuple(span_1b)] == "span_1b extension"
+    assert (
+        doc.user_data[
+            (
+                "._.",
+                "span_extension",
+                span_2a.start_char + len(doc_1.text) + 1,
+                span_2a.end_char + len(doc_1.text) + 1,
+                span_2a.label,
+                span_2a.kb_id,
+                span_2a.id,
+            )
+        ]
+        == "span_2a extension"
+    )
+
+
+def test_underscore_for_unique_span_as_span(en_tokenizer):
+    """Test that spans in the user_data keep the same data structure when using Span.as_doc"""
+    Span.set_extension(name="span_extension", default=None)
+
+    # Initialize doc
+    text = "Hello, world!"
+    doc = en_tokenizer(text)
+    span_1 = Span(doc, 0, 2, "SPAN_1")
+    span_2 = Span(doc, 0, 2, "SPAN_2")
+
+    # Set custom extensions
+    span_1._.span_extension = "span_1 extension"
+    span_2._.span_extension = "span_2 extension"
+
+    span_doc = span_1.as_doc(copy_user_data=True)
+
+    # Assert extensions
+    assert span_doc.user_data[_get_tuple(span_1)] == "span_1 extension"
+    assert span_doc.user_data[_get_tuple(span_2)] == "span_2 extension"

spacy/tests/lang/ru/test_lemmatizer.py

@@ -81,6 +81,7 @@ def test_ru_lemmatizer_punct(ru_lemmatizer):
 
 
 def test_ru_doc_lookup_lemmatization(ru_lookup_lemmatizer):
+    assert ru_lookup_lemmatizer.mode == "pymorphy3_lookup"
     words = ["мама", "мыла", "раму"]
     pos = ["NOUN", "VERB", "NOUN"]
     morphs = [
@@ -92,3 +93,17 @@ def test_ru_doc_lookup_lemmatization(ru_lookup_lemmatizer):
     doc = ru_lookup_lemmatizer(doc)
     lemmas = [token.lemma_ for token in doc]
     assert lemmas == ["мама", "мыла", "раму"]
+
+
+@pytest.mark.parametrize(
+    "word,lemma",
+    (
+        ("бременем", "бремя"),
+        ("будешь", "быть"),
+        ("какая-то", "какой-то"),
+    ),
+)
+def test_ru_lookup_lemmatizer(ru_lookup_lemmatizer, word, lemma):
+    assert ru_lookup_lemmatizer.mode == "pymorphy3_lookup"
+    doc = Doc(ru_lookup_lemmatizer.vocab, words=[word])
+    assert ru_lookup_lemmatizer(doc)[0].lemma_ == lemma

spacy/tests/lang/uk/test_lemmatizer.py

@@ -8,12 +8,20 @@ pytestmark = pytest.mark.filterwarnings("ignore::DeprecationWarning")
 def test_uk_lemmatizer(uk_lemmatizer):
     """Check that the default uk lemmatizer runs."""
     doc = Doc(uk_lemmatizer.vocab, words=["a", "b", "c"])
+    assert uk_lemmatizer.mode == "pymorphy3"
     uk_lemmatizer(doc)
     assert [token.lemma for token in doc]
 
 
-def test_uk_lookup_lemmatizer(uk_lookup_lemmatizer):
+@pytest.mark.parametrize(
-    """Check that the lookup uk lemmatizer runs."""
+    "word,lemma",
-    doc = Doc(uk_lookup_lemmatizer.vocab, words=["a", "b", "c"])
+    (
-    uk_lookup_lemmatizer(doc)
+        ("якийсь", "якийсь"),
-    assert [token.lemma for token in doc]
+        ("розповідають", "розповідати"),
+        ("розповіси", "розповісти"),
+    ),
+)
+def test_uk_lookup_lemmatizer(uk_lookup_lemmatizer, word, lemma):
+    assert uk_lookup_lemmatizer.mode == "pymorphy3_lookup"
+    doc = Doc(uk_lookup_lemmatizer.vocab, words=[word])
+    assert uk_lookup_lemmatizer(doc)[0].lemma_ == lemma

spacy/tests/matcher/test_matcher_api.py

@@ -50,8 +50,6 @@ def test_matcher_from_usage_docs(en_vocab):
 
     def label_sentiment(matcher, doc, i, matches):
         match_id, start, end = matches[i]
-        if doc.vocab.strings[match_id] == "HAPPY":
-            doc.sentiment += 0.1
         span = doc[start:end]
         with doc.retokenize() as retokenizer:
             retokenizer.merge(span)
@@ -61,7 +59,6 @@ def test_matcher_from_usage_docs(en_vocab):
     matcher = Matcher(en_vocab)
     matcher.add("HAPPY", pos_patterns, on_match=label_sentiment)
     matcher(doc)
-    assert doc.sentiment != 0
     assert doc[1].norm_ == "happy emoji"
 
 

spacy/tests/parser/_search.pyx

@@ -0,0 +1,119 @@
+# cython: infer_types=True, binding=True
+from spacy.pipeline._parser_internals.search cimport Beam, MaxViolation
+from spacy.typedefs cimport class_t, weight_t
+from cymem.cymem cimport Pool
+
+from ..conftest import cytest
+import pytest
+
+cdef struct TestState:
+    int length
+    int x
+    Py_UNICODE* string
+
+
+cdef int transition(void* dest, void* src, class_t clas, void* extra_args) except -1:
+    dest_state = <TestState*>dest
+    src_state = <TestState*>src
+    dest_state.length = src_state.length
+    dest_state.x = src_state.x
+    dest_state.x += clas
+    if extra_args != NULL:
+        dest_state.string = <Py_UNICODE*>extra_args
+    else:
+        dest_state.string = src_state.string
+
+
+cdef void* initialize(Pool mem, int n, void* extra_args) except NULL:
+    state = <TestState*>mem.alloc(1, sizeof(TestState))
+    state.length = n
+    state.x = 1
+    if extra_args == NULL:
+        state.string = u'default'
+    else:
+        state.string = <Py_UNICODE*>extra_args
+    return state
+
+
+cdef int destroy(Pool mem, void* state, void* extra_args) except -1:
+    state = <TestState*>state
+    mem.free(state)
+
+@cytest
+@pytest.mark.parametrize("nr_class,beam_width",
+    [
+        (2, 3),
+        (3, 6),
+        (4, 20),
+    ]
+)
+def test_init(nr_class, beam_width):
+    b = Beam(nr_class, beam_width)
+    assert b.size == 1
+    assert b.width == beam_width
+    assert b.nr_class == nr_class
+
+@cytest
+def test_init_violn():
+    MaxViolation()
+
+@cytest
+@pytest.mark.parametrize("nr_class,beam_width,length",
+    [
+        (2, 3, 3),
+        (3, 6, 15),
+        (4, 20, 32),
+    ]
+)
+def test_initialize(nr_class, beam_width, length):
+    b = Beam(nr_class, beam_width)
+    b.initialize(initialize, destroy, length, NULL)
+    for i in range(b.width):
+        s = <TestState*>b.at(i)
+        assert s.length == length, s.length
+        assert s.string == 'default'
+
+
+@cytest
+@pytest.mark.parametrize("nr_class,beam_width,length,extra",
+    [
+        (2, 3, 4, None),
+        (3, 6, 15, u"test beam 1"),
+    ]
+)
+def test_initialize_extra(nr_class, beam_width, length, extra):
+    b = Beam(nr_class, beam_width)
+    if extra is None:
+        b.initialize(initialize, destroy, length, NULL)
+    else:
+        b.initialize(initialize, destroy, length, <void*><Py_UNICODE*>extra)
+    for i in range(b.width):
+        s = <TestState*>b.at(i)
+        assert s.length == length
+
+
+@cytest
+@pytest.mark.parametrize("nr_class,beam_width,length",
+    [
+        (3, 6, 15),
+        (4, 20, 32),
+    ]
+)
+def test_transition(nr_class, beam_width, length):
+    b = Beam(nr_class, beam_width)
+    b.initialize(initialize, destroy, length, NULL)
+    b.set_cell(0, 2, 30, True, 0)
+    b.set_cell(0, 1, 42, False, 0)
+    b.advance(transition, NULL, NULL)
+    assert b.size == 1, b.size
+    assert b.score == 30, b.score
+    s = <TestState*>b.at(0)
+    assert s.x == 3
+    assert b._states[0].score == 30, b._states[0].score
+    b.set_cell(0, 1, 10, True, 0)
+    b.set_cell(0, 2, 20, True, 0)
+    b.advance(transition, NULL, NULL)
+    assert b._states[0].score == 50, b._states[0].score
+    assert b._states[1].score == 40
+    s = <TestState*>b.at(0)
+    assert s.x == 5

spacy/tests/parser/test_search.py

@@ -0,0 +1,3 @@
+from ..conftest import register_cython_tests
+
+register_cython_tests("spacy.tests.parser._search", __name__)

spacy/tests/pipeline/test_edit_tree_lemmatizer.py

@@ -62,10 +62,45 @@ def test_initialize_from_labels():
     nlp2 = Language()
     lemmatizer2 = nlp2.add_pipe("trainable_lemmatizer")
     lemmatizer2.initialize(
-        get_examples=lambda: train_examples,
+        # We want to check that the strings in replacement nodes are
+        # added to the string store. Avoid that they get added through
+        # the examples.
+        get_examples=lambda: train_examples[:1],
         labels=lemmatizer.label_data,
     )
     assert lemmatizer2.tree2label == {1: 0, 3: 1, 4: 2, 6: 3}
+    assert lemmatizer2.label_data == {
+        "trees": [
+            {"orig": "S", "subst": "s"},
+            {
+                "prefix_len": 1,
+                "suffix_len": 0,
+                "prefix_tree": 0,
+                "suffix_tree": 4294967295,
+            },
+            {"orig": "s", "subst": ""},
+            {
+                "prefix_len": 0,
+                "suffix_len": 1,
+                "prefix_tree": 4294967295,
+                "suffix_tree": 2,
+            },
+            {
+                "prefix_len": 0,
+                "suffix_len": 0,
+                "prefix_tree": 4294967295,
+                "suffix_tree": 4294967295,
+            },
+            {"orig": "E", "subst": "e"},
+            {
+                "prefix_len": 1,
+                "suffix_len": 0,
+                "prefix_tree": 5,
+                "suffix_tree": 4294967295,
+            },
+        ],
+        "labels": (1, 3, 4, 6),
+    }
 
 
 def test_no_data():

spacy/tests/pipeline/test_pipe_methods.py

@@ -529,17 +529,6 @@ def test_pipe_label_data_no_labels(pipe):
         assert "labels" not in get_arg_names(initialize)
 
 
-def test_warning_pipe_begin_training():
-    with pytest.warns(UserWarning, match="begin_training"):
-
-        class IncompatPipe(TrainablePipe):
-            def __init__(self):
-                ...
-
-            def begin_training(*args, **kwargs):
-                ...
-
-
 def test_pipe_methods_initialize():
     """Test that the [initialize] config reflects the components correctly."""
     nlp = Language()
@@ -615,20 +604,18 @@ def test_enable_disable_conflict_with_config():
 
     with make_tempdir() as tmp_dir:
         nlp.to_disk(tmp_dir)
-        # Expected to fail, as config and arguments conflict.
+        # Expected to succeed, as config and arguments do not conflict.
-        with pytest.raises(ValueError):
+        assert spacy.load(
-            spacy.load(
+            tmp_dir, enable=["tagger"], config={"nlp": {"disabled": ["senter"]}}
-                tmp_dir, enable=["tagger"], config={"nlp": {"disabled": ["senter"]}}
+        ).disabled == ["senter", "sentencizer"]
-            )
         # Expected to succeed without warning due to the lack of a conflicting config option.
         spacy.load(tmp_dir, enable=["tagger"])
-        # Expected to succeed with a warning, as disable=[] should override the config setting.
+        # Expected to fail due to conflict between enable and disabled.
-        with pytest.warns(UserWarning):
+        with pytest.raises(ValueError):
             spacy.load(
                 tmp_dir,
-                enable=["tagger"],
+                enable=["senter"],
-                disable=[],
+                config={"nlp": {"disabled": ["senter", "tagger"]}},
-                config={"nlp": {"disabled": ["senter"]}},
             )
 
 

spacy/tests/pipeline/test_spancat.py

@@ -372,24 +372,39 @@ def test_overfitting_IO_overlapping():
 
 
 def test_zero_suggestions():
-    # Test with a suggester that returns 0 suggestions
+    # Test with a suggester that can return 0 suggestions
 
-    @registry.misc("test_zero_suggester")
+    @registry.misc("test_mixed_zero_suggester")
-    def make_zero_suggester():
+    def make_mixed_zero_suggester():
-        def zero_suggester(docs, *, ops=None):
+        def mixed_zero_suggester(docs, *, ops=None):
             if ops is None:
                 ops = get_current_ops()
-            return Ragged(
+            spans = []
-                ops.xp.zeros((0, 0), dtype="i"), ops.xp.zeros((len(docs),), dtype="i")
+            lengths = []
-            )
+            for doc in docs:
+                if len(doc) > 0 and len(doc) % 2 == 0:
+                    spans.append((0, 1))
+                    lengths.append(1)
+                else:
+                    lengths.append(0)
+            spans = ops.asarray2i(spans)
+            lengths_array = ops.asarray1i(lengths)
+            if len(spans) > 0:
+                output = Ragged(ops.xp.vstack(spans), lengths_array)
+            else:
+                output = Ragged(ops.xp.zeros((0, 0), dtype="i"), lengths_array)
+            return output
 
-        return zero_suggester
+        return mixed_zero_suggester
 
     fix_random_seed(0)
     nlp = English()
     spancat = nlp.add_pipe(
         "spancat",
-        config={"suggester": {"@misc": "test_zero_suggester"}, "spans_key": SPAN_KEY},
+        config={
+            "suggester": {"@misc": "test_mixed_zero_suggester"},
+            "spans_key": SPAN_KEY,
+        },
     )
     train_examples = make_examples(nlp)
     optimizer = nlp.initialize(get_examples=lambda: train_examples)
@@ -397,6 +412,16 @@ def test_zero_suggestions():
     assert set(spancat.labels) == {"LOC", "PERSON"}
 
     nlp.update(train_examples, sgd=optimizer)
+    # empty doc
+    nlp("")
+    # single doc with zero suggestions
+    nlp("one")
+    # single doc with one suggestion
+    nlp("two two")
+    # batch with mixed zero/one suggestions
+    list(nlp.pipe(["one", "two two", "three three three", "", "four four four four"]))
+    # batch with no suggestions
+    list(nlp.pipe(["", "one", "three three three"]))
 
 
 def test_set_candidates():

spacy/tests/pipeline/test_textcat.py

@@ -361,6 +361,30 @@ def test_label_types(name):
         nlp.initialize()
 
 
+@pytest.mark.parametrize(
+    "name,get_examples",
+    [
+        ("textcat", make_get_examples_single_label),
+        ("textcat_multilabel", make_get_examples_multi_label),
+    ],
+)
+def test_invalid_label_value(name, get_examples):
+    nlp = Language()
+    textcat = nlp.add_pipe(name)
+    example_getter = get_examples(nlp)
+
+    def invalid_examples():
+        # make one example with an invalid score
+        examples = example_getter()
+        ref = examples[0].reference
+        key = list(ref.cats.keys())[0]
+        ref.cats[key] = 2.0
+        return examples
+
+    with pytest.raises(ValueError):
+        nlp.initialize(get_examples=invalid_examples)
+
+
 @pytest.mark.parametrize("name", ["textcat", "textcat_multilabel"])
 def test_no_label(name):
     nlp = Language()
@@ -815,8 +839,8 @@ def test_textcat_loss(multi_label: bool, expected_loss: float):
         textcat = nlp.add_pipe("textcat_multilabel")
     else:
         textcat = nlp.add_pipe("textcat")
-    textcat.initialize(lambda: train_examples)
     assert isinstance(textcat, TextCategorizer)
+    textcat.initialize(lambda: train_examples)
     scores = textcat.model.ops.asarray(
         [[0.0, 0.0, 0.0, 1.0], [0.0, 0.0, 1.0, 1.0]], dtype="f"  # type: ignore
     )

spacy/tests/serialize/test_serialize_pipeline.py

@@ -361,11 +361,10 @@ def test_serialize_pipeline_disable_enable():
     assert nlp3.component_names == ["ner", "tagger"]
     with make_tempdir() as d:
         nlp3.to_disk(d)
-        with pytest.warns(UserWarning):
+        nlp4 = spacy.load(d, disable=["ner"])
-            nlp4 = spacy.load(d, disable=["ner"])
+    assert nlp4.pipe_names == []
-    assert nlp4.pipe_names == ["tagger"]
     assert nlp4.component_names == ["ner", "tagger"]
-    assert nlp4.disabled == ["ner"]
+    assert nlp4.disabled == ["ner", "tagger"]
     with make_tempdir() as d:
         nlp.to_disk(d)
         nlp5 = spacy.load(d, exclude=["tagger"])

spacy/tests/test_cli.py

@@ -1,8 +1,12 @@
 import os
 import math
-from random import sample
+from collections import Counter
-from typing import Counter
+from typing import Tuple, List, Dict, Any
+import pkg_resources
+import time
 
+import spacy
+import numpy
 import pytest
 import srsly
 from click import NoSuchOption
@@ -15,6 +19,7 @@ from spacy.cli._util import is_subpath_of, load_project_config
 from spacy.cli._util import parse_config_overrides, string_to_list
 from spacy.cli._util import substitute_project_variables
 from spacy.cli._util import validate_project_commands
+from spacy.cli._util import upload_file, download_file
 from spacy.cli.debug_data import _compile_gold, _get_labels_from_model
 from spacy.cli.debug_data import _get_labels_from_spancat
 from spacy.cli.debug_data import _get_distribution, _get_kl_divergence
@@ -25,12 +30,16 @@ from spacy.cli.download import get_compatibility, get_version
 from spacy.cli.init_config import RECOMMENDATIONS, init_config, fill_config
 from spacy.cli.package import get_third_party_dependencies
 from spacy.cli.package import _is_permitted_package_name
+from spacy.cli.project.remote_storage import RemoteStorage
+from spacy.cli.project.run import _check_requirements
 from spacy.cli.validate import get_model_pkgs
+from spacy.cli.apply import apply
+from spacy.cli.find_threshold import find_threshold
 from spacy.lang.en import English
 from spacy.lang.nl import Dutch
 from spacy.language import Language
 from spacy.schemas import ProjectConfigSchema, RecommendationSchema, validate
-from spacy.tokens import Doc
+from spacy.tokens import Doc, DocBin
 from spacy.tokens.span import Span
 from spacy.training import Example, docs_to_json, offsets_to_biluo_tags
 from spacy.training.converters import conll_ner_to_docs, conllu_to_docs
@@ -116,6 +125,25 @@ def test_issue7055():
     assert "model" in filled_cfg["components"]["ner"]
 
 
+@pytest.mark.issue(11235)
+def test_issue11235():
+    """
+    Test that the cli handles interpolation in the directory names correctly when loading project config.
+    """
+    lang_var = "en"
+    variables = {"lang": lang_var}
+    commands = [{"name": "x", "script": ["hello ${vars.lang}"]}]
+    directories = ["cfg", "${vars.lang}_model"]
+    project = {"commands": commands, "vars": variables, "directories": directories}
+    with make_tempdir() as d:
+        srsly.write_yaml(d / "project.yml", project)
+        cfg = load_project_config(d)
+        # Check that the directories are interpolated and created correctly
+        assert os.path.exists(d / "cfg")
+        assert os.path.exists(d / f"{lang_var}_model")
+    assert cfg["commands"][0]["script"][0] == f"hello {lang_var}"
+
+
 def test_cli_info():
     nlp = Dutch()
     nlp.add_pipe("textcat")
@@ -589,6 +617,7 @@ def test_string_to_list_intify(value):
     assert string_to_list(value, intify=True) == [1, 2, 3]
 
 
+@pytest.mark.skip(reason="Temporarily skip for dev version")
 def test_download_compatibility():
     spec = SpecifierSet("==" + about.__version__)
     spec.prereleases = False
@@ -599,6 +628,7 @@ def test_download_compatibility():
         assert get_minor_version(about.__version__) == get_minor_version(version)
 
 
+@pytest.mark.skip(reason="Temporarily skip for dev version")
 def test_validate_compatibility_table():
     spec = SpecifierSet("==" + about.__version__)
     spec.prereleases = False
@@ -855,3 +885,303 @@ def test_span_length_freq_dist_output_must_be_correct():
     span_freqs = _get_spans_length_freq_dist(sample_span_lengths, threshold)
     assert sum(span_freqs.values()) >= threshold
     assert list(span_freqs.keys()) == [3, 1, 4, 5, 2]
+
+
+def test_applycli_empty_dir():
+    with make_tempdir() as data_path:
+        output = data_path / "test.spacy"
+        apply(data_path, output, "blank:en", "text", 1, 1)
+
+
+def test_applycli_docbin():
+    with make_tempdir() as data_path:
+        output = data_path / "testout.spacy"
+        nlp = spacy.blank("en")
+        doc = nlp("testing apply cli.")
+        # test empty DocBin case
+        docbin = DocBin()
+        docbin.to_disk(data_path / "testin.spacy")
+        apply(data_path, output, "blank:en", "text", 1, 1)
+        docbin.add(doc)
+        docbin.to_disk(data_path / "testin.spacy")
+        apply(data_path, output, "blank:en", "text", 1, 1)
+
+
+def test_applycli_jsonl():
+    with make_tempdir() as data_path:
+        output = data_path / "testout.spacy"
+        data = [{"field": "Testing apply cli.", "key": 234}]
+        data2 = [{"field": "234"}]
+        srsly.write_jsonl(data_path / "test.jsonl", data)
+        apply(data_path, output, "blank:en", "field", 1, 1)
+        srsly.write_jsonl(data_path / "test2.jsonl", data2)
+        apply(data_path, output, "blank:en", "field", 1, 1)
+
+
+def test_applycli_txt():
+    with make_tempdir() as data_path:
+        output = data_path / "testout.spacy"
+        with open(data_path / "test.foo", "w") as ftest:
+            ftest.write("Testing apply cli.")
+        apply(data_path, output, "blank:en", "text", 1, 1)
+
+
+def test_applycli_mixed():
+    with make_tempdir() as data_path:
+        output = data_path / "testout.spacy"
+        text = "Testing apply cli"
+        nlp = spacy.blank("en")
+        doc = nlp(text)
+        jsonl_data = [{"text": text}]
+        srsly.write_jsonl(data_path / "test.jsonl", jsonl_data)
+        docbin = DocBin()
+        docbin.add(doc)
+        docbin.to_disk(data_path / "testin.spacy")
+        with open(data_path / "test.txt", "w") as ftest:
+            ftest.write(text)
+        apply(data_path, output, "blank:en", "text", 1, 1)
+        # Check whether it worked
+        result = list(DocBin().from_disk(output).get_docs(nlp.vocab))
+        assert len(result) == 3
+        for doc in result:
+            assert doc.text == text
+
+
+def test_applycli_user_data():
+    Doc.set_extension("ext", default=0)
+    val = ("ext", 0)
+    with make_tempdir() as data_path:
+        output = data_path / "testout.spacy"
+        nlp = spacy.blank("en")
+        doc = nlp("testing apply cli.")
+        doc._.ext = val
+        docbin = DocBin(store_user_data=True)
+        docbin.add(doc)
+        docbin.to_disk(data_path / "testin.spacy")
+        apply(data_path, output, "blank:en", "", 1, 1)
+        result = list(DocBin().from_disk(output).get_docs(nlp.vocab))
+        assert result[0]._.ext == val
+
+
+def test_local_remote_storage():
+    with make_tempdir() as d:
+        filename = "a.txt"
+
+        content_hashes = ("aaaa", "cccc", "bbbb")
+        for i, content_hash in enumerate(content_hashes):
+            # make sure that each subsequent file has a later timestamp
+            if i > 0:
+                time.sleep(1)
+            content = f"{content_hash} content"
+            loc_file = d / "root" / filename
+            if not loc_file.parent.exists():
+                loc_file.parent.mkdir(parents=True)
+            with loc_file.open(mode="w") as file_:
+                file_.write(content)
+
+            # push first version to remote storage
+            remote = RemoteStorage(d / "root", str(d / "remote"))
+            remote.push(filename, "aaaa", content_hash)
+
+            # retrieve with full hashes
+            loc_file.unlink()
+            remote.pull(filename, command_hash="aaaa", content_hash=content_hash)
+            with loc_file.open(mode="r") as file_:
+                assert file_.read() == content
+
+            # retrieve with command hash
+            loc_file.unlink()
+            remote.pull(filename, command_hash="aaaa")
+            with loc_file.open(mode="r") as file_:
+                assert file_.read() == content
+
+            # retrieve with content hash
+            loc_file.unlink()
+            remote.pull(filename, content_hash=content_hash)
+            with loc_file.open(mode="r") as file_:
+                assert file_.read() == content
+
+            # retrieve with no hashes
+            loc_file.unlink()
+            remote.pull(filename)
+            with loc_file.open(mode="r") as file_:
+                assert file_.read() == content
+
+
+def test_local_remote_storage_pull_missing():
+    # pulling from a non-existent remote pulls nothing gracefully
+    with make_tempdir() as d:
+        filename = "a.txt"
+        remote = RemoteStorage(d / "root", str(d / "remote"))
+        assert remote.pull(filename, command_hash="aaaa") is None
+        assert remote.pull(filename) is None
+
+
+def test_cli_find_threshold(capsys):
+    thresholds = numpy.linspace(0, 1, 10)
+
+    def make_examples(nlp: Language) -> List[Example]:
+        docs: List[Example] = []
+
+        for t in [
+            (
+                "I am angry and confused in the Bank of America.",
+                {
+                    "cats": {"ANGRY": 1.0, "CONFUSED": 1.0, "HAPPY": 0.0},
+                    "spans": {"sc": [(31, 46, "ORG")]},
+                },
+            ),
+            (
+                "I am confused but happy in New York.",
+                {
+                    "cats": {"ANGRY": 0.0, "CONFUSED": 1.0, "HAPPY": 1.0},
+                    "spans": {"sc": [(27, 35, "GPE")]},
+                },
+            ),
+        ]:
+            doc = nlp.make_doc(t[0])
+            docs.append(Example.from_dict(doc, t[1]))
+
+        return docs
+
+    def init_nlp(
+        components: Tuple[Tuple[str, Dict[str, Any]], ...] = ()
+    ) -> Tuple[Language, List[Example]]:
+        new_nlp = English()
+        new_nlp.add_pipe(  # type: ignore
+            factory_name="textcat_multilabel",
+            name="tc_multi",
+            config={"threshold": 0.9},
+        )
+
+        # Append additional components to pipeline.
+        for cfn, comp_config in components:
+            new_nlp.add_pipe(cfn, config=comp_config)
+
+        new_examples = make_examples(new_nlp)
+        new_nlp.initialize(get_examples=lambda: new_examples)
+        for i in range(5):
+            new_nlp.update(new_examples)
+
+        return new_nlp, new_examples
+
+    with make_tempdir() as docs_dir:
+        # Check whether find_threshold() identifies lowest threshold above 0 as (first) ideal threshold, as this matches
+        # the current model behavior with the examples above. This can break once the model behavior changes and serves
+        # mostly as a smoke test.
+        nlp, examples = init_nlp()
+        DocBin(docs=[example.reference for example in examples]).to_disk(
+            docs_dir / "docs.spacy"
+        )
+        with make_tempdir() as nlp_dir:
+            nlp.to_disk(nlp_dir)
+            res = find_threshold(
+                model=nlp_dir,
+                data_path=docs_dir / "docs.spacy",
+                pipe_name="tc_multi",
+                threshold_key="threshold",
+                scores_key="cats_macro_f",
+                silent=True,
+            )
+            assert res[0] != thresholds[0]
+            assert thresholds[0] < res[0] < thresholds[9]
+            assert res[1] == 1.0
+            assert res[2][1.0] == 0.0
+
+        # Test with spancat.
+        nlp, _ = init_nlp((("spancat", {}),))
+        with make_tempdir() as nlp_dir:
+            nlp.to_disk(nlp_dir)
+            res = find_threshold(
+                model=nlp_dir,
+                data_path=docs_dir / "docs.spacy",
+                pipe_name="spancat",
+                threshold_key="threshold",
+                scores_key="spans_sc_f",
+                silent=True,
+            )
+            assert res[0] != thresholds[0]
+            assert thresholds[0] < res[0] < thresholds[8]
+            assert res[1] >= 0.6
+            assert res[2][1.0] == 0.0
+
+        # Having multiple textcat_multilabel components should work, since the name has to be specified.
+        nlp, _ = init_nlp((("textcat_multilabel", {}),))
+        with make_tempdir() as nlp_dir:
+            nlp.to_disk(nlp_dir)
+            assert find_threshold(
+                model=nlp_dir,
+                data_path=docs_dir / "docs.spacy",
+                pipe_name="tc_multi",
+                threshold_key="threshold",
+                scores_key="cats_macro_f",
+                silent=True,
+            )
+
+        # Specifying the name of an non-existing pipe should fail.
+        nlp, _ = init_nlp()
+        with make_tempdir() as nlp_dir:
+            nlp.to_disk(nlp_dir)
+            with pytest.raises(AttributeError):
+                find_threshold(
+                    model=nlp_dir,
+                    data_path=docs_dir / "docs.spacy",
+                    pipe_name="_",
+                    threshold_key="threshold",
+                    scores_key="cats_macro_f",
+                    silent=True,
+                )
+
+
+@pytest.mark.parametrize(
+    "reqs,output",
+    [
+        [
+            """
+            spacy
+
+            # comment
+
+            thinc""",
+            (False, False),
+        ],
+        [
+            """# comment
+            --some-flag
+            spacy""",
+            (False, False),
+        ],
+        [
+            """# comment
+            --some-flag
+            spacy; python_version >= '3.6'""",
+            (False, False),
+        ],
+        [
+            """# comment
+             spacyunknowndoesnotexist12345""",
+            (True, False),
+        ],
+    ],
+)
+def test_project_check_requirements(reqs, output):
+    # excessive guard against unlikely package name
+    try:
+        pkg_resources.require("spacyunknowndoesnotexist12345")
+    except pkg_resources.DistributionNotFound:
+        assert output == _check_requirements([req.strip() for req in reqs.split("\n")])
+
+
+def test_upload_download_local_file():
+    with make_tempdir() as d1, make_tempdir() as d2:
+        filename = "f.txt"
+        content = "content"
+        local_file = d1 / filename
+        remote_file = d2 / filename
+        with local_file.open(mode="w") as file_:
+            file_.write(content)
+        upload_file(local_file, remote_file)
+        local_file.unlink()
+        download_file(remote_file, local_file)
+        with local_file.open(mode="r") as file_:
+            assert file_.read() == content

spacy/tests/test_displacy.py

@@ -203,6 +203,16 @@ def test_displacy_parse_spans_different_spans_key(en_vocab):
     ]
 
 
+def test_displacy_parse_empty_spans_key(en_vocab):
+    """Test that having an unset spans key doesn't raise an error"""
+    doc = Doc(en_vocab, words=["Welcome", "to", "the", "Bank", "of", "China"])
+    doc.spans["custom"] = [Span(doc, 3, 6, "BANK")]
+    with pytest.warns(UserWarning, match="W117"):
+        spans = displacy.parse_spans(doc)
+
+    assert isinstance(spans, dict)
+
+
 def test_displacy_parse_ents(en_vocab):
     """Test that named entities on a Doc are converted into displaCy's format."""
     doc = Doc(en_vocab, words=["But", "Google", "is", "starting", "from", "behind"])

spacy/tests/training/test_training.py

@@ -2,6 +2,7 @@ import random
 
 import numpy
 import pytest
+import spacy
 import srsly
 from spacy.lang.en import English
 from spacy.tokens import Doc, DocBin
@@ -11,9 +12,10 @@ from spacy.training import offsets_to_biluo_tags
 from spacy.training.alignment_array import AlignmentArray
 from spacy.training.align import get_alignments
 from spacy.training.converters import json_to_docs
+from spacy.training.loop import train_while_improving
 from spacy.util import get_words_and_spaces, load_model_from_path, minibatch
 from spacy.util import load_config_from_str
-from thinc.api import compounding
+from thinc.api import compounding, Adam
 
 from ..util import make_tempdir
 
@@ -1112,3 +1114,39 @@ def test_retokenized_docs(doc):
         retokenizer.merge(doc1[0:2])
         retokenizer.merge(doc1[5:7])
     assert example.get_aligned("ORTH", as_string=True) == expected2
+
+
+def test_training_before_update(doc):
+    def before_update(nlp, args):
+        assert args["step"] == 0
+        assert args["epoch"] == 1
+
+        # Raise an error here as the rest of the loop
+        # will not run to completion due to uninitialized
+        # models.
+        raise ValueError("ran_before_update")
+
+    def generate_batch():
+        yield 1, [Example(doc, doc)]
+
+    nlp = spacy.blank("en")
+    nlp.add_pipe("tagger")
+    optimizer = Adam()
+    generator = train_while_improving(
+        nlp,
+        optimizer,
+        generate_batch(),
+        lambda: None,
+        dropout=0.1,
+        eval_frequency=100,
+        accumulate_gradient=10,
+        patience=10,
+        max_steps=100,
+        exclude=[],
+        annotating_components=[],
+        before_update=before_update,
+    )
+
+    with pytest.raises(ValueError, match="ran_before_update"):
+        for _ in generator:
+            pass

spacy/tests/vocab_vectors/test_vectors.py

@@ -626,3 +626,23 @@ def test_floret_vectors(floret_vectors_vec_str, floret_vectors_hashvec_str):
             OPS.to_numpy(vocab_r[word].vector),
             decimal=6,
         )
+
+
+def test_equality():
+    vectors1 = Vectors(shape=(10, 10))
+    vectors2 = Vectors(shape=(10, 8))
+
+    assert vectors1 != vectors2
+
+    vectors2 = Vectors(shape=(10, 10))
+    assert vectors1 == vectors2
+
+    vectors1.add("hello", row=2)
+    assert vectors1 != vectors2
+
+    vectors2.add("hello", row=2)
+    assert vectors1 == vectors2
+
+    vectors1.resize((5, 9))
+    vectors2.resize((5, 9))
+    assert vectors1 == vectors2

spacy/tests/vocab_vectors/test_vocab_api.py

@@ -1,8 +1,13 @@
+import os
+
 import pytest
 from spacy.attrs import IS_ALPHA, LEMMA, ORTH
+from spacy.lang.en import English
 from spacy.parts_of_speech import NOUN, VERB
 from spacy.vocab import Vocab
 
+from ..util import make_tempdir
+
 
 @pytest.mark.issue(1868)
 def test_issue1868():
@@ -59,3 +64,19 @@ def test_vocab_api_contains(en_vocab, text):
 def test_vocab_writing_system(en_vocab):
     assert en_vocab.writing_system["direction"] == "ltr"
     assert en_vocab.writing_system["has_case"] is True
+
+
+def test_to_disk():
+    nlp = English()
+    with make_tempdir() as d:
+        nlp.vocab.to_disk(d)
+        assert "vectors" in os.listdir(d)
+        assert "lookups.bin" in os.listdir(d)
+
+
+def test_to_disk_exclude():
+    nlp = English()
+    with make_tempdir() as d:
+        nlp.vocab.to_disk(d, exclude=("vectors", "lookups"))
+        assert "vectors" not in os.listdir(d)
+        assert "lookups.bin" not in os.listdir(d)

spacy/tokens/doc.pxd

@@ -48,8 +48,6 @@ cdef class Doc:
 
     cdef TokenC* c
 
-    cdef public float sentiment
-
     cdef public dict activations
 
     cdef public dict user_hooks

spacy/tokens/doc.pyi

@@ -21,7 +21,6 @@ class Doc:
     spans: SpanGroups
     max_length: int
     length: int
-    sentiment: float
     activations: Dict[str, Dict[str, Union[ArrayXd, Ragged]]]
     cats: Dict[str, float]
     user_hooks: Dict[str, Callable[..., Any]]

spacy/tokens/doc.pyx

@@ -243,7 +243,6 @@ cdef class Doc:
         self.c = data_start + PADDING
         self.max_length = size
         self.length = 0
-        self.sentiment = 0.0
         self.cats = {}
         self.activations = {}
         self.user_hooks = {}
@@ -360,6 +359,7 @@ cdef class Doc:
             for annot in annotations:
                 if annot:
                     if annot is heads or annot is sent_starts or annot is ent_iobs:
+                        annot = numpy.array(annot, dtype=numpy.int32).astype(numpy.uint64)
                         for i in range(len(words)):
                             if attrs.ndim == 1:
                                 attrs[i] = annot[i]
@@ -1178,13 +1178,22 @@ cdef class Doc:
 
             if "user_data" not in exclude:
                 for key, value in doc.user_data.items():
-                    if isinstance(key, tuple) and len(key) == 4 and key[0] == "._.":
+                    if isinstance(key, tuple) and len(key) >= 4 and key[0] == "._.":
-                        data_type, name, start, end = key
+                        data_type = key[0]
+                        name = key[1]
+                        start = key[2]
+                        end = key[3]
                         if start is not None or end is not None:
                             start += char_offset
                             if end is not None:
                                 end += char_offset
-                            concat_user_data[(data_type, name, start, end)] = copy.copy(value)
+                                _label = key[4]
+                                _kb_id = key[5]
+                                _span_id = key[6]
+                                concat_user_data[(data_type, name, start, end, _label, _kb_id, _span_id)] = copy.copy(value)
+                            else:
+                                concat_user_data[(data_type, name, start, end)] = copy.copy(value)
+
                         else:
                             warnings.warn(Warnings.W101.format(name=name))
                     else:
@@ -1270,7 +1279,6 @@ cdef class Doc:
         other.tensor = copy.deepcopy(self.tensor)
         other.cats = copy.deepcopy(self.cats)
         other.user_data = copy.deepcopy(self.user_data)
-        other.sentiment = self.sentiment
         other.has_unknown_spaces = self.has_unknown_spaces
         other.user_hooks = dict(self.user_hooks)
         other.user_token_hooks = dict(self.user_token_hooks)
@@ -1367,7 +1375,6 @@ cdef class Doc:
             "text": lambda: self.text,
             "array_head": lambda: array_head,
             "array_body": lambda: self.to_array(array_head),
-            "sentiment": lambda: self.sentiment,
             "tensor": lambda: self.tensor,
             "cats": lambda: self.cats,
             "spans": lambda: self.spans.to_bytes(),
@@ -1405,8 +1412,6 @@ cdef class Doc:
             for key, value in zip(user_data_keys, user_data_values):
                 self.user_data[key] = value
         cdef int i, start, end, has_space
-        if "sentiment" not in exclude and "sentiment" in msg:
-            self.sentiment = msg["sentiment"]
         if "tensor" not in exclude and "tensor" in msg:
             self.tensor = msg["tensor"]
         if "cats" not in exclude and "cats" in msg:
@@ -1569,6 +1574,7 @@ cdef class Doc:
 
             for j, (attr, annot) in enumerate(token_annotations.items()):
                 if attr is HEAD:
+                    annot = numpy.array(annot, dtype=numpy.int32).astype(numpy.uint64)
                     for i in range(len(words)):
                         array[i, j] = annot[i]
                 elif attr is MORPH:
@@ -1632,7 +1638,11 @@ cdef class Doc:
                 Span.set_extension(span_attr)
             for span_data in doc_json["underscore_span"][span_attr]:
                 value = span_data["value"]
-                self.char_span(span_data["start"], span_data["end"])._.set(span_attr, value)
+                span = self.char_span(span_data["start"], span_data["end"])
+                span.label = span_data["label"]
+                span.kb_id = span_data["kb_id"]
+                span.id = span_data["id"]
+                span._.set(span_attr, value)
         return self
 
     def to_json(self, underscore=None):
@@ -1679,6 +1689,20 @@ cdef class Doc:
 
         if underscore:
             user_keys = set()
+            # Handle doc attributes with .get to include values from getters
+            # and not only values stored in user_data, for backwards
+            # compatibility
+            for attr in underscore:
+                if self.has_extension(attr):
+                    if "_" not in data:
+                        data["_"] = {}
+                    value = self._.get(attr)
+                    if not srsly.is_json_serializable(value):
+                        raise ValueError(Errors.E107.format(attr=attr, value=repr(value)))
+                    data["_"][attr] = value
+                    user_keys.add(attr)
+            # Token and span attributes only include values stored in user_data
+            # and not values generated by getters
             if self.user_data:
                 for data_key, value in self.user_data.copy().items():
                     if type(data_key) == tuple and len(data_key) >= 4 and data_key[0] == "._.":
@@ -1689,25 +1713,23 @@ cdef class Doc:
                             user_keys.add(attr)
                             if not srsly.is_json_serializable(value):
                                 raise ValueError(Errors.E107.format(attr=attr, value=repr(value)))
-                            # Check if doc attribute
+                            # Token attribute
-                            if start is None:
+                            if start is not None and end is None:
-                                if "_" not in data:
-                                    data["_"] = {}
-                                data["_"][attr] = value
-                            # Check if token attribute
-                            elif end is None:
                                 if "underscore_token" not in data:
                                     data["underscore_token"] = {}
                                 if attr not in data["underscore_token"]:
                                     data["underscore_token"][attr] = []
                                 data["underscore_token"][attr].append({"start": start, "value": value})
                             # Else span attribute
-                            else:
+                            elif end is not None:
+                                _label = data_key[4]
+                                _kb_id = data_key[5]
+                                _span_id = data_key[6]
                                 if "underscore_span" not in data:
                                     data["underscore_span"] = {}
                                 if attr not in data["underscore_span"]:
                                     data["underscore_span"][attr] = []
-                                data["underscore_span"][attr].append({"start": start, "end": end, "value": value})
+                                data["underscore_span"][attr].append({"start": start, "end": end, "value": value, "label": _label, "kb_id": _kb_id, "id":_span_id})
 
             for attr in underscore:
                 if attr not in user_keys:

spacy/tokens/span.pyi

@@ -82,8 +82,6 @@ class Span:
     @property
     def tensor(self) -> FloatsXd: ...
     @property
-    def sentiment(self) -> float: ...
-    @property
     def text(self) -> str: ...
     @property
     def text_with_ws(self) -> str: ...
@@ -95,8 +93,8 @@ class Span:
         self,
         start_idx: int,
         end_idx: int,
-        label: int = ...,
+        label: Union[int, str] = ...,
-        kb_id: int = ...,
+        kb_id: Union[int, str] = ...,
         vector: Optional[Floats1d] = ...,
     ) -> Span: ...
     @property

spacy/tokens/span.pyx

@@ -218,11 +218,10 @@ cdef class Span:
         cdef SpanC* span_c = self.span_c()
         """Custom extension attributes registered via `set_extension`."""
         return Underscore(Underscore.span_extensions, self,
-                          start=span_c.start_char, end=span_c.end_char)
+                          start=span_c.start_char, end=span_c.end_char, label=self.label, kb_id=self.kb_id, span_id=self.id)
 
     def as_doc(self, *, bint copy_user_data=False, array_head=None, array=None):
         """Create a `Doc` object with a copy of the `Span`'s data.
-
         copy_user_data (bool): Whether or not to copy the original doc's user data.
         array_head (tuple): `Doc` array attrs, can be passed in to speed up computation.
         array (ndarray): `Doc` as array, can be passed in to speed up computation.
@@ -275,12 +274,22 @@ cdef class Span:
             char_offset = self.start_char
             for key, value in self.doc.user_data.items():
                 if isinstance(key, tuple) and len(key) == 4 and key[0] == "._.":
-                    data_type, name, start, end = key
+                    data_type = key[0]
+                    name = key[1]
+                    start = key[2]
+                    end = key[3]
                     if start is not None or end is not None:
                         start -= char_offset
+                        # Check if Span object
                         if end is not None:
                             end -= char_offset
-                        user_data[(data_type, name, start, end)] = copy.copy(value)
+                            _label = key[4]
+                            _kb_id = key[5]
+                            _span_id = key[6]
+                            user_data[(data_type, name, start, end, _label, _kb_id, _span_id)] = copy.copy(value)
+                        # Else Token object
+                        else:
+                            user_data[(data_type, name, start, end)] = copy.copy(value)
                 else:
                     user_data[key] = copy.copy(value)
             doc.user_data = user_data
@@ -309,7 +318,7 @@ cdef class Span:
                     for ancestor in ancestors:
                         ancestor_i = ancestor.i - span_c.start
                         if ancestor_i in range(length):
-                            array[i, head_col] = ancestor_i - i
+                            array[i, head_col] = numpy.int32(ancestor_i - i).astype(numpy.uint64)
 
                 # if there is no appropriate ancestor, define a new artificial root
                 value = array[i, head_col]
@@ -317,7 +326,7 @@ cdef class Span:
                     new_root = old_to_new_root.get(ancestor_i, None)
                     if new_root is not None:
                         # take the same artificial root as a previous token from the same sentence
-                        array[i, head_col] = new_root - i
+                        array[i, head_col] = numpy.int32(new_root - i).astype(numpy.uint64)
                     else:
                         # set this token as the new artificial root
                         array[i, head_col] = 0
@@ -566,16 +575,6 @@ cdef class Span:
             return None
         return self.doc.tensor[self.start : self.end]
 
-    @property
-    def sentiment(self):
-        """RETURNS (float): A scalar value indicating the positivity or
-            negativity of the span.
-        """
-        if "sentiment" in self.doc.user_span_hooks:
-            return self.doc.user_span_hooks["sentiment"](self)
-        else:
-            return sum([token.sentiment for token in self]) / len(self)
-
     @property
     def text(self):
         """RETURNS (str): The original verbatim text of the span."""
@@ -791,21 +790,36 @@ cdef class Span:
             return self.span_c().label
 
         def __set__(self, attr_t label):
-            self.span_c().label = label
+            if label != self.span_c().label :
+                old_label = self.span_c().label
+                self.span_c().label = label
+                new = Underscore(Underscore.span_extensions, self, start=self.span_c().start_char, end=self.span_c().end_char, label=self.label, kb_id=self.kb_id, span_id=self.id)
+                old = Underscore(Underscore.span_extensions, self, start=self.span_c().start_char, end=self.span_c().end_char, label=old_label, kb_id=self.kb_id, span_id=self.id)
+                Underscore._replace_keys(old, new)
 
     property kb_id:
         def __get__(self):
             return self.span_c().kb_id
 
         def __set__(self, attr_t kb_id):
-            self.span_c().kb_id = kb_id
+            if kb_id != self.span_c().kb_id :
+                old_kb_id = self.span_c().kb_id
+                self.span_c().kb_id = kb_id
+                new = Underscore(Underscore.span_extensions, self, start=self.span_c().start_char, end=self.span_c().end_char, label=self.label, kb_id=self.kb_id, span_id=self.id)
+                old = Underscore(Underscore.span_extensions, self, start=self.span_c().start_char, end=self.span_c().end_char, label=self.label, kb_id=old_kb_id, span_id=self.id)
+                Underscore._replace_keys(old, new)
 
     property id:
         def __get__(self):
             return self.span_c().id
 
         def __set__(self, attr_t id):
-            self.span_c().id = id
+            if id != self.span_c().id :
+                old_id = self.span_c().id
+                self.span_c().id = id
+                new = Underscore(Underscore.span_extensions, self, start=self.span_c().start_char, end=self.span_c().end_char, label=self.label, kb_id=self.kb_id, span_id=self.id)
+                old = Underscore(Underscore.span_extensions, self, start=self.span_c().start_char, end=self.span_c().end_char, label=self.label, kb_id=self.kb_id, span_id=old_id)
+                Underscore._replace_keys(old, new)
 
     property ent_id:
         """Alias for the span's ID."""

spacy/tokens/span_group.pyi

@@ -18,6 +18,7 @@ class SpanGroup:
     def doc(self) -> Doc: ...
     @property
     def has_overlap(self) -> bool: ...
+    def __iter__(self): ...
     def __len__(self) -> int: ...
     def append(self, span: Span) -> None: ...
     def extend(self, spans: Iterable[Span]) -> None: ...

spacy/tokens/span_group.pyx

@@ -159,6 +159,16 @@ cdef class SpanGroup:
             return self._concat(other)
         return NotImplemented
 
+    def __iter__(self):
+        """
+        Iterate over the spans in this SpanGroup.
+        YIELDS (Span): A span in this SpanGroup.
+
+        DOCS: https://spacy.io/api/spangroup#iter
+        """
+        for i in range(self.c.size()):
+            yield self[i]
+
     def append(self, Span span):
         """Add a span to the group. The span must refer to the same Doc
         object as the span group.

spacy/tokens/token.pyi

@@ -79,8 +79,6 @@ class Token:
     @property
     def prob(self) -> float: ...
     @property
-    def sentiment(self) -> float: ...
-    @property
     def lang(self) -> int: ...
     @property
     def idx(self) -> int: ...

spacy/tokens/token.pyx

@@ -283,14 +283,6 @@ cdef class Token:
         """RETURNS (float): Smoothed log probability estimate of token type."""
         return self.vocab[self.c.lex.orth].prob
 
-    @property
-    def sentiment(self):
-        """RETURNS (float): A scalar value indicating the positivity or
-            negativity of the token."""
-        if "sentiment" in self.doc.user_token_hooks:
-            return self.doc.user_token_hooks["sentiment"](self)
-        return self.vocab[self.c.lex.orth].sentiment
-
     @property
     def lang(self):
         """RETURNS (uint64): ID of the language of the parent document's

spacy/tokens/underscore.py

@@ -2,10 +2,10 @@ from typing import Dict, Any, List, Optional, Tuple, Union, TYPE_CHECKING
 import functools
 import copy
 from ..errors import Errors
+from .span import Span
 
 if TYPE_CHECKING:
     from .doc import Doc
-    from .span import Span
     from .token import Token
 
 
@@ -25,6 +25,9 @@ class Underscore:
         obj: Union["Doc", "Span", "Token"],
         start: Optional[int] = None,
         end: Optional[int] = None,
+        label: int = 0,
+        kb_id: int = 0,
+        span_id: int = 0,
     ):
         object.__setattr__(self, "_extensions", extensions)
         object.__setattr__(self, "_obj", obj)
@@ -36,6 +39,10 @@ class Underscore:
         object.__setattr__(self, "_doc", obj.doc)
         object.__setattr__(self, "_start", start)
         object.__setattr__(self, "_end", end)
+        if type(obj) == Span:
+            object.__setattr__(self, "_label", label)
+            object.__setattr__(self, "_kb_id", kb_id)
+            object.__setattr__(self, "_span_id", span_id)
 
     def __dir__(self) -> List[str]:
         # Hack to enable autocomplete on custom extensions
@@ -88,8 +95,39 @@ class Underscore:
     def has(self, name: str) -> bool:
         return name in self._extensions
 
-    def _get_key(self, name: str) -> Tuple[str, str, Optional[int], Optional[int]]:
+    def _get_key(
-        return ("._.", name, self._start, self._end)
+        self, name: str
+    ) -> Union[
+        Tuple[str, str, Optional[int], Optional[int]],
+        Tuple[str, str, Optional[int], Optional[int], int, int, int],
+    ]:
+        if hasattr(self, "_label"):
+            return (
+                "._.",
+                name,
+                self._start,
+                self._end,
+                self._label,
+                self._kb_id,
+                self._span_id,
+            )
+        else:
+            return "._.", name, self._start, self._end
+
+    @staticmethod
+    def _replace_keys(old_underscore: "Underscore", new_underscore: "Underscore"):
+        """
+        This function is called by Span when its kb_id or label are re-assigned.
+        It checks if any user_data is stored for this span and replaces the keys
+        """
+        for name in old_underscore._extensions:
+            old_key = old_underscore._get_key(name)
+            old_doc = old_underscore._doc
+            new_key = new_underscore._get_key(name)
+            if old_key != new_key and old_key in old_doc.user_data:
+                old_underscore._doc.user_data[
+                    new_key
+                ] = old_underscore._doc.user_data.pop(old_key)
 
     @classmethod
     def get_state(cls) -> Tuple[Dict[Any, Any], Dict[Any, Any], Dict[Any, Any]]:

spacy/training/batchers.py

@@ -2,11 +2,12 @@ from typing import Union, Iterable, Sequence, TypeVar, List, Callable, Iterator
 from typing import Optional, Any
 from functools import partial
 import itertools
+from thinc.schedules import Schedule, constant as constant_schedule
 
 from ..util import registry, minibatch
 
 
-Sizing = Union[Sequence[int], int]
+Sizing = Union[Sequence[int], int, Schedule[int]]
 ItemT = TypeVar("ItemT")
 BatcherT = Callable[[Iterable[ItemT]], Iterable[List[ItemT]]]
 
@@ -111,12 +112,13 @@ def minibatch_by_padded_size(
         The `len` function is used by default.
     """
     if isinstance(size, int):
-        size_ = itertools.repeat(size)  # type: Iterator[int]
+        size_ = constant_schedule(size)
     else:
-        size_ = iter(size)
+        assert isinstance(size, Schedule)
-    for outer_batch in minibatch(seqs, size=buffer):
+        size_ = size
+    for step, outer_batch in enumerate(minibatch(seqs, size=buffer)):
         outer_batch = list(outer_batch)
-        target_size = next(size_)
+        target_size = size_(step)
         for indices in _batch_by_length(outer_batch, target_size, get_length):
             subbatch = [outer_batch[i] for i in indices]
             padded_size = max(len(seq) for seq in subbatch) * len(subbatch)
@@ -147,10 +149,12 @@ def minibatch_by_words(
         item. The `len` function is used by default.
     """
     if isinstance(size, int):
-        size_ = itertools.repeat(size)  # type: Iterator[int]
+        size_ = constant_schedule(size)
     else:
-        size_ = iter(size)
+        assert isinstance(size, Schedule)
-    target_size = next(size_)
+        size_ = size
+    step = 0
+    target_size = size_(step)
     tol_size = target_size * tolerance
     batch = []
     overflow = []
@@ -175,7 +179,8 @@ def minibatch_by_words(
         else:
             if batch:
                 yield batch
-            target_size = next(size_)
+            step += 1
+            target_size = size_(step)
             tol_size = target_size * tolerance
             batch = overflow
             batch_size = overflow_size
@@ -193,7 +198,8 @@ def minibatch_by_words(
             else:
                 if batch:
                     yield batch
-                target_size = next(size_)
+                step += 1
+                target_size = size_(step)
                 tol_size = target_size * tolerance
                 batch = [seq]
                 batch_size = n_words

spacy/training/example.pyx

@@ -443,26 +443,27 @@ def _annot2array(vocab, tok_annot, doc_annot):
         if key not in IDS:
             raise ValueError(Errors.E974.format(obj="token", key=key))
         elif key in ["ORTH", "SPACY"]:
-            pass
+            continue
         elif key == "HEAD":
             attrs.append(key)
-            values.append([h-i if h is not None else 0 for i, h in enumerate(value)])
+            row = [h-i if h is not None else 0 for i, h in enumerate(value)]
         elif key == "DEP":
             attrs.append(key)
-            values.append([vocab.strings.add(h) if h is not None else MISSING_DEP for h in value])
+            row = [vocab.strings.add(h) if h is not None else MISSING_DEP for h in value]
         elif key == "SENT_START":
             attrs.append(key)
-            values.append([to_ternary_int(v) for v in value])
+            row = [to_ternary_int(v) for v in value]
         elif key == "MORPH":
             attrs.append(key)
-            values.append([vocab.morphology.add(v) for v in value])
+            row = [vocab.morphology.add(v) for v in value]
         else:
             attrs.append(key)
             if not all(isinstance(v, str) for v in value):
                 types = set([type(v) for v in value])
                 raise TypeError(Errors.E969.format(field=key, types=types)) from None
-            values.append([vocab.strings.add(v) for v in value])
+            row = [vocab.strings.add(v) for v in value]
-    array = numpy.asarray(values, dtype="uint64")
+        values.append([numpy.array(v, dtype=numpy.int32).astype(numpy.uint64) if v < 0 else v for v in row])
+    array = numpy.array(values, dtype=numpy.uint64)
     return attrs, array.T
 
 

spacy/training/loop.py

@@ -59,6 +59,7 @@ def train(
     batcher = T["batcher"]
     train_logger = T["logger"]
     before_to_disk = create_before_to_disk_callback(T["before_to_disk"])
+    before_update = T["before_update"]
 
     # Helper function to save checkpoints. This is a closure for convenience,
     # to avoid passing in all the args all the time.
@@ -89,6 +90,7 @@ def train(
         eval_frequency=T["eval_frequency"],
         exclude=frozen_components,
         annotating_components=annotating_components,
+        before_update=before_update,
     )
     clean_output_dir(output_path)
     stdout.write(msg.info(f"Pipeline: {nlp.pipe_names}") + "\n")
@@ -150,6 +152,7 @@ def train_while_improving(
     max_steps: int,
     exclude: List[str],
     annotating_components: List[str],
+    before_update: Optional[Callable[["Language", Dict[str, Any]], None]],
 ):
     """Train until an evaluation stops improving. Works as a generator,
     with each iteration yielding a tuple `(batch, info, is_best_checkpoint)`,
@@ -198,7 +201,10 @@ def train_while_improving(
     words_seen = 0
     start_time = timer()
     for step, (epoch, batch) in enumerate(train_data):
-        dropout = next(dropouts)  # type: ignore
+        if before_update:
+            before_update_args = {"step": step, "epoch": epoch}
+            before_update(nlp, before_update_args)
+        dropout = dropouts(optimizer.step)  # type: ignore
         for subbatch in subdivide_batch(batch, accumulate_gradient):
             nlp.update(
                 subbatch,
@@ -224,6 +230,7 @@ def train_while_improving(
                     score, other_scores = evaluate()
             else:
                 score, other_scores = evaluate()
+            optimizer.last_score = score
             results.append((score, step))
             is_best_checkpoint = score == max(results)[0]
         else:

spacy/util.py

@@ -9,7 +9,7 @@ import re
 from pathlib import Path
 import thinc
 from thinc.api import NumpyOps, get_current_ops, Adam, Config, Optimizer
-from thinc.api import ConfigValidationError, Model
+from thinc.api import ConfigValidationError, Model, constant as constant_schedule
 import functools
 import itertools
 import numpy
@@ -41,13 +41,12 @@ except ImportError:
 
 from .symbols import ORTH
 from .compat import cupy, CudaStream, is_windows, importlib_metadata
-from .errors import Errors, Warnings, OLD_MODEL_SHORTCUTS
+from .errors import Errors, Warnings
 from . import about
 
 if TYPE_CHECKING:
     # This lets us add type hints for mypy etc. without causing circular imports
-    from .language import Language  # noqa: F401
+    from .language import Language, PipeCallable  # noqa: F401
-    from .pipeline import Pipe  # noqa: F401
     from .tokens import Doc, Span  # noqa: F401
     from .vocab import Vocab  # noqa: F401
 
@@ -425,8 +424,6 @@ def load_model(
             return load_model_from_path(Path(name), **kwargs)  # type: ignore[arg-type]
     elif hasattr(name, "exists"):  # Path or Path-like to model data
         return load_model_from_path(name, **kwargs)  # type: ignore[arg-type]
-    if name in OLD_MODEL_SHORTCUTS:
-        raise IOError(Errors.E941.format(name=name, full=OLD_MODEL_SHORTCUTS[name]))  # type: ignore[index]
     raise IOError(Errors.E050.format(name=name))
 
 
@@ -434,9 +431,9 @@ def load_model_from_package(
     name: str,
     *,
     vocab: Union["Vocab", bool] = True,
-    disable: Union[str, Iterable[str]] = SimpleFrozenList(),
+    disable: Union[str, Iterable[str]] = _DEFAULT_EMPTY_PIPES,
-    enable: Union[str, Iterable[str]] = SimpleFrozenList(),
+    enable: Union[str, Iterable[str]] = _DEFAULT_EMPTY_PIPES,
-    exclude: Union[str, Iterable[str]] = SimpleFrozenList(),
+    exclude: Union[str, Iterable[str]] = _DEFAULT_EMPTY_PIPES,
     config: Union[Dict[str, Any], Config] = SimpleFrozenDict(),
 ) -> "Language":
     """Load a model from an installed package.
@@ -610,9 +607,9 @@ def load_model_from_init_py(
     init_file: Union[Path, str],
     *,
     vocab: Union["Vocab", bool] = True,
-    disable: Union[str, Iterable[str]] = SimpleFrozenList(),
+    disable: Union[str, Iterable[str]] = _DEFAULT_EMPTY_PIPES,
-    enable: Union[str, Iterable[str]] = SimpleFrozenList(),
+    enable: Union[str, Iterable[str]] = _DEFAULT_EMPTY_PIPES,
-    exclude: Union[str, Iterable[str]] = SimpleFrozenList(),
+    exclude: Union[str, Iterable[str]] = _DEFAULT_EMPTY_PIPES,
     config: Union[Dict[str, Any], Config] = SimpleFrozenDict(),
 ) -> "Language":
     """Helper function to use in the `load()` method of a model package's
@@ -1582,12 +1579,12 @@ def minibatch(items, size):
     so that batch-size can vary on each step.
     """
     if isinstance(size, int):
-        size_ = itertools.repeat(size)
+        size_ = constant_schedule(size)
     else:
         size_ = size
     items = iter(items)
-    while True:
+    for step in itertools.count():
-        batch_size = next(size_)
+        batch_size = size_(step)
         batch = list(itertools.islice(items, int(batch_size)))
         if len(batch) == 0:
             break
@@ -1633,9 +1630,11 @@ def check_bool_env_var(env_var: str) -> bool:
 
 def _pipe(
     docs: Iterable["Doc"],
-    proc: "Pipe",
+    proc: "PipeCallable",
     name: str,
-    default_error_handler: Callable[[str, "Pipe", List["Doc"], Exception], NoReturn],
+    default_error_handler: Callable[
+        [str, "PipeCallable", List["Doc"], Exception], NoReturn
+    ],
     kwargs: Mapping[str, Any],
 ) -> Iterator["Doc"]:
     if hasattr(proc, "pipe"):

spacy/vectors.pyx

@@ -243,6 +243,15 @@ cdef class Vectors:
         else:
             return key in self.key2row
 
+    def __eq__(self, other):
+        # Check for equality, with faster checks first
+        return (
+                self.shape == other.shape
+                and self.key2row == other.key2row
+                and self.to_bytes(exclude=["strings"])
+                  == other.to_bytes(exclude=["strings"])
+               )
+
     def resize(self, shape, inplace=False):
         """Resize the underlying vectors array. If inplace=True, the memory
         is reallocated. This may cause other references to the data to become

spacy/vocab.pyx

@@ -467,9 +467,9 @@ cdef class Vocab:
         setters = ["strings", "vectors"]
         if "strings" not in exclude:
             self.strings.to_disk(path / "strings.json")
-        if "vectors" not in "exclude":
+        if "vectors" not in exclude:
             self.vectors.to_disk(path, exclude=["strings"])
-        if "lookups" not in "exclude":
+        if "lookups" not in exclude:
             self.lookups.to_disk(path)
 
     def from_disk(self, path, *, exclude=tuple()):

website/README.md

@@ -1,531 +1,11 @@
-<Comment>
-
 # spacy.io website and docs
 
 ![Netlify Status](https://api.netlify.com/api/v1/badges/d65fe97d-99ab-47f8-a339-1d8987251da0/deploy-status)
 
-_This page contains the documentation and styleguide for the spaCy website. Its
+The styleguide for the spaCy website is available at
-rendered version is available at https://spacy.io/styleguide._
+[spacy.io/styleguide](https://spacy.io/styleguide).
 
----
+## Setup and installation
-
-</Comment>
-
-The [spacy.io](https://spacy.io) website is implemented using
-[Gatsby](https://www.gatsbyjs.org) with
-[Remark](https://github.com/remarkjs/remark) and [MDX](https://mdxjs.com/). This
-allows authoring content in **straightforward Markdown** without the usual
-limitations. Standard elements can be overwritten with powerful
-[React](http://reactjs.org/) components and wherever Markdown syntax isn't
-enough, JSX components can be used.
-
-> #### Contributing to the site
->
-> The docs can always use another example or more detail, and they should always
-> be up to date and not misleading. We always appreciate a
-> [pull request](https://github.com/explosion/spaCy/pulls). To quickly find the
-> correct file to edit, simply click on the "Suggest edits" button at the bottom
-> of a page.
->
-> For more details on editing the site locally, see the installation
-> instructions and markdown reference below.
-
-## Logo {#logo source="website/src/images/logo.svg"}
-
-import { Logos } from 'widgets/styleguide'
-
-If you would like to use the spaCy logo on your site, please get in touch and
-ask us first. However, if you want to show support and tell others that your
-project is using spaCy, you can grab one of our
-[spaCy badges](/usage/spacy-101#faq-project-with-spacy).
-
-<Logos />
-
-## Colors {#colors}
-
-import { Colors, Patterns } from 'widgets/styleguide'
-
-<Colors />
-
-### Patterns
-
-<Patterns />
-
-## Typography {#typography}
-
-import { H1, H2, H3, H4, H5, Label, InlineList, Comment } from
-'components/typography'
-
-> #### Markdown
->
-> ```markdown_
-> ## Headline 2
-> ## Headline 2 {#some_id}
-> ## Headline 2 {#some_id tag="method"}
-> ```
->
-> #### JSX
->
-> ```jsx
-> <H2>Headline 2</H2>
-> <H2 id="some_id">Headline 2</H2>
-> <H2 id="some_id" tag="method">Headline 2</H2>
-> ```
-
-Headlines are set in
-[HK Grotesk](http://cargocollective.com/hanken/HK-Grotesk-Open-Source-Font) by
-Hanken Design. All other body text and code uses the best-matching default
-system font to provide a "native" reading experience. All code uses the
-[JetBrains Mono](https://www.jetbrains.com/lp/mono/) typeface by JetBrains.
-
-<Infobox title="Important note" variant="warning">
-
-Level 2 headings are automatically wrapped in `<section>` elements at compile
-time, using a custom
-[Markdown transformer](https://github.com/explosion/spaCy/tree/master/website/plugins/remark-wrap-section.js).
-This makes it easier to highlight the section that's currently in the viewpoint
-in the sidebar menu.
-
-</Infobox>
-
-<div>
-<H1>Headline 1</H1>
-<H2>Headline 2</H2>
-<H3>Headline 3</H3>
-<H4>Headline 4</H4>
-<H5>Headline 5</H5>
-<Label>Label</Label>
-</div>
-
----
-
-The following optional attributes can be set on the headline to modify it. For
-example, to add a tag for the documented type or mark features that have been
-introduced in a specific version or require statistical models to be loaded.
-Tags are also available as standalone `<Tag />` components.
-
-| Argument | Example                    | Result                                    |
-| -------- | -------------------------- | ----------------------------------------- |
-| `tag`    | `{tag="method"}`           | <Tag>method</Tag>                         |
-| `new`    | `{new="3"}`                | <Tag variant="new">3</Tag>                |
-| `model`  | `{model="tagger, parser"}` | <Tag variant="model">tagger, parser</Tag> |
-| `hidden` | `{hidden="true"}`          |                                           |
-
-## Elements {#elements}
-
-### Links {#links}
-
-> #### Markdown
->
-> ```markdown
-> [I am a link](https://spacy.io)
-> ```
->
-> #### JSX
->
-> ```jsx
-> <Link to="https://spacy.io">I am a link</Link>
-> ```
-
-Special link styles are used depending on the link URL.
-
-- [I am a regular external link](https://explosion.ai)
-- [I am a link to the documentation](/api/doc)
-- [I am a link to an architecture](/api/architectures#HashEmbedCNN)
-- [I am a link to a model](/models/en#en_core_web_sm)
-- [I am a link to GitHub](https://github.com/explosion/spaCy)
-
-### Abbreviations {#abbr}
-
-import { Abbr } from 'components/typography'
-
-> #### JSX
->
-> ```jsx
-> <Abbr title="Explanation">Abbreviation</Abbr>
-> ```
-
-Some text with <Abbr title="Explanation here">an abbreviation</Abbr>. On small
-screens, I collapse and the explanation text is displayed next to the
-abbreviation.
-
-### Tags {#tags}
-
-import Tag from 'components/tag'
-
-> ```jsx
-> <Tag>method</Tag>
-> <Tag variant="new">2.1</Tag>
-> <Tag variant="model">tagger, parser</Tag>
-> ```
-
-Tags can be used together with headlines, or next to properties across the
-documentation, and combined with tooltips to provide additional information. An
-optional `variant` argument can be used for special tags. `variant="new"` makes
-the tag take a version number to mark new features. Using the component,
-visibility of this tag can later be toggled once the feature isn't considered
-new anymore. Setting `variant="model"` takes a description of model capabilities
-and can be used to mark features that require a respective model to be
-installed.
-
-<InlineList>
-
-<Tag>method</Tag> <Tag variant="new">2</Tag> <Tag variant="model">tagger,
-parser</Tag>
-
-</InlineList>
-
-### Buttons {#buttons}
-
-import Button from 'components/button'
-
-> ```jsx
-> <Button to="#" variant="primary">Primary small</Button>
-> <Button to="#" variant="secondary">Secondary small</Button>
-> ```
-
-Link buttons come in two variants, `primary` and `secondary` and two sizes, with
-an optional `large` size modifier. Since they're mostly used as enhanced links,
-the buttons are implemented as styled links instead of native button elements.
-
-<InlineList><Button to="#" variant="primary">Primary small</Button>
-<Button to="#" variant="secondary">Secondary small</Button></InlineList>
-
-<br />
-
-<InlineList><Button to="#" variant="primary" large>Primary large</Button>
-<Button to="#" variant="secondary" large>Secondary large</Button></InlineList>
-
-## Components
-
-### Table {#table}
-
-> #### Markdown
->
-> ```markdown_
-> | Header 1 | Header 2 |
-> | -------- | -------- |
-> | Column 1 | Column 2 |
-> ```
->
-> #### JSX
->
-> ```markup
-> <Table>
->     <Tr><Th>Header 1</Th><Th>Header 2</Th></Tr></thead>
->     <Tr><Td>Column 1</Td><Td>Column 2</Td></Tr>
-> </Table>
-> ```
-
-Tables are used to present data and API documentation. Certain keywords can be
-used to mark a footer row with a distinct style, for example to visualize the
-return values of a documented function.
-
-| Header 1    | Header 2 | Header 3 | Header 4 |
-| ----------- | -------- | :------: | -------: |
-| Column 1    | Column 2 | Column 3 | Column 4 |
-| Column 1    | Column 2 | Column 3 | Column 4 |
-| Column 1    | Column 2 | Column 3 | Column 4 |
-| Column 1    | Column 2 | Column 3 | Column 4 |
-| **RETURNS** | Column 2 | Column 3 | Column 4 |
-
-Tables also support optional "divider" rows that are typically used to denote
-keyword-only arguments in API documentation. To turn a row into a dividing
-headline, it should only include content in its first cell, and its value should
-be italicized:
-
-> #### Markdown
->
-> ```markdown_
-> | Header 1 | Header 2 | Header 3 |
-> | -------- | -------- | -------- |
-> | Column 1 | Column 2 | Column 3 |
-> | _Hello_  |          |          |
-> | Column 1 | Column 2 | Column 3 |
-> ```
-
-| Header 1 | Header 2 | Header 3 |
-| -------- | -------- | -------- |
-| Column 1 | Column 2 | Column 3 |
-| _Hello_  |          |          |
-| Column 1 | Column 2 | Column 3 |
-
-### Type Annotations {#type-annotations}
-
-> #### Markdown
->
-> ```markdown_
-> ~~Model[List[Doc], Floats2d]~~
-> ```
->
-> #### JSX
->
-> ```markup
-> <TypeAnnotation>Model[List[Doc], Floats2d]</Typeannotation>
-> ```
-
-Type annotations are special inline code blocks are used to describe Python
-types in the [type hints](https://docs.python.org/3/library/typing.html) format.
-The special component will split the type, apply syntax highlighting and link
-all types that specify links in `meta/type-annotations.json`. Types can link to
-internal or external documentation pages. To make it easy to represent the type
-annotations in Markdown, the rendering "hijacks" the `~~` tags that would
-typically be converted to a `<del>` element – but in this case, text surrounded
-by `~~` becomes a type annotation.
-
-- ~~Dict[str, List[Union[Doc, Span]]]~~
-- ~~Model[List[Doc], List[numpy.ndarray]]~~
-
-Type annotations support a special visual style in tables and will render as a
-separate row, under the cell text. This allows the API docs to display complex
-types without taking up too much space in the cell. The type annotation should
-always be the **last element** in the row.
-
-> #### Markdown
->
-> ```markdown_
-> | Header 1 | Header 2                |
-> | -------- | ----------------------- |
-> | Column 1 | Column 2 ~~List[Doc]~~ |
-> ```
-
-| Name                    | Description                                                                                                                                                                 |
-| ----------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
-| `vocab`                 | The shared vocabulary. ~~Vocab~~                                                                                                                                            |
-| `model`                 | The Thinc [`Model`](https://thinc.ai/docs/api-model) wrapping the transformer. ~~Model[List[Doc], FullTransformerBatch]~~                                                   |
-| `set_extra_annotations` | Function that takes a batch of `Doc` objects and transformer outputs and can set additional annotations on the `Doc`. ~~Callable[[List[Doc], FullTransformerBatch], None]~~ |
-
-### List {#list}
-
-> #### Markdown
->
-> ```markdown_
-> 1. One
-> 2. Two
-> ```
->
-> #### JSX
->
-> ```markup
-> <Ol>
->     <Li>One</Li>
->     <Li>Two</Li>
-> </Ol>
-> ```
-
-Lists are available as bulleted and numbered. Markdown lists are transformed
-automatically.
-
-- I am a bulleted list
-- I have nice bullets
-- Lorem ipsum dolor
-- consectetur adipiscing elit
-
-1. I am an ordered list
-2. I have nice numbers
-3. Lorem ipsum dolor
-4. consectetur adipiscing elit
-
-### Aside {#aside}
-
-> #### Markdown
->
-> ```markdown_
-> > #### Aside title
-> > This is aside text.
-> ```
->
-> #### JSX
->
-> ```jsx
-> <Aside title="Aside title">This is aside text.</Aside>
-> ```
-
-Asides can be used to display additional notes and content in the right-hand
-column. Asides can contain text, code and other elements if needed. Visually,
-asides are moved to the side on the X-axis, and displayed at the same level they
-were inserted. On small screens, they collapse and are rendered in their
-original position, in between the text.
-
-To make them easier to use in Markdown, paragraphs formatted as blockquotes will
-turn into asides by default. Level 4 headlines (with a leading `####`) will
-become aside titles.
-
-### Code Block {#code-block}
-
-> #### Markdown
->
-> ````markdown_
-> ```python
-> ### This is a title
-> import spacy
-> ```
-> ````
->
-> #### JSX
->
-> ```jsx
-> <CodeBlock title="This is a title" lang="python">
->   import spacy
-> </CodeBlock>
-> ```
-
-Code blocks use the [Prism](http://prismjs.com/) syntax highlighter with a
-custom theme. The language can be set individually on each block, and defaults
-to raw text with no highlighting. An optional label can be added as the first
-line with the prefix `####` (Python-like) and `///` (JavaScript-like). the
-indented block as plain text and preserve whitespace.
-
-```python
-### Using spaCy
-import spacy
-nlp = spacy.load("en_core_web_sm")
-doc = nlp("This is a sentence.")
-for token in doc:
-    print(token.text, token.pos_)
-```
-
-Code blocks and also specify an optional range of line numbers to highlight by
-adding `{highlight="..."}` to the headline. Acceptable ranges are spans like
-`5-7`, but also `5-7,10` or `5-7,10,13-14`.
-
-> #### Markdown
->
-> ````markdown_
-> ```python
-> ### This is a title {highlight="1-2"}
-> import spacy
-> nlp = spacy.load("en_core_web_sm")
-> ```
-> ````
-
-```python
-### Using the matcher {highlight="5-7"}
-import spacy
-from spacy.matcher import Matcher
-
-nlp = spacy.load('en_core_web_sm')
-matcher = Matcher(nlp.vocab)
-pattern = [{"LOWER": "hello"}, {"IS_PUNCT": True}, {"LOWER": "world"}]
-matcher.add("HelloWorld", None, pattern)
-doc = nlp("Hello, world! Hello world!")
-matches = matcher(doc)
-```
-
-Adding `{executable="true"}` to the title turns the code into an executable
-block, powered by [Binder](https://mybinder.org) and
-[Juniper](https://github.com/ines/juniper). If JavaScript is disabled, the
-interactive widget defaults to a regular code block.
-
-> #### Markdown
->
-> ````markdown_
-> ```python
-> ### {executable="true"}
-> import spacy
-> nlp = spacy.load("en_core_web_sm")
-> ```
-> ````
-
-```python
-### {executable="true"}
-import spacy
-nlp = spacy.load("en_core_web_sm")
-doc = nlp("This is a sentence.")
-for token in doc:
-    print(token.text, token.pos_)
-```
-
-If a code block only contains a URL to a GitHub file, the raw file contents are
-embedded automatically and syntax highlighting is applied. The link to the
-original file is shown at the top of the widget.
-
-> #### Markdown
->
-> ````markdown_
-> ```python
-> https://github.com/...
-> ```
-> ````
->
-> #### JSX
->
-> ```jsx
-> <GitHubCode url="https://github.com/..." lang="python" />
-> ```
-
-```python
-https://github.com/explosion/spaCy/tree/master/spacy/language.py
-```
-
-### Infobox {#infobox}
-
-import Infobox from 'components/infobox'
-
-> #### JSX
->
-> ```jsx
-> <Infobox title="Information">Regular infobox</Infobox>
-> <Infobox title="Important note" variant="warning">This is a warning.</Infobox>
-> <Infobox title="Be careful!" variant="danger">This is dangerous.</Infobox>
-> ```
-
-Infoboxes can be used to add notes, updates, warnings or additional information
-to a page or section. Semantically, they're implemented and interpreted as an
-`aside` element. Infoboxes can take an optional `title` argument, as well as an
-optional `variant` (either `"warning"` or `"danger"`).
-
-<Infobox title="This is an infobox">
-
-If needed, an infobox can contain regular text, `inline code`, lists and other
-blocks.
-
-</Infobox>
-
-<Infobox title="This is a warning" variant="warning">
-
-If needed, an infobox can contain regular text, `inline code`, lists and other
-blocks.
-
-</Infobox>
-
-<Infobox title="This is dangerous" variant="danger">
-
-If needed, an infobox can contain regular text, `inline code`, lists and other
-blocks.
-
-</Infobox>
-
-### Accordion {#accordion}
-
-import Accordion from 'components/accordion'
-
-> #### JSX
->
-> ```jsx
-> <Accordion title="This is an accordion">
->   Accordion content goes here.
-> </Accordion>
-> ```
-
-Accordions are collapsible sections that are mostly used for lengthy tables,
-like the tag and label annotation schemes for different languages. They all need
-to be presented – but chances are the user doesn't actually care about _all_ of
-them, especially not at the same time. So it's fairly reasonable to hide them
-begin a click. This particular implementation was inspired by the amazing
-[Inclusive Components blog](https://inclusive-components.design/collapsible-sections/).
-
-<Accordion title="This is an accordion">
-
-Lorem ipsum dolor sit amet, consectetur adipiscing elit. Quisque enim ante,
-pretium a orci eget, varius dignissim augue. Nam eu dictum mauris, id tincidunt
-nisi. Integer commodo pellentesque tincidunt. Nam at turpis finibus tortor
-gravida sodales tincidunt sit amet est. Nullam euismod arcu in tortor auctor,
-sit amet dignissim justo congue.
-
-</Accordion>
-
-## Setup and installation {#setup}
 
 Before running the setup, make sure your versions of
 [Node](https://nodejs.org/en/) and [npm](https://www.npmjs.com/) are up to date.
@@ -554,14 +34,14 @@ extensions for your code editor. The
 [`.prettierrc`](https://github.com/explosion/spaCy/tree/master/website/.prettierrc)
 file in the root defines the settings used in this codebase.
 
-## Building & developing the site with Docker {#docker}
+## Building & developing the site with Docker
-Sometimes it's hard to get a local environment working due to rapid updates to node dependencies,
-so it may be easier to use docker for building the docs.
 
-If you'd like to do this,
+Sometimes it's hard to get a local environment working due to rapid updates to
-**be sure you do *not* include your local `node_modules` folder**,
+node dependencies, so it may be easier to use docker for building the docs.
-since there are some dependencies that need to be built for the image system. 
+
-Rename it before using.
+If you'd like to do this, **be sure you do _not_ include your local
+`node_modules` folder**, since there are some dependencies that need to be built
+for the image system. Rename it before using.
 
 ```bash
 docker run -it \
@@ -571,16 +51,16 @@ docker run -it \
   gatsby develop -H 0.0.0.0
 ```
 
-This will allow you to access the built website at http://0.0.0.0:8000/ 
+This will allow you to access the built website at http://0.0.0.0:8000/ in your
-in your browser, and still edit code in your editor while having the site 
+browser, and still edit code in your editor while having the site reflect those
-reflect those changes.
+changes.
 
-**Note**: If you're working on a Mac with an M1 processor, 
+**Note**: If you're working on a Mac with an M1 processor, you might see
-you might see segfault errors from `qemu` if you use the default image. 
+segfault errors from `qemu` if you use the default image. To fix this use the
-To fix this use the `arm64` tagged image in the `docker run` command 
+`arm64` tagged image in the `docker run` command
 (ghcr.io/explosion/spacy-io:arm64).
 
-### Building the Docker image {#docker-build}
+### Building the Docker image
 
 If you'd like to build the image locally, you can do so like this:
 
@@ -588,67 +68,21 @@ If you'd like to build the image locally, you can do so like this:
 docker build -t spacy-io .
 ```
 
-This will take some time, so if you want to use the prebuilt image you'll save a bit of time.
+This will take some time, so if you want to use the prebuilt image you'll save a
+bit of time.
 
-## Markdown reference {#markdown}
+## Project structure
-
-All page content and page meta lives in the `.md` files in the `/docs`
-directory. The frontmatter block at the top of each file defines the page title
-and other settings like the sidebar menu.
-
-````markdown
----
-title: Page title
----
-
-## Headline starting a section {#some_id}
-
-This is a regular paragraph with a [link](https://spacy.io) and **bold text**.
-
-> #### This is an aside title
->
-> This is aside text.
-
-### Subheadline
-
-| Header 1 | Header 2 |
-| -------- | -------- |
-| Column 1 | Column 2 |
-
-```python
-### Code block title {highlight="2-3"}
-import spacy
-nlp = spacy.load("en_core_web_sm")
-doc = nlp("Hello world")
-```
-
-<Infobox title="Important note" variant="warning">
-
-This is content in the infobox.
-
-</Infobox>
-````
-
-In addition to the native markdown elements, you can use the components
-[`<Infobox />`][infobox], [`<Accordion />`][accordion], [`<Abbr />`][abbr] and
-[`<Tag />`][tag] via their JSX syntax.
-
-[infobox]: https://spacy.io/styleguide#infobox
-[accordion]: https://spacy.io/styleguide#accordion
-[abbr]: https://spacy.io/styleguide#abbr
-[tag]: https://spacy.io/styleguide#tag
-
-## Project structure {#structure}
 
 ```yaml
-### Directory structure
 ├── docs                 # the actual markdown content
 ├── meta                 # JSON-formatted site metadata
 |   ├── languages.json   # supported languages and statistical models
 |   ├── sidebars.json    # sidebar navigations for different sections
 |   ├── site.json        # general site metadata
+|   ├── type-annotations.json # Type annotations
 |   └── universe.json    # data for the spaCy universe section
 ├── public               # compiled site
+├── setup                # Jinja setup
 ├── src                  # source
 |   ├── components       # React components
 |   ├── fonts            # webfonts
@@ -661,54 +95,10 @@ In addition to the native markdown elements, you can use the components
 |   |   ├── models.js    # layout template for model pages
 |   |   └── universe.js  # layout templates for universe
 |   └── widgets          # non-reusable components with content, e.g. changelog
+├── .eslintrc.json       # ESLint config file
+├── .prettierrc          # Prettier config file
 ├── gatsby-browser.js    # browser-specific hooks for Gatsby
 ├── gatsby-config.js     # Gatsby configuration
 ├── gatsby-node.js       # Node-specific hooks for Gatsby
 └── package.json         # package settings and dependencies
 ```
-
-## Editorial {#editorial}
-
-- "spaCy" should always be spelled with a lowercase "s" and a capital "C",
-  unless it specifically refers to the Python package or Python import `spacy`
-  (in which case it should be formatted as code).
-  - ✅ spaCy is a library for advanced NLP in Python.
-  - ❌ Spacy is a library for advanced NLP in Python.
-  - ✅ First, you need to install the `spacy` package from pip.
-- Mentions of code, like function names, classes, variable names etc. in inline
-  text should be formatted as `code`.
-  - ✅ "Calling the `nlp` object on a text returns a `Doc`."
-- Objects that have pages in the [API docs](/api) should be linked – for
-  example, [`Doc`](/api/doc) or [`Language.to_disk`](/api/language#to_disk). The
-  mentions should still be formatted as code within the link. Links pointing to
-  the API docs will automatically receive a little icon. However, if a paragraph
-  includes many references to the API, the links can easily get messy. In that
-  case, we typically only link the first mention of an object and not any
-  subsequent ones.
-  - ✅ The [`Span`](/api/span) and [`Token`](/api/token) objects are views of a
-    [`Doc`](/api/doc). [`Span.as_doc`](/api/span#as_doc) creates a `Doc` object
-    from a `Span`.
-  - ❌ The [`Span`](/api/span) and [`Token`](/api/token) objects are views of a
-    [`Doc`](/api/doc). [`Span.as_doc`](/api/span#as_doc) creates a
-    [`Doc`](/api/doc) object from a [`Span`](/api/span).
-
-* Other things we format as code are: references to trained pipeline packages
-  like `en_core_web_sm` or file names like `code.py` or `meta.json`.
-
-  - ✅ After training, the `config.cfg` is saved to disk.
-
-* [Type annotations](#type-annotations) are a special type of code formatting,
-  expressed by wrapping the text in `~~` instead of backticks. The result looks
-  like this: ~~List[Doc]~~. All references to known types will be linked
-  automatically.
-
-  - ✅ The model has the input type ~~List[Doc]~~ and it outputs a
-    ~~List[Array2d]~~.
-
-* We try to keep links meaningful but short.
-  - ✅ For details, see the usage guide on
-    [training with custom code](/usage/training#custom-code).
-  - ❌ For details, see
-    [the usage guide on training with custom code](/usage/training#custom-code).
-  - ❌ For details, see the usage guide on training with custom code
-    [here](/usage/training#custom-code).

website/UNIVERSE.md

@@ -51,7 +51,7 @@ markup is correct.
         "import spacy",
         "import package_name",
         "",
-        "nlp = spacy.load('en')",
+        "nlp = spacy.load('en_core_web_sm')",
         "nlp.add_pipe(package_name)"
     ],
     "code_language": "python",

website/docs/api/cli.md

@@ -12,10 +12,11 @@ menu:
   - ['train', 'train']
   - ['pretrain', 'pretrain']
   - ['evaluate', 'evaluate']
+  - ['apply', 'apply']
+  - ['find-threshold', 'find-threshold']
   - ['assemble', 'assemble']
   - ['package', 'package']
   - ['project', 'project']
-  - ['ray', 'ray']
   - ['huggingface-hub', 'huggingface-hub']
 ---
 
@@ -53,7 +54,7 @@ $ python -m spacy download [model] [--direct] [--sdist] [pip_args]
 | `--direct`, `-D`                           | Force direct download of exact package version. ~~bool (flag)~~                                                                                                                                                                           |
 | `--sdist`, `-S` <Tag variant="new">3</Tag> | Download the source package (`.tar.gz` archive) instead of the default pre-built binary wheel. ~~bool (flag)~~                                                                                                                            |
 | `--help`, `-h`                             | Show help message and available arguments. ~~bool (flag)~~                                                                                                                                                                                |
-| pip args <Tag variant="new">2.1</Tag>      | Additional installation options to be passed to `pip install` when installing the pipeline package. For example, `--user` to install to the user home directory or `--no-deps` to not install package dependencies. ~~Any (option/flag)~~ |
+| pip args                                   | Additional installation options to be passed to `pip install` when installing the pipeline package. For example, `--user` to install to the user home directory or `--no-deps` to not install package dependencies. ~~Any (option/flag)~~ |
 | **CREATES**                                | The installed pipeline package in your `site-packages` directory.                                                                                                                                                                         |
 
 ## info {#info tag="command"}
@@ -77,15 +78,15 @@ $ python -m spacy info [--markdown] [--silent] [--exclude]
 $ python -m spacy info [model] [--markdown] [--silent] [--exclude]
 ```
 
-| Name                                             | Description                                                                                                             |
+| Name                                         | Description                                                                                                             |
-| ------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------- |
+| -------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------- |
-| `model`                                          | A trained pipeline, i.e. package name or path (optional). ~~Optional[str] \(option)~~                                   |
+| `model`                                      | A trained pipeline, i.e. package name or path (optional). ~~Optional[str] \(option)~~                                   |
-| `--markdown`, `-md`                              | Print information as Markdown. ~~bool (flag)~~                                                                          |
+| `--markdown`, `-md`                          | Print information as Markdown. ~~bool (flag)~~                                                                          |
-| `--silent`, `-s` <Tag variant="new">2.0.12</Tag> | Don't print anything, just return the values. ~~bool (flag)~~                                                           |
+| `--silent`, `-s`                             | Don't print anything, just return the values. ~~bool (flag)~~                                                           |
-| `--exclude`, `-e`                                | Comma-separated keys to exclude from the print-out. Defaults to `"labels"`. ~~Optional[str]~~                           |
+| `--exclude`, `-e`                            | Comma-separated keys to exclude from the print-out. Defaults to `"labels"`. ~~Optional[str]~~                           |
-| `--url`, `-u` <Tag variant="new">3.5.0</Tag>     | Print the URL to download the most recent compatible version of the pipeline. Requires a pipeline name. ~~bool (flag)~~ |
+| `--url`, `-u` <Tag variant="new">3.5.0</Tag> | Print the URL to download the most recent compatible version of the pipeline. Requires a pipeline name. ~~bool (flag)~~ |
-| `--help`, `-h`                                   | Show help message and available arguments. ~~bool (flag)~~                                                              |
+| `--help`, `-h`                               | Show help message and available arguments. ~~bool (flag)~~                                                              |
-| **PRINTS**                                       | Information about your spaCy installation.                                                                              |
+| **PRINTS**                                   | Information about your spaCy installation.                                                                              |
 
 ## validate {#validate new="2" tag="command"}
 
@@ -260,22 +261,22 @@ chosen based on the file extension of the input file.
 $ python -m spacy convert [input_file] [output_dir] [--converter] [--file-type] [--n-sents] [--seg-sents] [--base] [--morphology] [--merge-subtokens] [--ner-map] [--lang]
 ```
 
-| Name                                             | Description                                                                                                                               |
+| Name                      | Description                                                                                                                               |
-| ------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------- |
+| ------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------- |
-| `input_path`                                     | Input file or directory. ~~Path (positional)~~                                                                                            |
+| `input_path`              | Input file or directory. ~~Path (positional)~~                                                                                            |
-| `output_dir`                                     | Output directory for converted file. Defaults to `"-"`, meaning data will be written to `stdout`. ~~Optional[Path] \(option)~~            |
+| `output_dir`              | Output directory for converted file. Defaults to `"-"`, meaning data will be written to `stdout`. ~~Optional[Path] \(option)~~            |
-| `--converter`, `-c` <Tag variant="new">2</Tag>   | Name of converter to use (see below). ~~str (option)~~                                                                                    |
+| `--converter`, `-c`       | Name of converter to use (see below). ~~str (option)~~                                                                                    |
-| `--file-type`, `-t` <Tag variant="new">2.1</Tag> | Type of file to create. Either `spacy` (default) for binary [`DocBin`](/api/docbin) data or `json` for v2.x JSON format. ~~str (option)~~ |
+| `--file-type`, `-t`       | Type of file to create. Either `spacy` (default) for binary [`DocBin`](/api/docbin) data or `json` for v2.x JSON format. ~~str (option)~~ |
-| `--n-sents`, `-n`                                | Number of sentences per document. Supported for: `conll`, `conllu`, `iob`, `ner` ~~int (option)~~                                         |
+| `--n-sents`, `-n`         | Number of sentences per document. Supported for: `conll`, `conllu`, `iob`, `ner` ~~int (option)~~                                         |
-| `--seg-sents`, `-s` <Tag variant="new">2.2</Tag> | Segment sentences. Supported for: `conll`, `ner` ~~bool (flag)~~                                                                          |
+| `--seg-sents`, `-s`       | Segment sentences. Supported for: `conll`, `ner` ~~bool (flag)~~                                                                          |
-| `--base`, `-b`, `--model`                        | Trained spaCy pipeline for sentence segmentation to use as base (for `--seg-sents`). ~~Optional[str](option)~~                            |
+| `--base`, `-b`, `--model` | Trained spaCy pipeline for sentence segmentation to use as base (for `--seg-sents`). ~~Optional[str](option)~~                            |
-| `--morphology`, `-m`                             | Enable appending morphology to tags. Supported for: `conllu` ~~bool (flag)~~                                                              |
+| `--morphology`, `-m`      | Enable appending morphology to tags. Supported for: `conllu` ~~bool (flag)~~                                                              |
-| `--merge-subtokens`, `-T`                        | Merge CoNLL-U subtokens ~~bool (flag)~~                                                                                                   |
+| `--merge-subtokens`, `-T` | Merge CoNLL-U subtokens ~~bool (flag)~~                                                                                                   |
-| `--ner-map`, `-nm`                               | NER tag mapping (as JSON-encoded dict of entity types). Supported for: `conllu` ~~Optional[Path](option)~~                                |
+| `--ner-map`, `-nm`        | NER tag mapping (as JSON-encoded dict of entity types). Supported for: `conllu` ~~Optional[Path](option)~~                                |
-| `--lang`, `-l` <Tag variant="new">2.1</Tag>      | Language code (if tokenizer required). ~~Optional[str] \(option)~~                                                                        |
+| `--lang`, `-l`            | Language code (if tokenizer required). ~~Optional[str] \(option)~~                                                                        |
-| `--concatenate`, `-C`                            | Concatenate output to a single file ~~bool (flag)~~                                                                                       |
+| `--concatenate`, `-C`     | Concatenate output to a single file ~~bool (flag)~~                                                                                       |
-| `--help`, `-h`                                   | Show help message and available arguments. ~~bool (flag)~~                                                                                |
+| `--help`, `-h`            | Show help message and available arguments. ~~bool (flag)~~                                                                                |
-| **CREATES**                                      | Binary [`DocBin`](/api/docbin) training data that can be used with [`spacy train`](/api/cli#train).                                       |
+| **CREATES**               | Binary [`DocBin`](/api/docbin) training data that can be used with [`spacy train`](/api/cli#train).                                       |
 
 ### Converters {#converters}
 
@@ -474,8 +475,7 @@ report span characteristics such as the average span length and the span (or
 span boundary) distinctiveness. The distinctiveness measure shows how different
 the tokens are with respect to the rest of the corpus using the KL-divergence of
 the token distributions. To learn more, you can check out Papay et al.'s work on
-[*Dissecting Span Identification Tasks with Performance Prediction* (EMNLP
+[_Dissecting Span Identification Tasks with Performance Prediction_ (EMNLP 2020)](https://aclanthology.org/2020.emnlp-main.396/).
-2020)](https://aclanthology.org/2020.emnlp-main.396/).
 
 </Infobox>
 
@@ -1163,6 +1163,76 @@ $ python -m spacy evaluate [model] [data_path] [--output] [--code] [--gold-prepr
 | `--help`, `-h`                            | Show help message and available arguments. ~~bool (flag)~~                                                                                                                           |
 | **CREATES**                               | Training results and optional metrics and visualizations.                                                                                                                            |
 
+## apply {#apply new="3.5" tag="command"}
+
+Applies a trained pipeline to data and stores the resulting annotated documents
+in a `DocBin`. The input can be a single file or a directory. The recognized
+input formats are:
+
+1. `.spacy`
+2. `.jsonl` containing a user specified `text_key`
+3. Files with any other extension are assumed to be plain text files containing
+   a single document.
+
+When a directory is provided it is traversed recursively to collect all files.
+
+```cli
+$ python -m spacy apply [model] [data-path] [output-file] [--code] [--text-key] [--force-overwrite] [--gpu-id] [--batch-size] [--n-process]
+```
+
+| Name                                      | Description                                                                                                                                                                          |
+| ----------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
+| `model`                                   | Pipeline to apply to the data. Can be a package or a path to a data directory. ~~str (positional)~~                                                                                  |
+| `data_path`                               | Location of data to be evaluated in spaCy's [binary format](/api/data-formats#training), jsonl, or plain text. ~~Path (positional)~~                                                 |
+| `output-file`, `-o`                       | Output `DocBin` path. ~~str (positional)~~                                                                                                                                           |
+| `--code`, `-c` <Tag variant="new">3</Tag> | Path to Python file with additional code to be imported. Allows [registering custom functions](/usage/training#custom-functions) for new architectures. ~~Optional[Path] \(option)~~ |
+| `--text-key`, `-tk`                       | The key for `.jsonl` files to use to grab the texts from. Defaults to `text`. ~~Optional[str] \(option)~~                                                                            |
+| `--force-overwrite`, `-F`                 | If the provided `output-file` already exists, then force `apply` to overwrite it. If this is `False` (default) then quits with a warning instead. ~~bool (flag)~~                    |
+| `--gpu-id`, `-g`                          | GPU to use, if any. Defaults to `-1` for CPU. ~~int (option)~~                                                                                                                       |
+| `--batch-size`, `-b`                      | Batch size to use for prediction. Defaults to `1`. ~~int (option)~~                                                                                                                  |
+| `--n-process`, `-n`                       | Number of processes to use for prediction. Defaults to `1`. ~~int (option)~~                                                                                                         |
+| `--help`, `-h`                            | Show help message and available arguments. ~~bool (flag)~~                                                                                                                           |
+| **CREATES**                               | A `DocBin` with the annotations from the `model` for all the files found in `data-path`.                                                                                             |
+
+## find-threshold {#find-threshold new="3.5" tag="command"}
+
+Runs prediction trials for a trained model with varying tresholds to maximize
+the specified metric. The search space for the threshold is traversed linearly
+from 0 to 1 in `n_trials` steps. Results are displayed in a table on `stdout`
+(the corresponding API call to `spacy.cli.find_threshold.find_threshold()`
+returns all results).
+
+This is applicable only for components whose predictions are influenced by
+thresholds - e.g. `textcat_multilabel` and `spancat`, but not `textcat`. Note
+that the full path to the corresponding threshold attribute in the config has to
+be provided.
+
+> #### Examples
+>
+> ```cli
+> # For textcat_multilabel:
+> $ python -m spacy find-threshold my_nlp data.spacy textcat_multilabel threshold cats_macro_f
+> ```
+>
+> ```cli
+> # For spancat:
+> $ python -m spacy find-threshold my_nlp data.spacy spancat threshold spans_sc_f
+> ```
+
+| Name                    | Description                                                                                                                                                                          |
+| ----------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
+| `model`                 | Pipeline to evaluate. Can be a package or a path to a data directory. ~~str (positional)~~                                                                                           |
+| `data_path`             | Path to file with DocBin with docs to use for threshold search. ~~Path (positional)~~                                                                                                |
+| `pipe_name`             | Name of pipe to examine thresholds for. ~~str (positional)~~                                                                                                                         |
+| `threshold_key`         | Key of threshold attribute in component's configuration. ~~str (positional)~~                                                                                                        |
+| `scores_key`            | Name of score to metric to optimize. ~~str (positional)~~                                                                                                                            |
+| `--n_trials`, `-n`      | Number of trials to determine optimal thresholds. ~~int (option)~~                                                                                                                   |
+| `--code`, `-c`          | Path to Python file with additional code to be imported. Allows [registering custom functions](/usage/training#custom-functions) for new architectures. ~~Optional[Path] \(option)~~ |
+| `--gpu-id`, `-g`        | GPU to use, if any. Defaults to `-1` for CPU. ~~int (option)~~                                                                                                                       |
+| `--gold-preproc`, `-G`  | Use gold preprocessing. ~~bool (flag)~~                                                                                                                                              |
+| `--silent`, `-V`, `-VV` | GPU to use, if any. Defaults to `-1` for CPU. ~~int (option)~~                                                                                                                       |
+| `--help`, `-h`          | Show help message and available arguments. ~~bool (flag)~~                                                                                                                           |
+
 ## assemble {#assemble tag="command"}
 
 Assemble a pipeline from a config file without additional training. Expects a
@@ -1229,19 +1299,19 @@ $ python -m spacy package [input_dir] [output_dir] [--code] [--meta-path] [--cre
 > $ pip install dist/en_pipeline-0.0.0.tar.gz
 > ```
 
-| Name                                             | Description                                                                                                                                                                                                                                                                      |
+| Name                                         | Description                                                                                                                                                                                                                                                                      |
-| ------------------------------------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| -------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
-| `input_dir`                                      | Path to directory containing pipeline data. ~~Path (positional)~~                                                                                                                                                                                                                |
+| `input_dir`                                  | Path to directory containing pipeline data. ~~Path (positional)~~                                                                                                                                                                                                                |
-| `output_dir`                                     | Directory to create package folder in. ~~Path (positional)~~                                                                                                                                                                                                                     |
+| `output_dir`                                 | Directory to create package folder in. ~~Path (positional)~~                                                                                                                                                                                                                     |
-| `--code`, `-c` <Tag variant="new">3</Tag>        | Comma-separated paths to Python files to be included in the package and imported in its `__init__.py`. This allows including [registering functions](/usage/training#custom-functions) and [custom components](/usage/processing-pipelines#custom-components). ~~str (option)~~  |
+| `--code`, `-c` <Tag variant="new">3</Tag>    | Comma-separated paths to Python files to be included in the package and imported in its `__init__.py`. This allows including [registering functions](/usage/training#custom-functions) and [custom components](/usage/processing-pipelines#custom-components). ~~str (option)~~  |
-| `--meta-path`, `-m` <Tag variant="new">2</Tag>   | Path to [`meta.json`](/api/data-formats#meta) file (optional). ~~Optional[Path] \(option)~~                                                                                                                                                                                      |
+| `--meta-path`, `-m`                          | Path to [`meta.json`](/api/data-formats#meta) file (optional). ~~Optional[Path] \(option)~~                                                                                                                                                                                      |
-| `--create-meta`, `-C` <Tag variant="new">2</Tag> | Create a `meta.json` file on the command line, even if one already exists in the directory. If an existing file is found, its entries will be shown as the defaults in the command line prompt. ~~bool (flag)~~                                                                  |
+| `--create-meta`, `-C`                        | Create a `meta.json` file on the command line, even if one already exists in the directory. If an existing file is found, its entries will be shown as the defaults in the command line prompt. ~~bool (flag)~~                                                                  |
-| `--build`, `-b` <Tag variant="new">3</Tag>       | Comma-separated artifact formats to build. Can be `sdist` (for a `.tar.gz` archive) and/or `wheel` (for a binary `.whl` file), or `none` if you want to run this step manually. The generated artifacts can be installed by `pip install`. Defaults to `sdist`. ~~str (option)~~ |
+| `--build`, `-b` <Tag variant="new">3</Tag>   | Comma-separated artifact formats to build. Can be `sdist` (for a `.tar.gz` archive) and/or `wheel` (for a binary `.whl` file), or `none` if you want to run this step manually. The generated artifacts can be installed by `pip install`. Defaults to `sdist`. ~~str (option)~~ |
-| `--name`, `-n` <Tag variant="new">3</Tag>        | Package name to override in meta. ~~Optional[str] \(option)~~                                                                                                                                                                                                                    |
+| `--name`, `-n` <Tag variant="new">3</Tag>    | Package name to override in meta. ~~Optional[str] \(option)~~                                                                                                                                                                                                                    |
-| `--version`, `-v` <Tag variant="new">3</Tag>     | Package version to override in meta. Useful when training new versions, as it doesn't require editing the meta template. ~~Optional[str] \(option)~~                                                                                                                             |
+| `--version`, `-v` <Tag variant="new">3</Tag> | Package version to override in meta. Useful when training new versions, as it doesn't require editing the meta template. ~~Optional[str] \(option)~~                                                                                                                             |
-| `--force`, `-f`                                  | Force overwriting of existing folder in output directory. ~~bool (flag)~~                                                                                                                                                                                                        |
+| `--force`, `-f`                              | Force overwriting of existing folder in output directory. ~~bool (flag)~~                                                                                                                                                                                                        |
-| `--help`, `-h`                                   | Show help message and available arguments. ~~bool (flag)~~                                                                                                                                                                                                                       |
+| `--help`, `-h`                               | Show help message and available arguments. ~~bool (flag)~~                                                                                                                                                                                                                       |
-| **CREATES**                                      | A Python package containing the spaCy pipeline.                                                                                                                                                                                                                                  |
+| **CREATES**                                  | A Python package containing the spaCy pipeline.                                                                                                                                                                                                                                  |
 
 ## project {#project new="3"}
 
@@ -1352,12 +1422,13 @@ If the contents are different, the new version of the file is uploaded. Deleting
 obsolete files is left up to you.
 
 Remotes can be defined in the `remotes` section of the
-[`project.yml`](/usage/projects#project-yml). Under the hood, spaCy uses the
+[`project.yml`](/usage/projects#project-yml). Under the hood, spaCy uses
-[`smart-open`](https://github.com/RaRe-Technologies/smart_open) library to
+[`Pathy`](https://github.com/justindujardin/pathy) to communicate with the
-communicate with the remote storages, so you can use any protocol that
+remote storages, so you can use any protocol that `Pathy` supports, including
-`smart-open` supports, including [S3](https://aws.amazon.com/s3/),
+[S3](https://aws.amazon.com/s3/),
-[Google Cloud Storage](https://cloud.google.com/storage), SSH and more, although
+[Google Cloud Storage](https://cloud.google.com/storage), and the local
-you may need to install extra dependencies to use certain protocols.
+filesystem, although you may need to install extra dependencies to use certain
+protocols.
 
 ```cli
 $ python -m spacy project push [remote] [project_dir]
@@ -1396,12 +1467,13 @@ outputs, so if you change the config back, you'll be able to fetch back the
 result.
 
 Remotes can be defined in the `remotes` section of the
-[`project.yml`](/usage/projects#project-yml). Under the hood, spaCy uses the
+[`project.yml`](/usage/projects#project-yml). Under the hood, spaCy uses
-[`smart-open`](https://github.com/RaRe-Technologies/smart_open) library to
+[`Pathy`](https://github.com/justindujardin/pathy) to communicate with the
-communicate with the remote storages, so you can use any protocol that
+remote storages, so you can use any protocol that `Pathy` supports, including
-`smart-open` supports, including [S3](https://aws.amazon.com/s3/),
+[S3](https://aws.amazon.com/s3/),
-[Google Cloud Storage](https://cloud.google.com/storage), SSH and more, although
+[Google Cloud Storage](https://cloud.google.com/storage), and the local
-you may need to install extra dependencies to use certain protocols.
+filesystem, although you may need to install extra dependencies to use certain
+protocols.
 
 ```cli
 $ python -m spacy project pull [remote] [project_dir]
@@ -1503,50 +1575,6 @@ $ python -m spacy project dvc [project_dir] [workflow] [--force] [--verbose] [--
 | `--help`, `-h`    | Show help message and available arguments. ~~bool (flag)~~                                                    |
 | **CREATES**       | A `dvc.yaml` file in the project directory, based on the steps defined in the given workflow.                 |
 
-## ray {#ray new="3"}
-
-The `spacy ray` CLI includes commands for parallel and distributed computing via
-[Ray](https://ray.io).
-
-<Infobox variant="warning">
-
-To use this command, you need the
-[`spacy-ray`](https://github.com/explosion/spacy-ray) package installed.
-Installing the package will automatically add the `ray` command to the spaCy
-CLI.
-
-</Infobox>
-
-### ray train {#ray-train tag="command"}
-
-Train a spaCy pipeline using [Ray](https://ray.io) for parallel training. The
-command works just like [`spacy train`](/api/cli#train). For more details and
-examples, see the usage guide on
-[parallel training](/usage/training#parallel-training) and the spaCy project
-[integration](/usage/projects#ray).
-
-```cli
-$ python -m spacy ray train [config_path] [--code] [--output] [--n-workers] [--address] [--gpu-id] [--verbose] [overrides]
-```
-
-> #### Example
->
-> ```cli
-> $ python -m spacy ray train config.cfg --n-workers 2
-> ```
-
-| Name                | Description                                                                                                                                                                                |
-| ------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
-| `config_path`       | Path to [training config](/api/data-formats#config) file containing all settings and hyperparameters. ~~Path (positional)~~                                                                |
-| `--code`, `-c`      | Path to Python file with additional code to be imported. Allows [registering custom functions](/usage/training#custom-functions) for new architectures. ~~Optional[Path] \(option)~~       |
-| `--output`, `-o`    | Directory or remote storage URL for saving trained pipeline. The directory will be created if it doesn't exist. ~~Optional[Path] \(option)~~                                               |
-| `--n-workers`, `-n` | The number of workers. Defaults to `1`. ~~int (option)~~                                                                                                                                   |
-| `--address`, `-a`   | Optional address of the Ray cluster. If not set (default), Ray will run locally. ~~Optional[str] \(option)~~                                                                               |
-| `--gpu-id`, `-g`    | GPU ID or `-1` for CPU. Defaults to `-1`. ~~int (option)~~                                                                                                                                 |
-| `--verbose`, `-V`   | Display more information for debugging purposes. ~~bool (flag)~~                                                                                                                           |
-| `--help`, `-h`      | Show help message and available arguments. ~~bool (flag)~~                                                                                                                                 |
-| overrides           | Config parameters to override. Should be options starting with `--` that correspond to the config section and value to override, e.g. `--paths.train ./train.spacy`. ~~Any (option/flag)~~ |
-
 ## huggingface-hub {#huggingface-hub new="3.1"}
 
 The `spacy huggingface-cli` CLI includes commands for uploading your trained

website/docs/api/data-formats.md

@@ -186,6 +186,7 @@ process that are used when you run [`spacy train`](/api/cli#train).
 | `accumulate_gradient`                                | Whether to divide the batch up into substeps. Defaults to `1`. ~~int~~                                                                                                                                                                                                                                                              |
 | `batcher`                                            | Callable that takes an iterator of [`Doc`](/api/doc) objects and yields batches of `Doc`s. Defaults to [`batch_by_words`](/api/top-level#batch_by_words). ~~Callable[[Iterator[Doc], Iterator[List[Doc]]]]~~                                                                                                                        |
 | `before_to_disk`                                     | Optional callback to modify `nlp` object right before it is saved to disk during and after training. Can be used to remove or reset config values or disable components. Defaults to `null`. ~~Optional[Callable[[Language], Language]]~~                                                                                           |
+| `before_update`                                      | Optional callback that is invoked at the start of each training step with the `nlp` object and a `Dict` containing the following entries: `step`, `epoch`. Can be used to make deferred changes to components. Defaults to `null`. ~~Optional[Callable[[Language, Dict[str, Any]], None]]~~                                         |
 | `dev_corpus`                                         | Dot notation of the config location defining the dev corpus. Defaults to `corpora.dev`. ~~str~~                                                                                                                                                                                                                                     |
 | `dropout`                                            | The dropout rate. Defaults to `0.1`. ~~float~~                                                                                                                                                                                                                                                                                      |
 | `eval_frequency`                                     | How often to evaluate during training (steps). Defaults to `200`. ~~int~~                                                                                                                                                                                                                                                           |

website/docs/api/dependencyparser.md

@@ -169,12 +169,6 @@ arguments it receives via the
 [`[initialize.components]`](/api/data-formats#config-initialize) block in the
 config.
 
-<Infobox variant="warning" title="Changed in v3.0" id="begin_training">
-
-This method was previously called `begin_training`.
-
-</Infobox>
-
 > #### Example
 >
 > ```python

website/docs/api/doc.md

@@ -209,15 +209,15 @@ alignment mode `"strict".
 > assert span.text == "New York"
 > ```
 
-| Name                                 | Description                                                                                                                                                                                                                                                                  |
+| Name             | Description                                                                                                                                                                                                                                                                  |
-| ------------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| ---------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
-| `start`                              | The index of the first character of the span. ~~int~~                                                                                                                                                                                                                        |
+| `start`          | The index of the first character of the span. ~~int~~                                                                                                                                                                                                                        |
-| `end`                                | The index of the last character after the span. ~~int~~                                                                                                                                                                                                                      |
+| `end`            | The index of the last character after the span. ~~int~~                                                                                                                                                                                                                      |
-| `label`                              | A label to attach to the span, e.g. for named entities. ~~Union[int, str]~~                                                                                                                                                                                                  |
+| `label`          | A label to attach to the span, e.g. for named entities. ~~Union[int, str]~~                                                                                                                                                                                                  |
-| `kb_id` <Tag variant="new">2.2</Tag> | An ID from a knowledge base to capture the meaning of a named entity. ~~Union[int, str]~~                                                                                                                                                                                    |
+| `kb_id`          | An ID from a knowledge base to capture the meaning of a named entity. ~~Union[int, str]~~                                                                                                                                                                                    |
-| `vector`                             | A meaning representation of the span. ~~numpy.ndarray[ndim=1, dtype=float32]~~                                                                                                                                                                                               |
+| `vector`         | A meaning representation of the span. ~~numpy.ndarray[ndim=1, dtype=float32]~~                                                                                                                                                                                               |
-| `alignment_mode`                     | How character indices snap to token boundaries. Options: `"strict"` (no snapping), `"contract"` (span of all tokens completely within the character span), `"expand"` (span of all tokens at least partially covered by the character span). Defaults to `"strict"`. ~~str~~ |
+| `alignment_mode` | How character indices snap to token boundaries. Options: `"strict"` (no snapping), `"contract"` (span of all tokens completely within the character span), `"expand"` (span of all tokens at least partially covered by the character span). Defaults to `"strict"`. ~~str~~ |
-| **RETURNS**                          | The newly constructed object or `None`. ~~Optional[Span]~~                                                                                                                                                                                                                   |
+| **RETURNS**      | The newly constructed object or `None`. ~~Optional[Span]~~                                                                                                                                                                                                                   |
 
 ## Doc.set_ents {#set_ents tag="method" new="3"}
 
@@ -757,11 +757,10 @@ The L2 norm of the document's vector representation.
 | `text_with_ws`                             | An alias of `Doc.text`, provided for duck-type compatibility with `Span` and `Token`. ~~str~~                                                  |
 | `mem`                                      | The document's local memory heap, for all C data it owns. ~~cymem.Pool~~                                                                       |
 | `vocab`                                    | The store of lexical types. ~~Vocab~~                                                                                                          |
-| `tensor` <Tag variant="new">2</Tag>        | Container for dense vector representations. ~~numpy.ndarray~~                                                                                  |
+| `tensor`                                   | Container for dense vector representations. ~~numpy.ndarray~~                                                                                  |
 | `user_data`                                | A generic storage area, for user custom data. ~~Dict[str, Any]~~                                                                               |
-| `lang` <Tag variant="new">2.1</Tag>        | Language of the document's vocabulary. ~~int~~                                                                                                 |
+| `lang`                                     | Language of the document's vocabulary. ~~int~~                                                                                                 |
-| `lang_` <Tag variant="new">2.1</Tag>       | Language of the document's vocabulary. ~~str~~                                                                                                 |
+| `lang_`                                    | Language of the document's vocabulary. ~~str~~                                                                                                 |
-| `sentiment`                                | The document's positivity/negativity score, if available. ~~float~~                                                                            |
 | `user_hooks`                               | A dictionary that allows customization of the `Doc`'s properties. ~~Dict[str, Callable]~~                                                      |
 | `user_token_hooks`                         | A dictionary that allows customization of properties of `Token` children. ~~Dict[str, Callable]~~                                              |
 | `user_span_hooks`                          | A dictionary that allows customization of properties of `Span` children. ~~Dict[str, Callable]~~                                               |
@@ -785,7 +784,6 @@ serialization by passing in the string names via the `exclude` argument.
 | Name               | Description                                   |
 | ------------------ | --------------------------------------------- |
 | `text`             | The value of the `Doc.text` attribute.        |
-| `sentiment`        | The value of the `Doc.sentiment` attribute.   |
 | `tensor`           | The value of the `Doc.tensor` attribute.      |
 | `user_data`        | The value of the `Doc.user_data` dictionary.  |
 | `user_data_keys`   | The keys of the `Doc.user_data` dictionary.   |

website/docs/api/entitylinker.md

@@ -200,12 +200,6 @@ knowledge base. This argument should be a function that takes a `Vocab` instance
 and creates the `KnowledgeBase`, ensuring that the strings of the knowledge base
 are synced with the current vocab.
 
-<Infobox variant="warning" title="Changed in v3.0" id="begin_training">
-
-This method was previously called `begin_training`.
-
-</Infobox>
-
 > #### Example
 >
 > ```python

website/docs/api/entityrecognizer.md

@@ -165,12 +165,6 @@ arguments it receives via the
 [`[initialize.components]`](/api/data-formats#config-initialize) block in the
 config.
 
-<Infobox variant="warning" title="Changed in v3.0" id="begin_training">
-
-This method was previously called `begin_training`.
-
-</Infobox>
-
 > #### Example
 >
 > ```python