Merge branch 'develop' into feature/component-requirements

examples/training/train_textcat.py

@@ -20,6 +20,7 @@ import spacy
 from spacy import util
 from spacy.util import minibatch, compounding
 from spacy.gold import Example
+from thinc.api import Config
 
 
 @plac.annotations(
@@ -42,8 +43,8 @@ def main(config_path, output_dir=None, n_iter=20, n_texts=2000, init_tok2vec=Non
             output_dir.mkdir()
 
     print(f"Loading nlp model from {config_path}")
-    nlp_config = util.load_config(config_path, create_objects=False)["nlp"]
-    nlp = util.load_model_from_config(nlp_config)
+    nlp_config = Config().from_disk(config_path)
+    nlp, _ = util.load_model_from_config(nlp_config, auto_fill=True)
 
     # ensure the nlp object was defined with a textcat component
     if "textcat" not in nlp.pipe_names:

examples/training/train_textcat_config.cfg

@@ -1,19 +1,14 @@
 [nlp]
 lang = "en"
+pipeline = ["textcat"]
 
-[nlp.pipeline.textcat]
+[components]
+
+[components.textcat]
 factory = "textcat"
 
-[nlp.pipeline.textcat.model]
-@architectures = "spacy.TextCatCNN.v1"
-exclusive_classes = false
-
-[nlp.pipeline.textcat.model.tok2vec]
-@architectures = "spacy.HashEmbedCNN.v1"
-pretrained_vectors = null
-width = 96
-depth = 4
-embed_size = 2000
-window_size = 1
-maxout_pieces = 3
-subword_features = true
+[components.textcat.model]
+@architectures = "spacy.TextCatBOW.v1"
+exclusive_classes = true
+ngram_size = 1
+no_output_layer = false

spacy/cli/debug_model.py

@@ -8,6 +8,7 @@ import typer
 from ._util import Arg, Opt, debug_cli, show_validation_error, parse_config_overrides
 from .. import util
 from ..lang.en import English
+from ..util import dot_to_object
 
 
 @debug_cli.command("model")
@@ -60,16 +61,7 @@ def debug_model_cli(
         msg.info(f"Fixing random seed: {seed}")
         fix_random_seed(seed)
 
-    component = config
-    parts = section.split(".")
-    for item in parts:
-        try:
-            component = component[item]
-        except KeyError:
-            msg.fail(
-                f"The section '{section}' is not a valid section in the provided config.",
-                exits=1,
-            )
+    component = dot_to_object(config, section)
     if hasattr(component, "model"):
         model = component.model
     else:

spacy/errors.py

@@ -483,7 +483,8 @@ class Errors:
     E199 = ("Unable to merge 0-length span at doc[{start}:{end}].")
 
     # TODO: fix numbering after merging develop into master
-    E952 = ("Invalid requirement specified by component '{name}': {req}")
+    E951 = ("Invalid requirement specified by component '{name}': {req}")
+    E952 = ("The section '{name}' is not a valid section in the provided config.")
     E953 = ("Mismatched IDs received by the Tok2Vec listener: {id1} vs. {id2}")
     E954 = ("The Tok2Vec listener did not receive a valid input.")
     E955 = ("Can't find table '{table}' for language '{lang}' in spacy-lookups-data.")
@@ -593,7 +594,7 @@ class Errors:
             "for the `nlp` pipeline with components {names}.")
     E993 = ("The config for 'nlp' needs to include a key 'lang' specifying "
             "the code of the language to initialize it with (for example "
-            "'en' for English).\n\n{config}")
+            "'en' for English) - this can't be 'None'.\n\n{config}")
     E996 = ("Could not parse {file}: {msg}")
     E997 = ("Tokenizer special cases are not allowed to modify the text. "
             "This would map '{chunk}' to '{orth}' given token attributes "

spacy/language.py

@@ -213,6 +213,9 @@ class Language:
         self._meta["labels"] = self.pipe_labels
         reqs = {p: self.get_pipe_meta(p).package_requirements for p in self.pipe_names}
         self._meta["requirements"].extend(util.merge_pipe_requirements(reqs))
+        # TODO: Adding this back to prevent breaking people's code etc., but
+        # we should consider removing it
+        self._meta["pipeline"] = self.pipe_names
         return self._meta
 
     @meta.setter

spacy/tests/test_util.py

@@ -2,7 +2,13 @@ import pytest
 
 from .util import get_random_doc
 
-from spacy.util import minibatch_by_words
+from spacy import util
+from spacy.util import minibatch_by_words, dot_to_object
+from thinc.api import Config, Optimizer
+
+from ..lang.en import English
+from ..lang.nl import Dutch
+from ..language import DEFAULT_CONFIG_PATH
 
 
 @pytest.mark.parametrize(
@@ -56,3 +62,49 @@ def test_util_minibatch_oversize(doc_sizes, expected_batches):
         minibatch_by_words(docs, size=batch_size, tolerance=tol, discard_oversize=False)
     )
     assert [len(batch) for batch in batches] == expected_batches
+
+
+def test_util_dot_section():
+    cfg_string = """
+    [nlp]
+    lang = "en"
+    pipeline = ["textcat"]
+    load_vocab_data = false
+
+    [components]
+
+    [components.textcat]
+    factory = "textcat"
+
+    [components.textcat.model]
+    @architectures = "spacy.TextCatBOW.v1"
+    exclusive_classes = true
+    ngram_size = 1
+    no_output_layer = false
+    """
+    nlp_config = Config().from_str(cfg_string)
+    en_nlp, en_config = util.load_model_from_config(nlp_config, auto_fill=True)
+
+    default_config = Config().from_disk(DEFAULT_CONFIG_PATH)
+    default_config["nlp"]["lang"] = "nl"
+    nl_nlp, nl_config = util.load_model_from_config(default_config, auto_fill=True)
+
+    # Test that creation went OK
+    assert isinstance(en_nlp, English)
+    assert isinstance(nl_nlp, Dutch)
+    assert nl_nlp.pipe_names == []
+    assert en_nlp.pipe_names == ["textcat"]
+    assert en_nlp.get_pipe("textcat").model.attrs["multi_label"] == False   # not exclusive_classes
+
+    # Test that default values got overwritten
+    assert not en_config["nlp"]["load_vocab_data"]
+    assert nl_config["nlp"]["load_vocab_data"]  # default value True
+
+    # Test proper functioning of 'dot_to_object'
+    with pytest.raises(KeyError):
+        obj = dot_to_object(en_config, "nlp.pipeline.tagger")
+    with pytest.raises(KeyError):
+        obj = dot_to_object(en_config, "nlp.unknownattribute")
+    assert not dot_to_object(en_config, "nlp.load_vocab_data")
+    assert dot_to_object(nl_config, "nlp.load_vocab_data")
+    assert isinstance(dot_to_object(nl_config, "training.optimizer"), Optimizer)

spacy/util.py

@@ -259,7 +259,7 @@ def load_model_from_config(
     if "nlp" not in config:
         raise ValueError(Errors.E985.format(config=config))
     nlp_config = config["nlp"]
-    if "lang" not in nlp_config:
+    if "lang" not in nlp_config or nlp_config["lang"] is None:
         raise ValueError(Errors.E993.format(config=nlp_config))
     # This will automatically handle all codes registered via the languages
     # registry, including custom subclasses provided via entry points
@@ -1136,6 +1136,25 @@ def dict_to_dot(obj: Dict[str, dict]) -> Dict[str, Any]:
     return {".".join(key): value for key, value in walk_dict(obj)}
 
 
+def dot_to_object(config: Config, section: str):
+    """Convert dot notation of a "section" to a specific part of the Config.
+    e.g. "training.optimizer" would return the Optimizer object.
+    Throws an error if the section is not defined in this config.
+
+    config (Config): The config.
+    section (str): The dot notation of the section in the config.
+    RETURNS: The object denoted by the section
+    """
+    component = config
+    parts = section.split(".")
+    for item in parts:
+        try:
+            component = component[item]
+        except (KeyError, TypeError):
+            raise KeyError(Errors.E952.format(name=section))
+    return component
+
+
 def walk_dict(
     node: Dict[str, Any], parent: List[str] = []
 ) -> Iterator[Tuple[List[str], Any]]: