Merge pull request #13282 from danieldk/maintenance/merge-main-into-develop-20240129

Merge main into `develop`

spacy/cli/benchmark_speed.py

@@ -13,7 +13,7 @@ from .. import util
 from ..language import Language
 from ..tokens import Doc
 from ..training import Corpus
-from ._util import Arg, Opt, benchmark_cli, setup_gpu
+from ._util import Arg, Opt, benchmark_cli, import_code, setup_gpu
 
 
 @benchmark_cli.command(
@@ -30,12 +30,14 @@ def benchmark_speed_cli(
     use_gpu: int = Opt(-1, "--gpu-id", "-g", help="GPU ID or -1 for CPU"),
     n_batches: int = Opt(50, "--batches", help="Minimum number of batches to benchmark", min=30,),
     warmup_epochs: int = Opt(3, "--warmup", "-w", min=0, help="Number of iterations over the data for warmup"),
+    code_path: Optional[Path] = Opt(None, "--code", "-c", help="Path to Python file with additional code (registered functions) to be imported"),
     # fmt: on
 ):
     """
     Benchmark a pipeline. Expects a loadable spaCy pipeline and benchmark
     data in the binary .spacy format.
     """
+    import_code(code_path)
     setup_gpu(use_gpu=use_gpu, silent=False)
 
     nlp = util.load_model(model)
@@ -171,5 +173,5 @@ def print_outliers(sample: numpy.ndarray):
 def warmup(
     nlp: Language, docs: List[Doc], warmup_epochs: int, batch_size: Optional[int]
 ) -> numpy.ndarray:
-    docs = warmup_epochs * docs
+    docs = [doc.copy() for doc in docs * warmup_epochs]
     return annotate(nlp, docs, batch_size)

spacy/tests/serialize/test_serialize_extension_attrs.py

@@ -15,7 +15,12 @@ def doc_w_attrs(en_tokenizer):
     Token.set_extension("_test_token", default="t0")
     doc[1]._._test_token = "t1"
 
-    return doc
+    yield doc
+
+    Doc.remove_extension("_test_attr")
+    Doc.remove_extension("_test_prop")
+    Doc.remove_extension("_test_method")
+    Token.remove_extension("_test_token")
 
 
 def test_serialize_ext_attrs_from_bytes(doc_w_attrs):

website/docs/api/cli.mdx

@@ -1268,20 +1268,21 @@ the [binary `.spacy` format](/api/data-formats#binary-training). The pipeline is
 warmed up before any measurements are taken.
 
 ```cli
-$ python -m spacy benchmark speed [model] [data_path] [--batch_size] [--no-shuffle] [--gpu-id] [--batches] [--warmup]
+$ python -m spacy benchmark speed [model] [data_path] [--code] [--batch_size] [--no-shuffle] [--gpu-id] [--batches] [--warmup]
 ```
 
-| Name                 | Description                                                                                              |
-| -------------------- | -------------------------------------------------------------------------------------------------------- |
-| `model`              | Pipeline to benchmark the speed of. Can be a package or a path to a data directory. ~~str (positional)~~ |
-| `data_path`          | Location of benchmark data in spaCy's [binary format](/api/data-formats#training). ~~Path (positional)~~ |
-| `--batch-size`, `-b` | Set the batch size. If not set, the pipeline's batch size is used. ~~Optional[int] \(option)~~           |
-| `--no-shuffle`       | Do not shuffle documents in the benchmark data. ~~bool (flag)~~                                          |
-| `--gpu-id`, `-g`     | GPU to use, if any. Defaults to `-1` for CPU. ~~int (option)~~                                           |
-| `--batches`          | Number of batches to benchmark on. Defaults to `50`. ~~Optional[int] \(option)~~                         |
-| `--warmup`, `-w`     | Iterations over the benchmark data for warmup. Defaults to `3` ~~Optional[int] \(option)~~               |
-| `--help`, `-h`       | Show help message and available arguments. ~~bool (flag)~~                                               |
-| **PRINTS**           | Pipeline speed in words per second with a 95% confidence interval.                                       |
+| Name                 | Description                                                                                                                                                                          |
+| -------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
+| `model`              | Pipeline to benchmark the speed of. Can be a package or a path to a data directory. ~~str (positional)~~                                                                             |
+| `data_path`          | Location of benchmark data in spaCy's [binary format](/api/data-formats#training). ~~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)~~ |
+| `--batch-size`, `-b` | Set the batch size. If not set, the pipeline's batch size is used. ~~Optional[int] \(option)~~                                                                                       |
+| `--no-shuffle`       | Do not shuffle documents in the benchmark data. ~~bool (flag)~~                                                                                                                      |
+| `--gpu-id`, `-g`     | GPU to use, if any. Defaults to `-1` for CPU. ~~int (option)~~                                                                                                                       |
+| `--batches`          | Number of batches to benchmark on. Defaults to `50`. ~~Optional[int] \(option)~~                                                                                                     |
+| `--warmup`, `-w`     | Iterations over the benchmark data for warmup. Defaults to `3` ~~Optional[int] \(option)~~                                                                                           |
+| `--help`, `-h`       | Show help message and available arguments. ~~bool (flag)~~                                                                                                                           |
+| **PRINTS**           | Pipeline speed in words per second with a 95% confidence interval.                                                                                                                   |
 
 ## apply {id="apply", version="3.5", tag="command"}
 
@@ -1296,6 +1297,9 @@ input formats are:
 
 When a directory is provided it is traversed recursively to collect all files.
 
+When loading a .spacy file, any potential annotations stored on the `Doc` that are not overwritten by the pipeline will be preserved.
+If you want to evaluate the pipeline on raw text only, make sure that the .spacy file does not contain any annotations.
+
 ```bash
 $ python -m spacy apply [model] [data-path] [output-file] [--code] [--text-key] [--force-overwrite] [--gpu-id] [--batch-size] [--n-process]
 ```

website/docs/api/vocab.mdx

@@ -13,7 +13,7 @@ between `Doc` objects.
 <Infobox variant ="warning">
 
 Note that a `Vocab` instance is not static. It increases in size as texts with
-new tokens are processed.
+new tokens are processed. Some models may have an empty vocab at initialization.
 
 </Infobox>
 
@@ -93,6 +93,7 @@ given string, you need to look it up in
 > #### Example
 >
 > ```python
+> nlp("I'm eating an apple")
 > apple = nlp.vocab.strings["apple"]
 > oov = nlp.vocab.strings["dskfodkfos"]
 > assert apple in nlp.vocab