refacto: deprecate batch_size argument of Pipeline

Author: percevalw

changelog.md

@@ -39,6 +39,7 @@
 
 - `eds.span_context_getter`'s parameter `context_sents` is no longer optional and must be explicitly set to 0 to disable sentence context
 - In multi-GPU setups, streams that contain torch components are now stripped of their parameter tensors when sent to CPU Workers since these workers only perform preprocessing and postprocessing and should therefore not need the model parameters.
+- The `batch_size` argument of `Pipeline` is deprecated and is not used anymore. Use the `batch_size` argument of `stream.map_pipeline` instead.
 
 ### Fixed
 

edsnlp/core/pipeline.py

@@ -101,7 +101,7 @@ def __init__(
         lang: str,
         create_tokenizer: Optional[Callable[[Self], Tokenizer]] = None,
         vocab: Union[bool, Vocab] = True,
-        batch_size: Optional[int] = 128,
+        batch_size: Optional[int] = None,
         vocab_config: Type[BaseDefaults] = None,
         meta: Dict[str, Any] = None,
         pipeline: Optional[Sequence[str]] = None,
@@ -119,8 +119,6 @@ def __init__(
             Function that creates a tokenizer for the pipeline
         vocab: Union[bool, Vocab]
             Whether to create a new vocab or use an existing one
-        batch_size: Optional[int]
-            Batch size to use in the `.pipe()` method
         vocab_config: Type[BaseDefaults]
             Configuration for the vocab
         meta: Dict[str, Any]
@@ -129,6 +127,12 @@ def __init__(
         spacy_blank_cls = get_lang_class(lang)
 
         self.Defaults = spacy_blank_cls.Defaults
+        if batch_size is not None:
+            warnings.warn(
+                "The 'batch_size' argument is deprecated. Use the 'batch_size' "
+                "argument in `stream.map_pipeline` instead.",
+                DeprecationWarning,
+            )
         self.batch_size = batch_size
         if (vocab is not True) and (vocab_config is not None):
             raise ValueError(
@@ -397,7 +401,6 @@ def __call__(self, text: Union[str, Doc]) -> Doc:
     def pipe(
         self,
         inputs: Union[Iterable, Stream],
-        batch_size: Optional[int] = None,
         n_process: int = None,
         **kwargs,
     ) -> Stream:
@@ -409,9 +412,6 @@ def pipe(
         ----------
         inputs: Iterable[Union[str, Doc]]
             The inputs to create the Docs from, or Docs directly.
-        batch_size: Optional[int]
-            The batch size to use. If not provided, the batch size of the pipeline
-            object will be used.
         n_process: int
             Deprecated. Use the ".set(num_cpu_workers=n_process)" method on the returned
             data stream instead.
@@ -423,10 +423,6 @@ def pipe(
         Stream
         """
 
-        if batch_size is None:
-            batch_size = self.batch_size
-            kwargs = {"batch_size": batch_size, **kwargs}
-
         stream = edsnlp.data.from_iterable(inputs)
         stream = stream.map_pipeline(self, **kwargs)
         if n_process is not None: