do not allow parameters for documents.map to simplify pipeline

Author: ArneBinder

src/pytorch_ie/pipeline.py

@@ -69,7 +69,6 @@ def __init__(
             self._dataloader_params,
             self._forward_params,
             self._postprocess_params,
-            self._dataset_map_params,
         ) = self._sanitize_parameters(**kwargs)
 
     def save_pretrained(self, save_directory: str):
@@ -163,7 +162,7 @@ def _ensure_tensor_on_device(self, inputs, device):
 
     def _sanitize_parameters(
         self, **pipeline_parameters
-    ) -> Tuple[Dict[str, Any], Dict[str, Any], Dict[str, Any], Dict[str, Any], Dict[str, Any]]:
+    ) -> Tuple[Dict[str, Any], Dict[str, Any], Dict[str, Any], Dict[str, Any]]:
         """
         _sanitize_parameters will be called with any excessive named arguments from either `__init__` or `__call__`
         methods. It should return 4 dictionaries of the resolved parameters used by the various `preprocess`,
@@ -177,7 +176,6 @@ def _sanitize_parameters(
         dataloader_params = {}
         forward_parameters = {}
         postprocess_parameters: Dict[str, Any] = {}
-        dataset_map_parameters = {}
 
         # set preprocess parameters
         field = pipeline_parameters.get("predict_field")
@@ -199,16 +197,11 @@ def _sanitize_parameters(
             if p_name in pipeline_parameters:
                 postprocess_parameters[p_name] = pipeline_parameters[p_name]
 
-        for p_name in ["document_batch_size"]:
-            if p_name in pipeline_parameters:
-                dataset_map_parameters["batch_size"] = pipeline_parameters[p_name]
-
         return (
             preprocess_parameters,
             dataloader_params,
             forward_parameters,
             postprocess_parameters,
-            dataset_map_parameters,
         )
 
     def preprocess(
@@ -342,7 +335,6 @@ def __call__(
             dataloader_params,
             forward_params,
             postprocess_params,
-            dataset_map_params,
         ) = self._sanitize_parameters(**kwargs)
 
         if "TOKENIZERS_PARALLELISM" not in os.environ:
@@ -356,7 +348,6 @@ def __call__(
         dataloader_params = {**self._dataloader_params, **dataloader_params}
         forward_params = {**self._forward_params, **forward_params}
         postprocess_params = {**self._postprocess_params, **postprocess_params}
-        dataset_map_params = {**self._dataset_map_params, **dataset_map_params}
 
         self.call_count += 1
         if self.call_count > 10 and self.device.type == "cuda":
@@ -394,7 +385,6 @@ def __call__(
                         postprocess_params=postprocess_params,
                     ),
                     batched=True,
-                    **dataset_map_params,
                 )
             finally:
                 if was_caching_enabled: