Fixing NER pipeline for list inputs. (#10184)

Fixes #10168

src/transformers/pipelines/token_classification.py

@@ -28,11 +28,14 @@ class TokenClassificationArgumentHandler(ArgumentHandler):
     Handles arguments for token classification.
     """
 
-    def __call__(self, *args, **kwargs):
+    def __call__(self, inputs: Union[str, List[str]], **kwargs):
 
-        if args is not None and len(args) > 0:
+        if inputs is not None and isinstance(inputs, (list, tuple)) and len(inputs) > 0:
-            inputs = list(args)
+            inputs = list(inputs)
             batch_size = len(inputs)
+        elif isinstance(inputs, str):
+            inputs = [inputs]
+            batch_size = 1
         else:
             raise ValueError("At least one input is required.")
 
@@ -137,11 +140,11 @@ class TokenClassificationPipeline(Pipeline):
               Only exists if the offsets are available within the tokenizer
         """
 
-        inputs, offset_mappings = self._args_parser(inputs, **kwargs)
+        _inputs, offset_mappings = self._args_parser(inputs, **kwargs)
 
         answers = []
 
-        for i, sentence in enumerate(inputs):
+        for i, sentence in enumerate(_inputs):
 
             # Manage correct placement of the tensors
             with self.device_placement():

tests/test_pipelines_ner.py

@@ -14,14 +14,17 @@
 
 import unittest
 
-from transformers import AutoTokenizer, pipeline
+from transformers import AutoTokenizer, is_torch_available, pipeline
 from transformers.pipelines import Pipeline, TokenClassificationArgumentHandler
 from transformers.testing_utils import require_tf, require_torch, slow
 
 from .test_pipelines_common import CustomInputPipelineCommonMixin
 
 
-VALID_INPUTS = ["A simple string", ["list of strings"]]
+if is_torch_available():
+    import numpy as np
+
+VALID_INPUTS = ["A simple string", ["list of strings", "A simple string that is quite a bit longer"]]
 
 
 class NerPipelineTests(CustomInputPipelineCommonMixin, unittest.TestCase):
@@ -334,17 +337,26 @@ class NerPipelineTests(CustomInputPipelineCommonMixin, unittest.TestCase):
     @require_torch
     def test_simple(self):
         nlp = pipeline(task="ner", model="dslim/bert-base-NER", grouped_entities=True)
-        output = nlp("Hello Sarah Jessica Parker who Jessica lives in New York")
+        sentence = "Hello Sarah Jessica Parker who Jessica lives in New York"
+        sentence2 = "This is a simple test"
+        output = nlp(sentence)
 
         def simplify(output):
-            for i in range(len(output)):
+            if isinstance(output, (list, tuple)):
-                output[i]["score"] = round(output[i]["score"], 3)
+                return [simplify(item) for item in output]
+            elif isinstance(output, dict):
+                return {simplify(k): simplify(v) for k, v in output.items()}
+            elif isinstance(output, (str, int, np.int64)):
                 return output
+            elif isinstance(output, float):
+                return round(output, 3)
+            else:
+                raise Exception(f"Cannot handle {type(output)}")
 
-        output = simplify(output)
+        output_ = simplify(output)
 
         self.assertEqual(
-            output,
+            output_,
             [
                 {
                     "entity_group": "PER",
@@ -358,6 +370,21 @@ class NerPipelineTests(CustomInputPipelineCommonMixin, unittest.TestCase):
             ],
         )
 
+        output = nlp([sentence, sentence2])
+        output_ = simplify(output)
+
+        self.assertEqual(
+            output_,
+            [
+                [
+                    {"entity_group": "PER", "score": 0.996, "word": "Sarah Jessica Parker", "start": 6, "end": 26},
+                    {"entity_group": "PER", "score": 0.977, "word": "Jessica", "start": 31, "end": 38},
+                    {"entity_group": "LOC", "score": 0.999, "word": "New York", "start": 48, "end": 56},
+                ],
+                [],
+            ],
+        )
+
     @require_torch
     def test_pt_small_ignore_subwords_available_for_fast_tokenizers(self):
         for model_name in self.small_models:
@@ -386,7 +413,7 @@ class TokenClassificationArgumentHandlerTestCase(unittest.TestCase):
         self.assertEqual(inputs, [string])
         self.assertEqual(offset_mapping, None)
 
-        inputs, offset_mapping = self.args_parser(string, string)
+        inputs, offset_mapping = self.args_parser([string, string])
         self.assertEqual(inputs, [string, string])
         self.assertEqual(offset_mapping, None)
 
@@ -394,25 +421,35 @@ class TokenClassificationArgumentHandlerTestCase(unittest.TestCase):
         self.assertEqual(inputs, [string])
         self.assertEqual(offset_mapping, [[(0, 1), (1, 2)]])
 
-        inputs, offset_mapping = self.args_parser(string, string, offset_mapping=[[(0, 1), (1, 2)], [(0, 2), (2, 3)]])
+        inputs, offset_mapping = self.args_parser(
+            [string, string], offset_mapping=[[(0, 1), (1, 2)], [(0, 2), (2, 3)]]
+        )
         self.assertEqual(inputs, [string, string])
         self.assertEqual(offset_mapping, [[(0, 1), (1, 2)], [(0, 2), (2, 3)]])
 
     def test_errors(self):
         string = "This is a simple input"
 
-        # 2 sentences, 1 offset_mapping
+        # 2 sentences, 1 offset_mapping, args
-        with self.assertRaises(ValueError):
+        with self.assertRaises(TypeError):
             self.args_parser(string, string, offset_mapping=[[(0, 1), (1, 2)]])
 
-        # 2 sentences, 1 offset_mapping
+        # 2 sentences, 1 offset_mapping, args
-        with self.assertRaises(ValueError):
+        with self.assertRaises(TypeError):
             self.args_parser(string, string, offset_mapping=[(0, 1), (1, 2)])
 
+        # 2 sentences, 1 offset_mapping, input_list
+        with self.assertRaises(ValueError):
+            self.args_parser([string, string], offset_mapping=[[(0, 1), (1, 2)]])
+
+        # 2 sentences, 1 offset_mapping, input_list
+        with self.assertRaises(ValueError):
+            self.args_parser([string, string], offset_mapping=[(0, 1), (1, 2)])
+
         # 1 sentences, 2 offset_mapping
         with self.assertRaises(ValueError):
             self.args_parser(string, offset_mapping=[[(0, 1), (1, 2)], [(0, 2), (2, 3)]])
 
         # 0 sentences, 1 offset_mapping
-        with self.assertRaises(ValueError):
+        with self.assertRaises(TypeError):
             self.args_parser(offset_mapping=[[(0, 1), (1, 2)]])