make fixup

tests/pipelines/test_pipelines_table_question_answering.py

@@ -14,6 +14,8 @@
 
 import unittest
 
+from huggingface_hub import TableQuestionAnsweringOutputElement
+
 from transformers import (
     MODEL_FOR_TABLE_QUESTION_ANSWERING_MAPPING,
     AutoModelForTableQuestionAnswering,
@@ -24,6 +26,7 @@ from transformers import (
     pipeline,
 )
 from transformers.testing_utils import (
+    compare_pipeline_output_to_hub_spec,
     is_pipeline_test,
     require_pandas,
     require_tensorflow_probability,
@@ -66,6 +69,7 @@ class TQAPipelineTests(unittest.TestCase):
             },
             query="how many movies has george clooney played in?",
         )
+        compare_pipeline_output_to_hub_spec(outputs, TableQuestionAnsweringOutputElement)
         self.assertEqual(
             outputs,
             {"answer": "AVERAGE > ", "coordinates": [], "cells": [], "aggregator": "AVERAGE"},
@@ -87,6 +91,8 @@ class TQAPipelineTests(unittest.TestCase):
                 {"answer": "AVERAGE > ", "coordinates": [], "cells": [], "aggregator": "AVERAGE"},
             ],
         )
+        for output in outputs:
+            compare_pipeline_output_to_hub_spec(output, TableQuestionAnsweringOutputElement)
         outputs = table_querier(
             table={
                 "Repository": ["Transformers", "Datasets", "Tokenizers"],
@@ -113,6 +119,8 @@ class TQAPipelineTests(unittest.TestCase):
                 {"answer": "AVERAGE > ", "coordinates": [], "cells": [], "aggregator": "AVERAGE"},
             ],
         )
+        for output in outputs:
+            compare_pipeline_output_to_hub_spec(output, TableQuestionAnsweringOutputElement)
 
         with self.assertRaises(ValueError):
             table_querier(query="What does it do with empty context ?", table=None)

tests/test_pipeline_mixin.py

@@ -34,6 +34,7 @@ from huggingface_hub import (
     ImageToTextInput,
     ObjectDetectionInput,
     QuestionAnsweringInput,
+    TableQuestionAnsweringInput,
     VideoClassificationInput,
     ZeroShotImageClassificationInput,
 )
@@ -48,6 +49,7 @@ from transformers.pipelines import (
     ImageToTextPipeline,
     ObjectDetectionPipeline,
     QuestionAnsweringPipeline,
+    TableQuestionAnsweringPipeline,
     VideoClassificationPipeline,
     ZeroShotImageClassificationPipeline,
 )
@@ -136,6 +138,7 @@ task_to_pipeline_and_spec_mapping = {
     "image-to-text": (ImageToTextPipeline, ImageToTextInput),
     "object-detection": (ObjectDetectionPipeline, ObjectDetectionInput),
     "question-answering": (QuestionAnsweringPipeline, QuestionAnsweringInput),
+    "table-question-answering": (TableQuestionAnsweringPipeline, TableQuestionAnsweringInput),
     "video-classification": (VideoClassificationPipeline, VideoClassificationInput),
     "zero-shot-image-classification": (ZeroShotImageClassificationPipeline, ZeroShotImageClassificationInput),
 }