fixed clarifai tests, everything importing

pliers/extractors/__init__.py

@@ -6,12 +6,17 @@
 from .base import Extractor, ExtractorResult, merge_results
 from .api import (IndicoAPITextExtractor,
                   IndicoAPIImageExtractor,
-                  ClarifaiAPIExtractor,
+                  ClarifaiAPIImageExtractor,
+                  ClarifaiAPIVideoExtractor,
                   GoogleVisionAPIFaceExtractor,
                   GoogleVisionAPILabelExtractor,
                   GoogleVisionAPIPropertyExtractor,
                   GoogleVisionAPISafeSearchExtractor,
                   GoogleVisionAPIWebEntitiesExtractor,
+                  GoogleVideoIntelligenceAPIExtractor,
+                  GoogleVideoAPILabelDetectionExtractor,
+                  GoogleVideoAPIShotDetectionExtractor,
+                  GoogleVideoAPIExplicitDetectionExtractor,
                   MicrosoftAPIFaceExtractor,
                   MicrosoftAPIFaceEmotionExtractor,
                   MicrosoftVisionAPIExtractor,
@@ -54,7 +59,8 @@
     'ExtractorResult',
     'IndicoAPITextExtractor',
     'IndicoAPIImageExtractor',
-    'ClarifaiAPIExtractor',
+    'ClarifaiAPIImageExtractor',
+    'ClarifaiAPIVideoExtractor',
     'STFTAudioExtractor',
     'MeanAmplitudeExtractor',
     'LibrosaFeatureExtractor',
@@ -77,6 +83,10 @@
     'GoogleVisionAPIPropertyExtractor',
     'GoogleVisionAPISafeSearchExtractor',
     'GoogleVisionAPIWebEntitiesExtractor',
+    'GoogleVideoIntelligenceAPIExtractor',
+    'GoogleVideoAPILabelDetectionExtractor',
+    'GoogleVideoAPIShotDetectionExtractor',
+    'GoogleVideoAPIExplicitDetectionExtractor',
     'BrightnessExtractor',
     'SaliencyExtractor',
     'SharpnessExtractor',

pliers/extractors/api/__init__.py

@@ -1,11 +1,16 @@
-from .clarifai import ClarifaiAPIExtractor
+from .clarifai import (ClarifaiAPIImageExtractor,
+                       ClarifaiAPIVideoExtractor)
 from .indico import (IndicoAPITextExtractor,
                      IndicoAPIImageExtractor)
 from .google import (GoogleVisionAPIFaceExtractor,
                      GoogleVisionAPILabelExtractor,
                      GoogleVisionAPIPropertyExtractor,
                      GoogleVisionAPISafeSearchExtractor,
-                     GoogleVisionAPIWebEntitiesExtractor)
+                     GoogleVisionAPIWebEntitiesExtractor,
+                     GoogleVideoIntelligenceAPIExtractor,
+                     GoogleVideoAPILabelDetectionExtractor,
+                     GoogleVideoAPIShotDetectionExtractor,
+                     GoogleVideoAPIExplicitDetectionExtractor)
 from .microsoft import (MicrosoftAPIFaceExtractor,
                         MicrosoftAPIFaceEmotionExtractor,
                         MicrosoftVisionAPIExtractor,
@@ -16,14 +21,19 @@
                         MicrosoftVisionAPIAdultExtractor)
 
 __all__ = [
-    'ClarifaiAPIExtractor',
+    'ClarifaiAPIImageExtractor',
+    'ClarifaiAPIVideoExtractor',
     'IndicoAPITextExtractor',
     'IndicoAPIImageExtractor',
     'GoogleVisionAPIFaceExtractor',
     'GoogleVisionAPILabelExtractor',
     'GoogleVisionAPIPropertyExtractor',
     'GoogleVisionAPISafeSearchExtractor',
     'GoogleVisionAPIWebEntitiesExtractor',
+    'GoogleVideoIntelligenceAPIExtractor',
+    'GoogleVideoAPILabelDetectionExtractor',
+    'GoogleVideoAPIShotDetectionExtractor',
+    'GoogleVideoAPIExplicitDetectionExtractor',
     'MicrosoftAPIFaceExtractor',
     'MicrosoftAPIFaceEmotionExtractor',
     'MicrosoftVisionAPIExtractor',

pliers/extractors/api/clarifai.py

@@ -26,7 +26,7 @@
 
 class ClarifaiAPIExtractor(APITransformer):
 
-    ''' Uses the Clarifai API to extract tags of images.
+    ''' Uses the Clarifai API to extract tags of visual stimuli.
 
     Args:
         api_key (str): A valid API_KEY for the Clarifai API. Only needs to be
@@ -41,7 +41,6 @@ class ClarifaiAPIExtractor(APITransformer):
             API. For example, ['food', 'animal'].
         rate_limit (int): The minimum number of seconds required between
             transform calls on this Transformer.
-        batch_size (int): Number of stims to send per batched API request.
     '''
 
     _log_attributes = ('api_key', 'model', 'model_name', 'min_value',
@@ -76,8 +75,7 @@ def __init__(self, api_key=None, model='general-v1.3', min_value=None,
             select_concepts = listify(select_concepts)
             self.select_concepts = [clarifai_client.Concept(concept_name=n)
                                     for n in select_concepts]
-        super(ClarifaiAPIExtractor, self).__init__(rate_limit=rate_limit,
-                                                   batch_size=batch_size)
+        super(ClarifaiAPIExtractor, self).__init__(rate_limit=rate_limit)
 
     @property
     def api_keys(self):
@@ -105,8 +103,38 @@ def _parse_annotations(self, annotation):
 class ClarifaiAPIImageExtractor(ClarifaiAPIExtractor, BatchTransformerMixin,
                                 ImageExtractor):
 
+    ''' Uses the Clarifai API to extract tags of images.
+
+    Args:
+        api_key (str): A valid API_KEY for the Clarifai API. Only needs to be
+            passed the first time the extractor is initialized.
+        model (str): The name of the Clarifai model to use. If None, defaults
+            to the general image tagger.
+        min_value (float): A value between 0.0 and 1.0 indicating the minimum
+            confidence required to return a prediction. Defaults to 0.0.
+        max_concepts (int): A value between 0 and 200 indicating the maximum
+            number of label predictions returned.
+        select_concepts (list): List of concepts (strings) to query from the
+            API. For example, ['food', 'animal'].
+        rate_limit (int): The minimum number of seconds required between
+            transform calls on this Transformer.
+        batch_size (int): Number of stims to send per batched API request.
+    '''
+
     _batch_size = 32
 
+    def __init__(self, api_key=None, model='general-v1.3', min_value=None,
+                 max_concepts=None, select_concepts=None, rate_limit=None,
+                 batch_size=None):
+        super(ClarifaiAPIImageExtractor,
+              self).__init__(api_key=api_key,
+                             model=model,
+                             min_value=min_value,
+                             max_concepts=max_concepts,
+                             select_concepts=select_concepts,
+                             rate_limit=rate_limit,
+                             batch_size=batch_size)
+
     def _extract(self, stims):
         verify_dependencies(['clarifai_client'])
 
@@ -140,17 +168,23 @@ def _extract(self, stim):
         return ExtractorResult(outputs, stim, self)
 
     def _to_df(self, result):
-        rows = []
-        for frame_res in result._data[0]['data']['frames']:
-            data_dict = self._parse_annotations(frame_res)
-            data_dict['onset_'] = frame_res['frame_info']['time'] / 1000.0
-            frame_num = frame_res['frame_info']['index']
-            if frame_num == 0:
-                est_duration = 0.0
+        onsets = []
+        durations = []
+        data = []
+        frames = result._data[0]['data']['frames']
+        for i, frame_res in enumerate(frames):
+            data.append(self._parse_annotations(frame_res))
+            onset = frame_res['frame_info']['time'] / 1000.0
+            if (i + 1) == len(frames):
+                end = result.stim.duration
+                print(end)
             else:
-                est_duration = data_dict['onset_'] / float(frame_num)
-            data_dict['duration_'] = min(est_duration,
-                                         result.stim.duration - data_dict['onset_'])
-            data_dict['order_'] = frame_num
-            rows.append(data_dict)
-        return pd.DataFrame(rows)
+                end = frames[i+1]['frame_info']['time'] / 1000.0
+            onsets.append(onset)
+            durations.append(end - onset)
+
+        result._onsets = onsets
+        result._durations = durations
+        df = pd.DataFrame(data)
+        result.features = list(df.columns)
+        return df

pliers/extractors/api/google.py

@@ -243,7 +243,10 @@ def _parse_label(self, features, label):
         for segment in label['segments']:
             onset, duration = self._get_onset_duration(segment['segment'])
             score = segment['confidence']
-            data.extend(self._enumerate_features(features, onset, duration, score))
+            data.extend(self._enumerate_features(features,
+                                                 onset,
+                                                 duration,
+                                                 score))
         return data
 
     def _parse_frame(self, features, annotation, score_key, max_duration):
@@ -257,7 +260,10 @@ def _parse_frame(self, features, annotation, score_key, max_duration):
                 end = float(frames[i+1]['timeOffset'][:-1])
             duration = end - onset
             score = frame[score_key]
-            data.extend(self._enumerate_features(features, onset, duration, score))
+            data.extend(self._enumerate_features(features,
+                                                 onset,
+                                                 duration,
+                                                 score))
         return data
 
     def _to_df(self, result):
@@ -272,8 +278,12 @@ def _to_df(self, result):
                         for category in annot.get('categoryEntities', []):
                             feats.append('category_' + category['description'])
                         if key == 'frameLabelAnnotations':
-                            data.extend(self._parse_frame(feats, annot, 'confidence', duration))
+                            data.extend(self._parse_frame(feats,
+                                                          annot,
+                                                          'confidence',
+                                                          duration))
                         else:
+                            # Good for shot or segment labels
                             data.extend(self._parse_label(feats, annot))
                 elif key == 'shotAnnotations':
                     for shot in res:
@@ -285,7 +295,10 @@ def _to_df(self, result):
                         })
                 elif key == 'explicitAnnotation':
                     feature = 'pornographyLikelihood'
-                    data.extend(self._parse_frame([feature], res, feature, duration))
+                    data.extend(self._parse_frame([feature],
+                                                  res,
+                                                  feature,
+                                                  duration))
 
         df = pd.DataFrame(data)
         result._onsets = df['onset']

pliers/tests/extractors/api/test_clarifai_extractors.py

@@ -1,56 +1,58 @@
 from os.path import join
 from ...utils import get_test_data_path
 from pliers import config
-from pliers.extractors import ClarifaiAPIExtractor
+from pliers.extractors import (ClarifaiAPIImageExtractor,
+                               ClarifaiAPIVideoExtractor)
 from pliers.extractors.base import merge_results
-from pliers.stimuli import ImageStim
+from pliers.stimuli import ImageStim, VideoStim
 import numpy as np
 import pytest
 
 IMAGE_DIR = join(get_test_data_path(), 'image')
+VIDEO_DIR = join(get_test_data_path(), 'video')
 
 
 @pytest.mark.skipif("'CLARIFAI_API_KEY' not in os.environ")
 def test_clarifai_api_extractor():
     stim = ImageStim(join(IMAGE_DIR, 'apple.jpg'))
-    ext = ClarifaiAPIExtractor()
+    ext = ClarifaiAPIImageExtractor()
     assert ext.validate_keys()
     result = ext.transform(stim).to_df()
     assert result['apple'][0] > 0.5
     assert result.ix[:, 5][0] > 0.0
 
-    result = ClarifaiAPIExtractor(max_concepts=5).transform(stim).to_df()
+    result = ClarifaiAPIImageExtractor(max_concepts=5).transform(stim).to_df()
     assert result.shape == (1, 9)
 
-    result = ClarifaiAPIExtractor(
+    result = ClarifaiAPIImageExtractor(
         min_value=0.9).transform(stim).to_df(object_id=False)
     assert all(np.isnan(d) or d > 0.9 for d in result.values[0, 3:])
 
     concepts = ['cat', 'dog']
-    result = ClarifaiAPIExtractor(select_concepts=concepts).transform(stim)
+    result = ClarifaiAPIImageExtractor(select_concepts=concepts).transform(stim)
     result = result.to_df()
     assert result.shape == (1, 6)
     assert 'cat' in result.columns and 'dog' in result.columns
 
     url = 'https://tuition.utexas.edu/sites/all/themes/tuition/logo.png'
     stim = ImageStim(url=url)
-    result = ClarifaiAPIExtractor(max_concepts=5).transform(stim).to_df()
+    result = ClarifaiAPIImageExtractor(max_concepts=5).transform(stim).to_df()
     assert result.shape == (1, 9)
     assert result['symbol'][0] > 0.8
 
-    ext = ClarifaiAPIExtractor(api_key='nogood')
+    ext = ClarifaiAPIImageExtractor(api_key='nogood')
     assert not ext.validate_keys()
 
 
 @pytest.mark.skipif("'CLARIFAI_API_KEY' not in os.environ")
 def test_clarifai_api_extractor_batch():
     stim = ImageStim(join(IMAGE_DIR, 'apple.jpg'))
     stim2 = ImageStim(join(IMAGE_DIR, 'obama.jpg'))
-    ext = ClarifaiAPIExtractor()
+    ext = ClarifaiAPIImageExtractor()
     results = ext.transform([stim, stim2])
     results = merge_results(results)
-    assert results['ClarifaiAPIExtractor#apple'][0] > 0.5 or \
-        results['ClarifaiAPIExtractor#apple'][1] > 0.5
+    assert results['ClarifaiAPIImageExtractor#apple'][0] > 0.5 or \
+        results['ClarifaiAPIImageExtractor#apple'][1] > 0.5
 
 
 @pytest.mark.skipif("'CLARIFAI_API_KEY' not in os.environ")
@@ -60,16 +62,29 @@ def test_clarifai_api_extractor_large():
     config.set_option('allow_large_jobs', False)
     config.set_option('large_job', 1)
 
-    ext = ClarifaiAPIExtractor()
+    ext = ClarifaiAPIImageExtractor()
     images = [ImageStim(join(IMAGE_DIR, 'apple.jpg')),
               ImageStim(join(IMAGE_DIR, 'obama.jpg'))]
     with pytest.raises(ValueError):
         merge_results(ext.transform(images))
 
     config.set_option('allow_large_jobs', True)
     results = merge_results(ext.transform(images))
-    assert 'ClarifaiAPIExtractor#apple' in results.columns
+    assert 'ClarifaiAPIImageExtractor#apple' in results.columns
     assert results.shape == (2, 49)
 
     config.set_option('allow_large_jobs', default)
     config.set_option('large_job', default_large)
+
+
+@pytest.mark.skipif("'CLARIFAI_API_KEY' not in os.environ")
+def test_clarifai_api_video_extractor():
+    stim = VideoStim(join(VIDEO_DIR, 'small.mp4'))
+    ext = ClarifaiAPIVideoExtractor()
+    assert ext.validate_keys()
+    result = ext.transform(stim).to_df()
+    assert result.shape == (6, 27)
+    assert result['toy'][0] > 0.5
+    assert result['onset'][1] == 1.0
+    assert result['duration'][0] == 1.0
+    assert np.isclose(result['duration'][5], 0.57)