Merge branch 'master' into master

photonix/classifiers/color/model.py

@@ -17,7 +17,7 @@ def __init__(self):
         self.colors = {
             # Name: ((red, green, blue), ordering)
 
-            'Red':                  ((120, 4, 20),     1),
+            'Red':                  ((120, 4, 20),      1),
             'Orange':               ((245, 133, 0),     2),
             'Amber':                ((234, 166, 30),    3),
             'Yellow':               ((240, 240, 39),    4),

photonix/classifiers/face/model.py

@@ -73,14 +73,17 @@ def load_graph(self, graph_file):
     def predict(self, image_file, min_score=0.99):
         # Detects face bounding boxes
         image = Image.open(image_file)
-
+
+        if image.mode != 'RGB':
+            image = image.convert('RGB')
+
         # Perform rotations if decalared in metadata
         metadata = PhotoMetadata(image_file)
         if metadata.get('Orientation') in ['Rotate 90 CW', 'Rotate 270 CCW']:
             image = image.rotate(-90, expand=True)
         elif metadata.get('Orientation') in ['Rotate 90 CCW', 'Rotate 270 CW']:
             image = image.rotate(90, expand=True)
-            
+
         image = np.asarray(image)
         results = self.graph['mtcnn'].detect_faces(image)
         return list(filter(lambda f: f['confidence'] > min_score, results))

photonix/classifiers/object/model.py

@@ -116,14 +116,17 @@ def format_output(self, output_dict, min_score):
 
     def predict(self, image_file, min_score=0.1):
         image = Image.open(image_file)
-
+
+        if image.mode != 'RGB':
+            image = image.convert('RGB')
+
         # Perform rotations if decalared in metadata
         metadata = PhotoMetadata(image_file)
         if metadata.get('Orientation') in ['Rotate 90 CW', 'Rotate 270 CCW']:
             image = image.rotate(-90, expand=True)
         elif metadata.get('Orientation') in ['Rotate 90 CCW', 'Rotate 270 CW']:
             image = image.rotate(90, expand=True)
-            
+
         # the array based representation of the image will be used later in order to prepare the
         # result image with boxes and labels on it.
         image_np = self.load_image_into_numpy_array(image)

photonix/classifiers/style/model.py

@@ -9,6 +9,7 @@
 
 from photonix.classifiers.base_model import BaseModel
 from photonix.photos.utils.redis import redis_connection
+from photonix.web.utils import logger
 
 
 GRAPH_FILE = os.path.join('style', 'graph.pb')
@@ -71,6 +72,10 @@ def predict(self, image_file, min_score=0.66):
             input_mean=input_mean,
             input_std=input_std)
 
+        if t is None:
+            logger.info(f'Skipping {image_file}, file format not supported by Tensorflow')
+            return None
+
         input_name = "import/" + input_layer
         output_name = "import/" + output_layer
         input_operation = self.graph.get_operation_by_name(input_name)
@@ -90,22 +95,25 @@ def predict(self, image_file, min_score=0.66):
 
     def read_tensor_from_image_file(self, file_name, input_height=299, input_width=299, input_mean=0, input_std=255):
         input_name = "file_reader"
-
-        file_reader = tf.io.read_file(file_name, input_name)
-        if file_name.endswith(".png"):
-            image_reader = tf.image.decode_png(file_reader, channels=3, name='png_reader')
-        elif file_name.endswith(".gif"):
-            image_reader = tf.squeeze(tf.image.decode_gif(file_reader, name='gif_reader'))
-        elif file_name.endswith(".bmp"):
-            image_reader = tf.image.decode_bmp(file_reader, name='bmp_reader')
-        else:
-            image_reader = tf.image.decode_jpeg(file_reader, channels=3, name='jpeg_reader')
-        float_caster = tf.cast(image_reader, tf.float32)
-        dims_expander = tf.expand_dims(float_caster, 0)
-        resized = tf.image.resize(dims_expander, [input_height, input_width], method=tf.image.ResizeMethod.BILINEAR, antialias=True)
-        normalized = tf.divide(tf.subtract(resized, [input_mean]), [input_std])
-        sess = tf.compat.v1.Session()
-        return sess.run(normalized)
+        try:
+
+            file_reader = tf.io.read_file(file_name, input_name)
+            if file_name.endswith(".png"):
+                image_reader = tf.image.decode_png(file_reader, channels=3, name='png_reader')
+            elif file_name.endswith(".gif"):
+                image_reader = tf.squeeze(tf.image.decode_gif(file_reader, name='gif_reader'))
+            elif file_name.endswith(".bmp"):
+                image_reader = tf.image.decode_bmp(file_reader, name='bmp_reader')
+            else:
+                image_reader = tf.image.decode_jpeg(file_reader, channels=3, name='jpeg_reader')
+            float_caster = tf.cast(image_reader, tf.float32)
+            dims_expander = tf.expand_dims(float_caster, 0)
+            resized = tf.image.resize(dims_expander, [input_height, input_width], method=tf.image.ResizeMethod.BILINEAR, antialias=True)
+            normalized = tf.divide(tf.subtract(resized, [input_mean]), [input_std])
+            sess = tf.compat.v1.Session()
+            return sess.run(normalized)
+        except:
+            return None
 
 
 def run_on_photo(photo_id):
@@ -114,7 +122,7 @@ def run_on_photo(photo_id):
     from photonix.classifiers.runners import results_for_model_on_photo, get_or_create_tag
     photo, results = results_for_model_on_photo(model, photo_id)
 
-    if photo:
+    if photo and results is not None:
         from photonix.photos.models import PhotoTag
         photo.clear_tags(source='C', type='S')
         for name, score in results:
@@ -132,5 +140,8 @@ def run_on_photo(photo_id):
 
     results = model.predict(sys.argv[1], min_score=0.01)
 
-    for label, score in results:
-        print('{} (score: {:0.5f})'.format(label, score))
+    if results is None:
+        print(f'{sys.argv[1]} could not be processed by style classifier')
+    else:
+        for label, score in results:
+            print('{} (score: {:0.5f})'.format(label, score))

photonix/photos/utils/redis.py

@@ -6,5 +6,6 @@
 redis_connection = redis.Redis(
     host=os.environ.get('REDIS_HOST', '127.0.0.1'),
     port=int(os.environ.get('REDIS_PORT', '6379')),
-    db=int(os.environ.get('REDIS_DB', '0'))
+    db=int(os.environ.get('REDIS_DB', '0')),
+    password=os.environ.get('REDIS_PASSWORD')
 )

tests/test_classifier_models.py

@@ -108,6 +108,11 @@ def test_style_predict():
     assert result[0][0] == 'serene'
     assert '{0:.3f}'.format(result[0][1]) == '0.962'
 
+    # Check that there is no error when running with non-RGB image
+    cmyk = str(Path(__file__).parent / 'photos' / 'cmyk.tif')
+    result = model.predict(cmyk)
+    assert result == None
+
 
 def test_face_predict():
     from photonix.classifiers.face.model import FaceModel