Update app.py

app.py

@@ -166,7 +166,7 @@ def get_imgs_from_folder(image_files, zipfile):
   return image_files
 
 
-def get_seo_tags(image_path, topical_map, new_imgs, attempts=0, max_attempts=10):
+def get_seo_tags(image_path, topical_map, new_imgs, attempts=0, max_attempts=6):
   '''
   Gets the seo tags and topic/sub-topic classification for an image using OpenAI GPT-4 Vision Preview
   Input: image path of desired file
@@ -194,7 +194,12 @@ def get_seo_tags(image_path, topical_map, new_imgs, attempts=0, max_attempts=10)
   topic_list.insert(0, "irrelevant")
   
   def encode_image(image_path):
-    # Check the file extension to determine if it is a HEIC file
+    # Check the file size (in bytes)
+    file_size = os.path.getsize(image_path)
+    
+    # Define the maximum file size for compression (20 MB)
+    max_size = 20 * 1024 * 1024  # 20 MB in bytes
+    
     if image_path.lower().endswith('.heic'):
         # Read the HEIC file
         heif_file = pyheif.read(image_path)
@@ -207,17 +212,23 @@ def get_seo_tags(image_path, topical_map, new_imgs, attempts=0, max_attempts=10)
             heif_file.mode,
             heif_file.stride,
         )
-        # Convert the image to JPEG format in memory
-        with BytesIO() as img_buffer:
-            image.save(img_buffer, format='JPEG')
-            # Seek to the beginning of the stream
-            img_buffer.seek(0)
-            # Read the JPEG image data and encode it in base64
-            return base64.b64encode(img_buffer.read()).decode('utf-8')
     else:
-        # Handle other image types directly
-        with open(image_path, "rb") as image_file:
-            return base64.b64encode(image_file.read()).decode('utf-8')
+        # Open other image types with PIL directly
+        image = Image.open(image_path)
+
+    # Use in-memory buffer for processing
+    with BytesIO() as img_buffer:
+        if file_size > max_size:
+            # Adjust the quality to reduce the file size
+            image.save(img_buffer, format='JPEG', quality=75)
+        else:
+            # Save the image without changing the quality if not needed
+            image.save(img_buffer, format='JPEG', quality=85)
+        
+        # Seek to the beginning of the stream
+        img_buffer.seek(0)
+        # Read the JPEG image data and encode it in base64
+        return base64.b64encode(img_buffer.read()).decode('utf-8')
 
   print(image_path)
   base64_image = encode_image(image_path)
@@ -773,7 +784,7 @@ def get_collection_dict():
 
   return collection_dict
 
-def import_client_data(client_name, zipfile, topical_map, password, project_bool, ai_bool, done_num, progress=gr.Progress(), create=False):
+def import_client_data(client_name, zipfile, topical_map, password, project_bool, ai_bool, progress=gr.Progress(), create=False):
   '''
   Takes the client neame and the client zipfile path to import all image files in the google drive into brandfolder under a collection
   with the client's name
@@ -832,9 +843,8 @@ def import_client_data(client_name, zipfile, topical_map, password, project_bool
       img_lists.append(unpack_list)
   img_list = sum(img_lists, [])
   new_imgs = []
-
+  error_imgs = []
   # iterates all images and puts them into brandfolder with AI elements
-  img_list = img_list[done_num:]
   for img in progress.tqdm(img_list, desc="Uploading..."):
     zip = img_dict[img]
     img = zip.extract(img)
@@ -847,9 +857,12 @@ def import_client_data(client_name, zipfile, topical_map, password, project_bool
         elif ai_bool.lower() == 'off':
             create_asset_no_ai(client_name, collection_id, img, project_bool=project_bool)
     except Exception as e:
+        error_imgs.append(f'{str(img)}; error: {e}\n')
         print(f'An unexpected error occured processing {img}: {e}')
   gr.Info('Images have been uploaded!')
-  return "Images Uploaded"
+  if error_imgs:
+      error_imgs_text = '\n'.join(error_imgs)
+  return "Images Uploaded", error_imgs_text
 
 
 def get_collection_names():
@@ -863,6 +876,7 @@ with gr.Blocks() as block:
     # Brandfolder Zipfile Dashboard
     This dashboard is for uploading photos from a zipfile to a brandfolder collection. 
     """)
+    
     with gr.Column(visible=True, elem_id='login') as login:
         password = gr.Textbox(label='Enter Password')
         # with gr.Row():
@@ -873,16 +887,15 @@ with gr.Blocks() as block:
             with gr.Column():
                 options = get_collection_names()
                 selection = gr.Dropdown(options, label='Choose Existing Collection', info='If creating a new section, select Create a Collection')
-                gr.Markdown('Upload zipfile containing client photos below')
+                gr.Markdown('## Upload zipfile containing client photos below')
                 zipfile = gr.File(label='Client Photos (must be zipfile)', file_count='multiple', file_types=['.zip'])
                 ai_bool = gr.Radio(choices=['On', 'Off'], label='AI Algorithm?', info = 'Would you like to use the AI Algorithm to upload these images?')
                 project_bool = gr.Radio(choices=['Yes', 'No'], label='Project Names?', info='Would you like to include project names for these images?')
-                gr.Markdown('Upload topical map document for the client below')
+                gr.Markdown('## Upload topical map document for the client below')
                 topical_map = gr.File(label='Topical Map (must be docx)', file_types=['.docx'])
-                
-                done_num = gr.Number(label='How many images were already processed?')
                 algorithm = gr.Button('Run Algorithm')
                 upload = gr.Label(label='Uploader')
+                err_imgs = gr.Textbox(label="Images Not Processed")
                 stop = gr.Button("Stop Run")
                 # with gr.Row():
                 #     zipfile_to_login_btn = gr.Button("Back to Login")
@@ -924,7 +937,7 @@ with gr.Blocks() as block:
         #     offline_to_trigger_btn = gr.Button("Brandfolder Trigger")
 
     # selection.select(fn=get_collection_names, outputs=[selection])
-    algo_event = algorithm.click(fn=import_client_data, inputs=[selection, zipfile, topical_map, password, project_bool, ai_bool, done_num], outputs=[upload])
+    algo_event = algorithm.click(fn=import_client_data, inputs=[selection, zipfile, topical_map, password, project_bool, ai_bool, done_num], outputs=[upload, err_imgs])
     bf_button.click(fn=bf_trigger.run_preprocess_ai, inputs=[bf_topical_map, bf_selection, section], outputs=[bf_upload])
     offline_event = offline_button.click(fn=offline_update.run_preprocess_ai, inputs=[offline_topical_map, offline_selection], outputs=[offline_upload])
     # stop.click(fn=None, inputs=None, outputs=None, cancels=[algo_event])