format app.py file

app.py

@@ -3,6 +3,7 @@ import os
 os.environ["HUGGINGFACE_DEMO"] = "1"  # set before import from app
 
 from dotenv import load_dotenv
+
 load_dotenv()
 ################################################################################################
 
@@ -20,10 +21,12 @@ settings = get_settings()
 IMAGE_MAX_SIZE = 1536
 
 
-async def forward_request(attributes, product_taxonomy, product_data, ai_model, pil_images):
+async def forward_request(
+    attributes, product_taxonomy, product_data, ai_model, pil_images
+):
     # prepare temp folder
     request_id = str(uuid.uuid4())
-    request_temp_folder = os.path.join('gradio_temp', request_id)
+    request_temp_folder = os.path.join("gradio_temp", request_id)
     os.makedirs(request_temp_folder, exist_ok=True)
 
     try:
@@ -32,9 +35,11 @@ async def forward_request(attributes, product_taxonomy, product_data, ai_model,
         try:
             attributes = exec(attributes, globals())
         except:
-            raise gr.Error("Invalid `Attribute Schema`. Please insert valid schema following the example.")
-        for key, value in attributes_object.items(): # type: ignore
-            attributes_object[key] = Attribute(**value) # type: ignore
+            raise gr.Error(
+                "Invalid `Attribute Schema`. Please insert valid schema following the example."
+            )
+        for key, value in attributes_object.items():  # type: ignore
+            attributes_object[key] = Attribute(**value)  # type: ignore
 
         if product_data == "":
             product_data = "{}"
@@ -43,39 +48,48 @@ async def forward_request(attributes, product_taxonomy, product_data, ai_model,
         try:
             exec(product_data_code, globals())
         except:
-            raise gr.Error('Invalid `Product Data`. Please insert valid dictionary or leave it empty.')
+            raise gr.Error(
+                "Invalid `Product Data`. Please insert valid dictionary or leave it empty."
+            )
 
         if pil_images is None:
-            raise gr.Error('Please upload image(s) of the product')
+            raise gr.Error("Please upload image(s) of the product")
         pil_images = [pil_image[0] for pil_image in pil_images]
         img_paths = []
         for i, pil_image in enumerate(pil_images):
             if max(pil_image.size) > IMAGE_MAX_SIZE:
                 ratio = IMAGE_MAX_SIZE / max(pil_image.size)
-                pil_image = pil_image.resize((int(pil_image.width * ratio), int(pil_image.height * ratio)))
-            img_path = os.path.join(request_temp_folder, f'{i}.jpg')
-            if pil_image.mode in ('RGBA', 'LA') or (pil_image.mode == 'P' and 'transparency' in pil_image.info):
+                pil_image = pil_image.resize(
+                    (int(pil_image.width * ratio), int(pil_image.height * ratio))
+                )
+            img_path = os.path.join(request_temp_folder, f"{i}.jpg")
+            if pil_image.mode in ("RGBA", "LA") or (
+                pil_image.mode == "P" and "transparency" in pil_image.info
+            ):
                 pil_image = pil_image.convert("RGBA")
-                if pil_image.getchannel("A").getextrema() == (255, 255):  # if fully opaque, save as JPEG
+                if pil_image.getchannel("A").getextrema() == (
+                    255,
+                    255,
+                ):  # if fully opaque, save as JPEG
                     pil_image = pil_image.convert("RGB")
-                    image_format = 'JPEG'
+                    image_format = "JPEG"
                 else:
-                    image_format = 'PNG'
+                    image_format = "PNG"
             else:
-                image_format = 'JPEG'
+                image_format = "JPEG"
             pil_image.save(img_path, image_format, quality=100, subsampling=0)
             img_paths.append(img_path)
 
         # mapping
         if ai_model in settings.OPENAI_MODELS:
-            ai_vendor = 'openai'
+            ai_vendor = "openai"
         elif ai_model in settings.ANTHROPIC_MODELS:
-            ai_vendor = 'anthropic'
+            ai_vendor = "anthropic"
         service = AIServiceFactory.get_service(ai_vendor)
 
         try:
             json_attributes = await service.extract_attributes_with_validation(
-                attributes_object, # type: ignore
+                attributes_object,  # type: ignore
                 ai_model,
                 None,
                 product_taxonomy,
@@ -83,12 +97,12 @@ async def forward_request(attributes, product_taxonomy, product_data, ai_model,
                 img_paths=img_paths,
             )
         except:
-            raise gr.Error('Failed to extract attributes. Something went wrong.')
+            raise gr.Error("Failed to extract attributes. Something went wrong.")
     finally:
         # remove temp folder anyway
         shutil.rmtree(request_temp_folder)
-    
-    gr.Info('Process completed!')
+
+    gr.Info("Process completed!")
     return json_attributes
 
 
@@ -209,7 +223,7 @@ with gr.Blocks(title="Internal Demo for Attribution") as demo:
                     #                     ka_value = gr.Textbox(placeholder="data", key=f"data-{i}", show_label=False)
                     #                     ka_names.append(ka_name)
                     #                     ka_values.append(ka_value)
-                    
+
                     # add_track_btn = gr.Button("Add Product Data")
                     # remove_track_btn = gr.Button("Remove Product Data")
                     # add_track_btn.click(lambda count: count + 1, track_count, track_count)
@@ -272,4 +286,8 @@ with gr.Blocks(title="Internal Demo for Attribution") as demo:
         outputs=output_json,
     )
 
-demo.launch()
+
+attr_user = os.getenv("ATTR_USER", "1")
+attr_pass = os.getenv("ATTR_PASS", "a")
+auth = (attr_user, attr_pass)
+demo.launch(auth=auth, debug=True)