fina p3

src/inference_logic.py

@@ -274,8 +274,20 @@ async def run_gemini_labeling_pipeline(video_path: str, caption: str, transcript
         uploaded_file = None
         is_text_only = False
         if video_path and os.path.exists(video_path):
+             yield f"data:   - Uploading video to Gemini API...\n\n"
              uploaded_file = await loop.run_in_executor(None, lambda: genai_legacy.upload_file(path=video_path))
-             while uploaded_file.state.name == "PROCESSING": await asyncio.sleep(2)
+             
+             # Continuously poll the API for the updated state
+             while True:
+                 curr_file = await loop.run_in_executor(None, lambda: genai_legacy.get_file(uploaded_file.name))
+                 if curr_file.state.name == "PROCESSING":
+                     yield f"data:   - Waiting for Gemini to process the video...\n\n"
+                     await asyncio.sleep(3)
+                 elif curr_file.state.name == "FAILED":
+                     yield f"data:   - Gemini Video Processing FAILED.\n\n"
+                     break
+                 else:
+                     break
         else: is_text_only = True
         
         active_tools =[]
@@ -313,8 +325,8 @@ async def run_gemini_labeling_pipeline(video_path: str, caption: str, transcript
                      f"{toon_schema}"
                 )
                 save_debug_log(request_id, 'prompt', prompt_text, attempt, 'reprompt')
-                inputs = [prompt_text]
-                if uploaded_file: inputs.append(uploaded_file)
+                inputs =[prompt_text]
+                if uploaded_file and uploaded_file.state.name != "FAILED": inputs.append(uploaded_file)
                 response = await loop.run_in_executor(None, lambda: model.generate_content(inputs, generation_config={"temperature": 0.2}))
                 raw_text = response.text
                 save_debug_log(request_id, 'response', raw_text, attempt, 'reprompt')
@@ -325,8 +337,8 @@ async def run_gemini_labeling_pipeline(video_path: str, caption: str, transcript
                     
                     macro_prompt = FCOT_MACRO_PROMPT.format(system_persona=system_persona, caption=caption, transcript=transcript)
                     save_debug_log(request_id, 'prompt', macro_prompt, attempt, 'fcot_macro')
-                    inputs1 =[macro_prompt]
-                    if uploaded_file: inputs1.insert(0, uploaded_file)
+                    inputs1 = [macro_prompt]
+                    if uploaded_file and uploaded_file.state.name != "FAILED": inputs1.insert(0, uploaded_file)
                     res1 = await loop.run_in_executor(None, lambda: chat.send_message(inputs1))
                     macro_hypothesis = res1.text
                     save_debug_log(request_id, 'response', macro_hypothesis, attempt, 'fcot_macro')
@@ -354,8 +366,8 @@ async def run_gemini_labeling_pipeline(video_path: str, caption: str, transcript
                     prompt_used = prompt_text
                     if is_text_only: prompt_text = "NOTE: Text Analysis Only.\n" + prompt_text
                     save_debug_log(request_id, 'prompt', prompt_text, attempt, f'standard_{reasoning_method}')
-                    inputs = [prompt_text]
-                    if uploaded_file: inputs.append(uploaded_file)
+                    inputs =[prompt_text]
+                    if uploaded_file and uploaded_file.state.name != "FAILED": inputs.append(uploaded_file)
                     response = await loop.run_in_executor(None, lambda: model.generate_content(inputs, generation_config={"temperature": 0.1}))
                     raw_text = response.text
                     save_debug_log(request_id, 'response', raw_text, attempt, f'standard_{reasoning_method}')
@@ -462,7 +474,7 @@ async def run_vertex_labeling_pipeline(video_path: str, caption: str, transcript
                 )
                 
                 save_debug_log(request_id, 'prompt', prompt_text, attempt, 'reprompt')
-                contents = [prompt_text]
+                contents =[prompt_text]
                 if video_part: contents.insert(0, video_part)
                 
                 response = await loop.run_in_executor(None, lambda: client.models.generate_content(model=model_name, contents=contents, config=config))