Update app.py

app.py

@@ -1,40 +1,10 @@
 import warnings
 warnings.filterwarnings("ignore")
 import gradio as gr
-import json
-from pydantic import BaseModel
-from typing import Dict
-import csv
-import os
 from src.video_model import describe_video  # Assuming this function processes the video and query
 
-# Define the Pydantic model for structured data
-class VideoObservation(BaseModel):
-    is_subject_standing: int
-    is_subject_holding_object_in_hands: int
-    is_subject_present_indoors: int
-    is_subject_interacting_with_screen: int
-
-# Function to save data to a CSV file and generate a download link
-def save_to_csv(video_name: str, observation: VideoObservation, output_dir: str = "outputs") -> str:
-    if not os.path.exists(output_dir):
-        os.makedirs(output_dir)
-    
-    csv_file = os.path.join(output_dir, "video_observations.csv")
-    
-    with open(csv_file, mode='a', newline='') as file:
-        writer = csv.writer(file)
-        # Write the header if the file is empty
-        if file.tell() == 0:
-            writer.writerow(["video_name", "is_subject_standing", "is_subject_holding_object_in_hands", "is_subject_present_indoors", "is_subject_interacting_with_screen"])
-        # Write the data
-        writer.writerow([video_name, observation.is_subject_standing, observation.is_subject_holding_object_in_hands, observation.is_subject_present_indoors, observation.is_subject_interacting_with_screen])
-    
-    return csv_file
-
-# --- Function to process the video and questions, then save results to CSV --- 
+# --- Function to construct the final query --- 
 def process_video_and_questions(video, sitting, hands, location, screen):
-    video_name = video.split("/")[-1]  # Extract video name from the path
     query = "Describe this video in detail and answer the questions"
     additional_info = []
     if sitting:
@@ -48,26 +18,9 @@ def process_video_and_questions(video, sitting, hands, location, screen):
     end_query = "Provide the results in JSON format with 0 being False and 1 being True"
     final_query = query + " " + " ".join(additional_info)
     final_prompt = final_query + " " + end_query
-    
     # Assuming your describe_video function handles the video processing
     response = describe_video(video, final_prompt)
-    
-    # Extract the JSON part from the response
-    start_index = response.find('{')
-    end_index = response.rfind('}') + 1
-    json_str = response[start_index:end_index]
-
-    try:
-        # Parse JSON string into a Pydantic model
-        observation = VideoObservation.parse_raw(json_str)
-        # Save to CSV and return the file path
-        csv_file = save_to_csv(video_name, observation)
-    except json.JSONDecodeError as e:
-        return f"Error parsing JSON: {e}", None
-    except Exception as e:
-        return f"An error occurred: {e}", None
-    
-    return response, csv_file
+    return response
 
 # Video and text inputs for the interface
 video = gr.Video(label="Video")
@@ -80,7 +33,6 @@ screen = gr.Checkbox(label="Screen Interaction")
 
 # Output for the interface
 response = gr.Textbox(label="Predicted answer", show_label=True, show_copy_button=True)
-download_link = gr.File(label="Download CSV")
 
 # Examples for the interface
 examples = [
@@ -102,16 +54,17 @@ title = "GSoC Super Raid Annotator"
 description = "Annotate Videos"
 article = "<p style='text-align: center'><a href='https://github.com/OpenBMB/MiniCPM-V' target='_blank'>Model GitHub Repo</a> | <a href='https://huggingface.co/openbmb/MiniCPM-V-2_6' target='_blank'>Model Page</a></p>"
 
+
 custom_theme = gr.themes.Soft(
+    # Set the primary hue of the Soft theme to your red color
     primary_hue="red", 
-    secondary_hue="red"
-)
+    secondary_hue="red")
 
 # Launch the interface
 interface = gr.Interface(
-    fn=process_video_and_questions,  # Updated function to handle the query construction and CSV saving
+    fn=process_video_and_questions, # Updated function to handle the query construction
     inputs=[video, sitting, hands, location, screen], 
-    outputs=[response, download_link],
+    outputs=response,
     examples=examples,
     title=title,
     description=description,
@@ -119,4 +72,4 @@ interface = gr.Interface(
     theme=custom_theme,
     allow_flagging="never",
 )
-interface.launch(debug=False)
+interface.launch(debug=False)