Rename old_app.py to multi_video_app.py

multi_video_app.py

@@ -0,0 +1,125 @@
+import warnings
+warnings.filterwarnings("ignore")
+import gradio as gr
+import re
+from typing import Dict, List
+import csv
+import os
+import torch
+from src.video_model import describe_video 
+from src.utils import parse_string, parse_annotations
+
+# Utility functions (from your provided utilities)
+
+# Function to save data to a CSV file
+def save_to_csv(observations: List[Dict], 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='w', newline='') as file:
+        writer = csv.writer(file)
+        writer.writerow(["video_name", "standing", "hands_free", "indoors", "screen_interaction_yes"])
+        for observation in observations:
+            writer.writerow([
+                observation['video_name'], 
+                observation['standing'], 
+                observation['hands_free'], 
+                observation['indoors'], 
+                observation['screen_interaction_yes']
+            ])
+    
+    return csv_file
+
+# Function to process a single video and return the observation data
+def process_single_video(video_path: str, sitting, hands, location, screen) -> Dict:
+    video_name = os.path.basename(video_path)  # Extract video name from the path
+    query = "Describe this video in detail and answer the questions"
+    additional_info = []
+    if sitting:
+        additional_info.append("Is the subject in the video standing or sitting?")
+    if hands:
+        additional_info.append("Is the subject holding any object in their hands, if so the hands are not free else they are free?")
+    if location:
+        additional_info.append("Is the subject present indoors or outdoors?")
+    if screen:
+        additional_info.append("Is the subject interacting with a screen in the background by facing the screen?")
+    
+    end_query = """Provide the results in <annotation> tags, where 0 indicates False, 1 indicates True, and None indicates that no information is present. Follow the below examples:
+        <annotation>indoors: 0</annotation>
+        <annotation>standing: 1</annotation>
+        <annotation>hands.free: 0</annotation>
+        <annotation>screen.interaction_yes: 0</annotation>
+        """
+    
+    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_path, final_prompt)
+
+    try:
+        # Parse the annotations from the response
+        tags = ["annotation"]
+        parsed_data = parse_string(response, tags)
+        annotations_list = parsed_data.get("annotation", [])
+        annotations_dict = parse_annotations(annotations_list)
+        
+        return {
+            "video_name": video_name,
+            "standing": annotations_dict.get("standing", 'N/A'),
+            "hands_free": annotations_dict.get("hands.free", 'N/A'),
+            "indoors": annotations_dict.get("indoors", 'N/A'),
+            "screen_interaction_yes": annotations_dict.get("screen.interaction_yes", 'N/A'),
+        }
+    except Exception as e:
+        return {"error": f"An error occurred with {video_name}: {e}"}
+
+# Function to process all videos in a folder
+def process_folder_of_videos(videos_folder: str, sitting, hands, location, screen):
+    all_observations = []
+    video_files = [os.path.join(videos_folder, f) for f in os.listdir(videos_folder) if f.endswith(('.mp4', '.avi', '.mkv'))]
+
+    for video_path in video_files:
+        observation = process_single_video(video_path, sitting, hands, location, screen)
+        if "error" not in observation:
+            all_observations.append(observation)
+        else:
+            print(observation["error"])  # Log any errors
+
+        # Clear GPU cache
+        torch.cuda.empty_cache()
+
+    # Save all observations to a CSV file and return the file path
+    csv_file = save_to_csv(all_observations)
+    return "Processing completed. Download the CSV file.", csv_file
+
+# Gradio interface
+def gradio_interface(videos_folder, sitting, hands, location, screen):
+    return process_folder_of_videos(videos_folder, sitting, hands, location, screen)
+
+# Inputs
+videos_folder = gr.Directory(label="Upload a folder of videos")
+sitting = gr.Checkbox(label="Sitting/Standing")
+hands = gr.Checkbox(label="Hands Free/Not Free")
+location = gr.Checkbox(label="Indoors/Outdoors")
+screen = gr.Checkbox(label="Screen Interaction")
+
+# Outputs
+response = gr.Textbox(label="Status")
+download_link = gr.File(label="Download CSV")
+
+# Gradio interface setup
+interface = gr.Interface(
+    fn=gradio_interface,
+    inputs=[videos_folder, sitting, hands, location, screen],
+    outputs=[response, download_link],
+    title="Batch Video Annotation",
+    description="Upload a folder of videos and process them sequentially, saving the results to a downloadable CSV file.",
+    theme=gr.themes.Soft(primary_hue="red", secondary_hue="red"),
+    allow_flagging="never"
+)
+
+# Launch interface
+interface.launch(debug=False)

old_app.py

@@ -1,102 +0,0 @@
-import warnings
-warnings.filterwarnings("ignore")
-import gradio as gr
-import pandas as pd
-from src.video_model import describe_video  
-from src.utils import parse_string, parse_annotations
-import os
-
-# --- Function to construct the final query --- 
-def process_video_and_questions(video, standing, hands, location, screen):
-    video_name = os.path.basename(video)
-    query = f"Answer the questions from the video\n"
-    additional_info = []
-    if standing:
-        additional_info.append("Is the subject in the video standing or sitting?\n")
-    if hands:
-        additional_info.append("Is the subject holding any object in their hands?\n")
-    if location:
-        additional_info.append("Is the subject present indoors?\n")
-    if screen:
-        additional_info.append("Is the subject interacting with a screen in the background by facing the screen?\n")
-    
-    end_query = """Provide the results in <annotation> tags, where 0 indicates False, 1 indicates True, and None indicates that no information is present. Follow the below examples\n:
-        <annotation>indoors: 0</annotation>
-        <annotation>standing: 1</annotation>
-        <annotation>hands.free: 0</annotation>
-        <annotation>screen.interaction_yes: 0</annotation>
-        """
-
-    final_query = query + " " + " ".join(additional_info)
-    final_prompt = final_query + " " + end_query
-    
-    response = describe_video(video, final_prompt)
-    final_response = f"<video_name>{video_name}</video_name>" + " \n" + response
-
-    conditions = {
-        'standing': (standing, 'standing: 1', 'standing: None'),
-        'hands': (hands, 'hands.free: 1', 'hands.free: None'),
-        'location': (location, 'indoors: 1', 'indoors: None'),
-        'screen': (screen, 'screen.interaction_yes: 1', 'screen.interaction_yes: None')
-    }
-    
-    for key, (condition, to_replace, replacement) in conditions.items():
-        if not condition:
-            final_response = final_response.replace(to_replace, replacement)
-    
-    return final_response 
-
-def process_multiple_videos(video_files, standing, hands, location, screen):
-    # Initialize an empty DataFrame to store results for all videos
-    all_results_df = pd.DataFrame()
-
-    for video in video_files:
-        final_response = process_video_and_questions(video.name, standing, hands, location, screen)
-        video_df = output_to_csv(final_response)
-        all_results_df = pd.concat([all_results_df, video_df], ignore_index=True)
-    
-    # Save the combined results as a CSV file
-    csv_file_path = "multiple_videos_annotations.csv"
-    all_results_df.to_csv(csv_file_path, index=False)
-    
-    return csv_file_path  # Return the path to the CSV file for download
-
-def output_to_csv(final_response):
-    parsed_content = parse_string(final_response, ["video_name", "annotation"])
-    video_name = parsed_content['video_name'][0] if parsed_content['video_name'] else None
-    annotations_dict = parse_annotations(parsed_content['annotation']) if parsed_content['annotation'] else {}
-    
-    df = pd.DataFrame([{'video_name': video_name, **annotations_dict}])
-    
-    return df
-
-title = "GSoC Super Raid Annotator"
-description = "Annotate Multiple 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(primary_hue="red", secondary_hue="red")
-
-with gr.Blocks(theme=custom_theme) as demo:
-    gr.Markdown(f"# {title}")
-    gr.Markdown(description)
-    gr.Markdown(article)
-    
-    with gr.Row():
-        with gr.Column():
-            video_files = gr.Files(label="Upload Videos", file_count="multiple")
-            standing = gr.Checkbox(label="Standing")
-            hands = gr.Checkbox(label="Hands Free")
-            location = gr.Checkbox(label="Indoors")
-            screen = gr.Checkbox(label="Screen Interaction")
-            generate_csv_btn = gr.Button("Process and Generate CSV")
-        
-        with gr.Column():
-            csv_output = gr.File(label="Download CSV", interactive=False)
-    
-    generate_csv_btn.click(
-        fn=process_multiple_videos,
-        inputs=[video_files, standing, hands, location, screen],
-        outputs=csv_output
-    )
-    
-demo.launch(debug=False)