Update app.py

app.py

@@ -1,7 +1,9 @@
 import warnings
 warnings.filterwarnings("ignore")
 import gradio as gr
-from src.video_model import describe_video  # Assuming this function processes the video and query
+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 --- 
@@ -33,31 +35,36 @@ def process_video_and_questions(video, sitting, hands, location, screen):
     final_response = f"<video_name>{video_name}</video_name>" + " " + response
     return final_response 
 
-# Video and text inputs for the interface
-video = gr.Video(label="Video")
 
-# Options as checkboxes
-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")
+def output_to_csv(final_response):
+    # Parse the string to get the content
+    parsed_content = parse_string(final_response, ["video_name", "annotation"])
+    
+    # Extract video name (assuming there's only one video name)
+    video_name = parsed_content['video_name'][0] if parsed_content['video_name'] else None
+    
+    # Convert the list of annotations into a dictionary
+    annotations_dict = parse_annotations(parsed_content['annotation']) if parsed_content['annotation'] else {}
+    
+    # Combine the video name and annotation dictionary into a single row
+    df = pd.DataFrame([{'video_name': video_name, **annotations_dict}])
+
+    return df
 
-# Output for the interface
-response = gr.Textbox(label="Predicted answer", show_label=True, show_copy_button=True)
 
 # Examples for the interface
 examples = [
-    ["videos/2016-01-01_0100_US_KNBC_Channel_4_News_1867.16-1871.38_now.mp4",],
-    ["videos/2016-01-01_0200_US_KNBC_Channel_4_News_1329.12-1333.29_tonight.mp4",],
-    ["videos/2016-01-01_0830_US_KNBC_Tonight_Show_with_Jimmy_Fallon_725.45-729.76_tonight.mp4",],
-    ["videos/2016-01-01_0200_US_KOCE_The_PBS_Newshour_577.03-581.31_tonight.mp4"],
-    ["videos/2016-01-01_1400_US_KTTV-FOX_Morning_News_at_6AM_1842.36-1846.68_this_year.mp4"],
-    ["videos/2016-01-02_0735_US_KCBS_Late_Show_with_Stephen_Colbert_285.94-290.67_this_year.mp4"],
-    ["videos/2016-01-13_2200_US_KTTV-FOX_The_Doctor_Oz_Show_1709.79-1714.17_this_month.mp4"],
-    ["videos/2016-01-01_1400_US_KTTV-FOX_Morning_News_at_6AM_1842.36-1846.68_this_year.mp4"],
-    ["videos/2016-01-01_1300_US_KNBC_Today_in_LA_at_5am_12.46-16.95_this_morning.mp4"],
-    ["videos/2016-01-05_0200_US_KNBC_Channel_4_News_1561.29-1565.95_next_week.mp4"],
-    ["videos/2016-01-28_0700_US_KNBC_Channel_4_News_at_11PM_629.56-633.99_in_the_future.mp4"]
+    ["videos/2016-01-01_0100_US_KNBC_Channel_4_News_1867.16-1871.38_now.mp4", True, False, True, False],
+    ["videos/2016-01-01_0200_US_KNBC_Channel_4_News_1329.12-1333.29_tonight.mp4", False, True, True, True],
+    ["videos/2016-01-01_0830_US_KNBC_Tonight_Show_with_Jimmy_Fallon_725.45-729.76_tonight.mp4", True, False, False, True],
+    ["videos/2016-01-01_0200_US_KOCE_The_PBS_Newshour_577.03-581.31_tonight.mp4", False, True, True, False],
+    ["videos/2016-01-01_1400_US_KTTV-FOX_Morning_News_at_6AM_1842.36-1846.68_this_year.mp4", True, True, False, False],
+    ["videos/2016-01-02_0735_US_KCBS_Late_Show_with_Stephen_Colbert_285.94-290.67_this_year.mp4", False, True, True, True],
+    ["videos/2016-01-13_2200_US_KTTV-FOX_The_Doctor_Oz_Show_1709.79-1714.17_this_month.mp4", True, False, False, True],
+    ["videos/2016-01-01_1400_US_KTTV-FOX_Morning_News_at_6AM_1842.36-1846.68_this_year.mp4", False, True, True, False],
+    ["videos/2016-01-01_1300_US_KNBC_Today_in_LA_at_5am_12.46-16.95_this_morning.mp4", True, False, False, True],
+    ["videos/2016-01-05_0200_US_KNBC_Channel_4_News_1561.29-1565.95_next_week.mp4", False, True, True, False],
+    ["videos/2016-01-28_0700_US_KNBC_Channel_4_News_at_11PM_629.56-633.99_in_the_future.mp4", True, False, False, True]
 ]
 
 # Title, description, and article for the interface
@@ -69,16 +76,39 @@ custom_theme = gr.themes.Soft(
     primary_hue="red", 
     secondary_hue="red")
 
-# Launch the interface
-interface = gr.Interface(
-    fn=process_video_and_questions, 
-    inputs=[video, sitting, hands, location, screen], 
-    outputs=response,
-    examples=examples,
-    title=title,
-    description=description,
-    article=article,
-    theme=custom_theme,
-    allow_flagging="never",
-)
-interface.launch(debug=False)
+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 = gr.Video(label="Video")
+            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")
+            submit_btn = gr.Button("Generate Annotations")
+            generate_csv_btn = gr.Button("Generate CSV")
+        
+        with gr.Column():
+            response = gr.Textbox(label="Video Description", show_label=True, show_copy_button=True)
+            csv_output = gr.DataFrame(label="CSV Output", headers=["video_name", "annotation"])
+    
+    # Event handling for the Submit button
+    submit_btn.click(
+        fn=process_video_and_questions,
+        inputs=[video, sitting, hands, location, screen],
+        outputs=response
+    )
+    
+    # Event handling for the Generate CSV button
+    generate_csv_btn.click(
+        fn=output_to_csv,
+        inputs=response,
+        outputs=csv_output
+    )
+    
+    gr.Examples(examples=examples, inputs=[video, sitting, hands, location, screen])
+    
+demo.launch(debug=False)