app.py

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 os

# --- Function to construct the final query --- 
def process_video_and_questions(video, sitting, hands, location, screen):
    # Extract the video name (filename)
    video_name = os.path.basename(video)
    
    # Construct the query with the video name included
    query = f"Describe the video '{video_name}' 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 for example with 0 being False and 1 being True\n
    <annotation>indoor: 0</annotation>
    <annotation>standing: 1</annotation>
    <annotation>hands_free: 1</annotation>
    <annotation>screen_interaction: 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, final_prompt)
    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")

# 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"]
]

# Title, description, and article for the interface
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(
    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)