Spaces:
Running
Running
| import os | |
| import streamlit as st | |
| from moviepy.video.io.VideoFileClip import VideoFileClip | |
| from pydub import AudioSegment | |
| import whisper | |
| from transformers import pipeline, MarianMTModel, MarianTokenizer | |
| import yt_dlp as youtube_dl | |
| # App Configuration | |
| st.set_page_config(page_title="Video-to-Text Summarization", layout="centered") | |
| # Header | |
| st.title("🎥 Smart Video-to-Text Summarization App") | |
| st.markdown(""" | |
| This app helps you: | |
| - Convert videos into text and summarize them. | |
| - Extract multilingual transcriptions and translations. | |
| - Process videos with multiple speakers. | |
| """) | |
| # Temporary video storage | |
| if "video_path" not in st.session_state: | |
| st.session_state.video_path = None | |
| # 1. Upload Video Section | |
| st.header("Upload Your Video") | |
| # Choose upload option | |
| upload_option = st.selectbox("Select Upload Method", ["Local", "YouTube URL"]) | |
| # Upload Local File | |
| if upload_option == "Local": | |
| video_file = st.file_uploader("Upload your video file", type=["mp4", "mkv", "avi"]) | |
| if video_file: | |
| with open("uploaded_video.mp4", "wb") as f: | |
| f.write(video_file.read()) | |
| st.session_state.video_path = "uploaded_video.mp4" | |
| st.success("Video uploaded successfully!") | |
| # Download Video from YouTube | |
| elif upload_option == "YouTube URL": | |
| youtube_url = st.text_input("Enter YouTube URL") | |
| if youtube_url: | |
| try: | |
| os.system(f"yt-dlp -o video.mp4 {youtube_url}") | |
| st.session_state.video_path = "video.mp4" | |
| st.success("YouTube video downloaded successfully!") | |
| except Exception as e: | |
| st.error(f"Error downloading video: {str(e)}") | |
| # 2. Process Video Section (After Upload) | |
| if st.session_state.video_path: | |
| st.header("Process Your Video") | |
| st.write(f"Processing {st.session_state.video_path}...") | |
| # Extract Audio from Video | |
| def extract_audio(video_path): | |
| try: | |
| audio = AudioSegment.from_file(video_path) | |
| audio.export("extracted_audio.mp3", format="mp3") | |
| st.success("Audio extracted successfully!") | |
| return "extracted_audio.mp3" | |
| except Exception as e: | |
| st.error(f"Error in extracting audio: {str(e)}") | |
| return None | |
| audio_path = extract_audio(st.session_state.video_path) | |
| # Real-time Audio Transcription | |
| def transcribe_audio(audio_path): | |
| try: | |
| model = whisper.load_model("base") | |
| result = model.transcribe(audio_path) | |
| st.text_area("Transcription", result['text'], height=200) | |
| return result['text'] | |
| except Exception as e: | |
| st.error(f"Error in transcription: {str(e)}") | |
| return None | |
| if audio_path: | |
| transcription = transcribe_audio(audio_path) | |
| # 3. Summarize and Translate | |
| if 'transcription' in locals(): | |
| st.header("Results") | |
| # Summarize Text | |
| def summarize_text(text): | |
| try: | |
| summarizer = pipeline("summarization", model="facebook/bart-large-cnn") | |
| summary = summarizer(text, max_length=150, min_length=30, do_sample=False) | |
| st.text_area("Summary", summary[0]['summary_text'], height=150) | |
| return summary[0]['summary_text'] | |
| except Exception as e: | |
| st.error(f"Error in summarization: {str(e)}") | |
| return None | |
| summary = summarize_text(transcription) | |
| # Translate Text | |
| def translate_text(text, src_lang="en", tgt_lang="es"): | |
| try: | |
| model_name = f"Helsinki-NLP/opus-mt-{src_lang}-{tgt_lang}" | |
| tokenizer = MarianTokenizer.from_pretrained(model_name) | |
| model = MarianMTModel.from_pretrained(model_name) | |
| translated = model.generate(**tokenizer(text, return_tensors="pt", padding=True)) | |
| translated_text = tokenizer.decode(translated[0], skip_special_tokens=True) | |
| st.text_area("Translated Summary", translated_text, height=150) | |
| return translated_text | |
| except Exception as e: | |
| st.error(f"Error in translation: {str(e)}") | |
| return None | |
| target_language = st.selectbox("Select Translation Language", ["es", "fr", "de", "zh"]) | |
| if target_language: | |
| translated_summary = translate_text(summary, tgt_lang=target_language) | |
| else: | |
| st.info("Please upload a video to start the process.") |