app.py

import gradio as gr
import os
import google.generativeai as genai
from elevenlabs.client import ElevenLabs
from tavily import TavilyClient
import requests
import subprocess
import json
import time
import random

# --- 1. CONFIGURE API KEYS FROM HUGGING FACE SECRETS ---
try:
    genai.configure(api_key=os.environ["GEMINI_API_KEY"])
    tavily_client = TavilyClient(api_key=os.environ["TAVILY_API_KEY"])
    RUNWAY_API_KEY = os.environ["RUNWAY_API_KEY"]
    elevenlabs_client = ElevenLabs(api_key=os.environ["ELEVENLABS_API_KEY"])
except KeyError as e:
    raise ValueError(f"API Key Error: Please set the {e} secret in your Hugging Face Space settings.")

# --- 2. DEFINE API ENDPOINTS AND HEADERS ---
RUNWAY_API_URL = "https://api.runwayml.com/v1/jobs"
RUNWAY_HEADERS = {
    "Authorization": f"Bearer {RUNWAY_API_KEY}",
    "Content-Type": "application/json"
}

# --- 3. THE CORE VIDEO GENERATION FUNCTION ---
def generate_video_from_topic(topic_prompt, progress=gr.Progress(track_tqdm=True)):
    job_id = f"{int(time.time())}_{random.randint(1000, 9999)}"
    print(f"--- Starting New Job: {job_id} for topic: '{topic_prompt}' ---")
    
    intermediate_files = []

    try:
        # STEP 1: RESEARCH (Tavily)
        progress(0.1, desc="🔍 Researching topic with Tavily...")
        try:
            research_results = tavily_client.search(query=f"Key facts and interesting points about {topic_prompt}", search_depth="basic")
            facts = "\n".join([res['content'] for res in research_results['results']])
            print(f"Tavily Research Results:\n{facts}")
        except Exception as e:
            print(f"Tavily API failed: {e}. Proceeding without research.")
            facts = "No research data available."

        # STEP 2: SCRIPT & SCENE PROMPTS (Gemini)
        progress(0.2, desc="✍️ Writing script with Gemini...")
        gemini_model = genai.GenerativeModel('gemini-1.5-flash')
        prompt = f"""
        You are a creative director for viral short-form videos.
        Based on the topic '{topic_prompt}' and the following research:
        ---
        {facts}
        ---
        Create a script for a 30-45 second video.
        Your output MUST be a valid JSON object with two keys:
        1. "narration_script": A string containing the full voiceover narration. Make it engaging and concise.
        2. "scene_prompts": A list of exactly 4 strings. Each string must be a highly detailed, visually rich, and cinematic prompt for a text-to-video AI like Runway Gen-2. Describe camera angles, lighting, and mood.
        """
        response = gemini_model.generate_content(prompt)
        
        try:
            cleaned_text = response.text.strip().replace("```json", "").replace("```", "")
            script_data = json.loads(cleaned_text)
            narration = script_data['narration_script']
            scene_prompts = script_data['scene_prompts']
            print(f"Gemini Script:\n{narration}\nPrompts:\n{scene_prompts}")
        except (json.JSONDecodeError, KeyError) as e:
            raise gr.Error(f"Gemini did not return valid JSON. Error: {e}. Response was: {response.text}")

        # STEP 3: VOICE OVER (ElevenLabs)
        progress(0.3, desc="🎙️ Recording voiceover with ElevenLabs...")
        audio_path = f"audio_{job_id}.mp3"
        intermediate_files.append(audio_path)
        
        # !!!!!!!!!!! THIS IS THE CORRECTED LINE !!!!!!!!!!!
        audio_bytes = elevenlabs_client.generate.from_text(
            text=narration,
            voice="Adam", # You can choose any available voice
            model="eleven_multilingual_v2"
        )
        # Save the audio file
        with open(audio_path, "wb") as f:
            for chunk in audio_bytes:
                f.write(chunk)
        print(f"Audio file saved: {audio_path}")

        # STEP 4: VISUALS (Runway)
        video_clip_paths = []
        for i, scene_prompt in enumerate(scene_prompts):
            progress(0.4 + (i * 0.12), desc=f"🎬 Generating video scene {i+1}/{len(scene_prompts)}...")
            runway_payload = {"text_prompt": scene_prompt}
            post_response = requests.post(RUNWAY_API_URL, headers=RUNWAY_HEADERS, json=runway_payload)
            if post_response.status_code != 200:
                raise gr.Error(f"Runway API Error (start job): {post_response.status_code} - {post_response.text}")
            
            task_id = post_response.json().get("uuid")
            if not task_id:
                raise gr.Error(f"Runway API did not return a task UUID. Response: {post_response.json()}")

            video_url = None
            for _ in range(60): # Poll for up to 10 minutes (60 * 10s)
                get_response = requests.get(f"{RUNWAY_API_URL}/{task_id}", headers=RUNWAY_HEADERS)
                status_details = get_response.json()
                status = status_details.get("status")
                
                if status == "succeeded":
                    video_url = status_details.get("outputs", {}).get("video")
                    break
                elif status == "failed":
                    raise gr.Error(f"Runway job failed. Details: {status_details.get('error_message')}")
                
                print(f"Scene {i+1} status: {status}. Waiting 10 seconds...")
                time.sleep(10)
            
            if not video_url:
                raise gr.Error(f"Runway job timed out for scene {i+1}.")

            clip_path = f"scene_{i+1}_{job_id}.mp4"
            intermediate_files.append(clip_path)
            video_clip_paths.append(clip_path)
            
            video_response = requests.get(video_url, stream=True)
            with open(clip_path, "wb") as f:
                for chunk in video_response.iter_content(chunk_size=1024):
                    if chunk:
                        f.write(chunk)
            print(f"Video clip saved: {clip_path}")

        # STEP 5: STITCHING (FFmpeg)
        progress(0.9, desc="✂️ Assembling final video with FFmpeg...")
        file_list_path = f"file_list_{job_id}.txt"
        intermediate_files.append(file_list_path)
        with open(file_list_path, "w") as f:
            for clip in video_clip_paths:
                f.write(f"file '{clip}'\n")

        combined_video_path = f"combined_video_{job_id}.mp4"
        intermediate_files.append(combined_video_path)
        subprocess.run(['ffmpeg', '-f', 'concat', '-safe', '0', '-i', file_list_path, '-c', 'copy', combined_video_path, '-y'], check=True)
        
        final_video_path = f"final_video_{job_id}.mp4"
        subprocess.run([
            'ffmpeg', '-i', combined_video_path, '-i', audio_path, '-c:v', 'copy',
            '-c:a', 'aac', '-shortest', final_video_path, '-y'
        ], check=True)
        print(f"Final video created at: {final_video_path}")
        
        progress(1.0, desc="✅ Done!")
        return final_video_path

    except Exception as e:
        print(f"--- JOB {job_id} FAILED --- \nError: {e}")
        raise gr.Error(f"An error occurred: {e}")
    
    finally:
        # STEP 6: CLEANUP
        print("Cleaning up intermediate files...")
        for file_path in intermediate_files:
            if os.path.exists(file_path):
                os.remove(file_path)
                print(f"Removed: {file_path}")

# --- 4. CREATE AND LAUNCH THE GRADIO INTERFACE ---
with gr.Blocks(theme=gr.themes.Soft()) as demo:
    gr.Markdown(
        """
        # 🤖 My Personal AI Video Studio
        Enter a topic to generate a short-form video. This private tool is used for fulfilling freelance orders.
        """
    )
    
    with gr.Row():
        topic_input = gr.Textbox(
            label="Video Topic", 
            placeholder="e.g., 'The history of coffee' or 'Benefits of our new CRM software'",
            scale=3
        )
        generate_button = gr.Button("Generate Video", variant="primary", scale=1)
        
    with gr.Row():
        video_output = gr.Video(label="Generated Video")
        
    generate_button.click(
        fn=generate_video_from_topic,
        inputs=topic_input,
        outputs=video_output
    )
    
    gr.Markdown("--- \n ### Examples of Good Topics:\n - A product: 'The new waterproof Chrono-Watch X1'\n - A concept: 'The science of sleep'\n - A story: 'The legend of El Dorado'")


if __name__ == "__main__":
    demo.launch()