app.py

import requests_proxy
import gradio as gr
from tts_module import get_voices, text_to_speech
from moviepy.editor import (
    AudioFileClip, ImageClip, CompositeAudioClip,
    concatenate_audioclips, concatenate_videoclips, vfx, CompositeVideoClip,
    ColorClip
)
import asyncio
import os
import json
import time
import requests
import random
from googleapiclient.discovery import build
from google.oauth2 import service_account
from googleapiclient.http import MediaFileUpload
from io import BytesIO
from PIL import Image
import numpy as np

MIN_WIDTH = 1920
MIN_HEIGHT = 1080
TARGET_ASPECT_RATIO = 16/9

output_folder = "outputs"
temp_dir = "temp_files"
os.makedirs(output_folder, exist_ok=True)
os.makedirs(temp_dir, exist_ok=True)

FOLDER_ID = "12S6adpanAXjf71pKKGRRPqpzbJa5XEh3"

def load_proxies(proxy_file="proxys.txt"):
    try:
        with open(proxy_file, 'r') as f:
            proxies = [line.strip() for line in f if line.strip()]
        print(f"Loaded {len(proxies)} proxies from file")
        return [requests_proxy.ProxyInfo(proxy=f"http://{proxy}") for proxy in proxies]
    except Exception as e:
        print(f"Error loading proxies: {e}")
        return []
        
def search_google_images(query, num_images=1):
    try:
        api_key = os.getenv('GOOGLE_API_KEY')
        cse_id = os.getenv('GOOGLE_CSE_ID')
        proxies = load_proxies()
        
        print(f"Buscando imágenes para: {query}")
        
        if not proxies:
            print("No proxies available, trying without proxy")
            with requests.Session() as session:
                result = session.get(
                    f"https://www.googleapis.com/customsearch/v1",
                    params={
                        "q": query,
                        "cx": cse_id,
                        "searchType": "image",
                        "num": num_images * 3,
                        "safe": 'off',
                        "imgSize": 'LARGE',
                        "imgType": 'imgTypeUndefined',
                        "rights": 'cc_publicdomain|cc_attribute|cc_sharealike',
                        "key": api_key
                    }
                ).json()
        else:
            with requests_proxy.ProxyManager(proxies) as proxy_manager:
                result = proxy_manager.get(
                    f"https://www.googleapis.com/customsearch/v1",
                    params={
                        "q": query,
                        "cx": cse_id,
                        "searchType": "image",
                        "num": num_images * 3,
                        "safe": 'off',
                        "imgSize": 'LARGE',
                        "imgType": 'imgTypeUndefined',
                        "rights": 'cc_publicdomain|cc_attribute|cc_sharealike',
                        "key": api_key
                    }
                ).json()
        
        if 'items' in result:
            image_urls = []
            for item in result.get('items', []):
                if 'image' in item:
                    width = int(item['image'].get('width', 0))
                    height = int(item['image'].get('height', 0))
                    if width >= MIN_WIDTH and height >= MIN_HEIGHT:
                        image_urls.append(item['link'])
                        if len(image_urls) >= num_images:
                            break
            
            print(f"Encontradas {len(image_urls)} imágenes de tamaño adecuado")
            return image_urls
        
        print("No se encontraron imágenes después de probar todos los proxies")
        return []
            
    except Exception as e:
        print(f"Error general en la búsqueda de imágenes: {str(e)}")
        return []

def process_image(image):
    try:
        width, height = image.size
        current_ratio = width / height
        
        if current_ratio > TARGET_ASPECT_RATIO:
            new_width = max(MIN_WIDTH, width)
            new_height = int(new_width / TARGET_ASPECT_RATIO)
        else:
            new_height = max(MIN_HEIGHT, height)
            new_width = int(new_height * TARGET_ASPECT_RATIO)
        
        image = image.resize((new_width, new_height), Image.Resampling.LANCZOS)
        background = Image.new('RGB', (max(new_width, MIN_WIDTH), max(new_height, MIN_HEIGHT)), 'black')
        
        offset = ((background.width - image.width) // 2,
                 (background.height - image.height) // 2)
        background.paste(image, offset)
        
        return background
    except Exception as e:
        print(f"Error processing image: {e}")
        return None

def download_image(url):
    proxies = load_proxies()
    if not proxies:
        proxies = [None]
    
    for proxy in proxies:
        try:
            response = requests.get(url, proxies=proxy, timeout=10)
            image = Image.open(BytesIO(response.content))
            
            if image.mode in ('RGBA', 'LA') or (image.mode == 'P' and 'transparency' in image.info):
                background = Image.new('RGB', image.size, (0, 0, 0))
                background.paste(image, mask=image.split()[-1])
                image = background
            
            processed_image = process_image(image)
            if processed_image:
                return processed_image
            
        except Exception as e:
            print(f"Error downloading image with proxy {proxy}: {e}")
            continue
    return None

def create_animated_clip(image, duration=5, zoom_factor=1.1):
    img_array = np.array(image)
    img_clip = ImageClip(img_array).set_duration(duration)
    
    if img_clip.size[0] < MIN_WIDTH or img_clip.size[1] < MIN_HEIGHT:
        img_clip = img_clip.resize(width=MIN_WIDTH, height=MIN_HEIGHT)
    
    return img_clip.resize(lambda t: 1 + (zoom_factor - 1) * t / duration)

def concatenate_google_images(keywords, clip_duration=5, num_images_per_keyword=1):
    keyword_list = [keyword.strip() for keyword in keywords.split(",") if keyword.strip()]
    if not keyword_list:
        keyword_list = ["nature"]
    
    video_clips = []
    for keyword in keyword_list:
        try:
            print(f"Searching images for keyword '{keyword}'...")
            image_urls = search_google_images(keyword, num_images=num_images_per_keyword)
            for url in image_urls:
                image = download_image(url)
                if image:
                    clip = create_animated_clip(image, duration=clip_duration)
                    video_clips.append(clip)
                    time.sleep(1)
        except Exception as e:
            print(f"Error processing keyword '{keyword}': {e}")
            continue
    
    if not video_clips:
        return ColorClip(size=(MIN_WIDTH, MIN_HEIGHT), color=[0, 0, 0], duration=5)
    
    random.shuffle(video_clips)
    return concatenate_videoclips(video_clips, method="compose")

def adjust_background_music(video_duration, music_file):
    try:
        music = AudioFileClip(music_file)
        if music.duration < video_duration:
            repetitions = int(video_duration / music.duration) + 1
            music_clips = [music] * repetitions
            music = concatenate_audioclips(music_clips)
        music = music.subclip(0, video_duration)
        return music.volumex(0.2)
    except Exception as e:
        print(f"Error adjusting music: {e}")
        return None

def combine_audio_video(audio_file, video_clip, music_clip=None):
    try:
        audio_clip = AudioFileClip(audio_file)
        total_duration = audio_clip.duration + 2
        
        video_clip = video_clip.loop(duration=total_duration)
        video_clip = video_clip.set_duration(total_duration).fadeout(2)
        
        final_clip = video_clip.set_audio(audio_clip)
        if music_clip:
            music_clip = music_clip.set_duration(total_duration).audio_fadeout(2)
            final_clip = final_clip.set_audio(CompositeAudioClip([audio_clip, music_clip]))
        
        output_filename = f"final_video_{int(time.time())}.mp4"
        output_path = os.path.join(output_folder, output_filename)
        
        final_clip.write_videofile(output_path, codec="libx264", audio_codec="aac", fps=24)
        
        final_clip.close()
        video_clip.close()
        audio_clip.close()
        if music_clip:
            music_clip.close()
            
        return output_path
    except Exception as e:
        print(f"Error combining audio and video: {e}")
        if 'final_clip' in locals():
            final_clip.close()
        return None

def upload_to_google_drive(file_path, folder_id):
    try:
        creds = service_account.Credentials.from_service_account_info(
            json.loads(os.getenv('GOOGLE_SERVICE_ACCOUNT')),
            scopes=['https://www.googleapis.com/auth/drive']
        )
        service = build('drive', 'v3', credentials=creds)

        file_metadata = {
            'name': os.path.basename(file_path),
            'parents': [folder_id]
        }
        media = MediaFileUpload(file_path, resumable=True)
        file = service.files().create(body=file_metadata, media_body=media, fields='id').execute()

        permission = {
            'type': 'anyone',
            'role': 'reader'
        }
        service.permissions().create(fileId=file['id'], body=permission).execute()

        file_id = file['id']
        download_link = f"https://drive.google.com/uc?export=download&id={file_id}"
        return download_link
    except Exception as e:
        print(f"Error uploading to Google Drive: {e}")
        return None

def process_input(text, txt_file, mp3_file, selected_voice, rate, pitch, keywords):
    try:
        if text.strip():
            final_text = text
        elif txt_file is not None:
            final_text = txt_file.decode("utf-8")
        else:
            raise ValueError("No text input provided")
        
        audio_file = asyncio.run(text_to_speech(final_text, selected_voice, rate, pitch))
        if not audio_file:
            raise ValueError("Failed to generate audio")
        
        video_clip = concatenate_google_images(keywords, clip_duration=5, num_images_per_keyword=1)
        if not video_clip:
            raise ValueError("Failed to generate video")
        
        music_clip = None
        if mp3_file is not None:
            music_clip = adjust_background_music(video_clip.duration, mp3_file.name)
        
        final_video_path = combine_audio_video(audio_file, video_clip, music_clip)
        if not final_video_path:
            raise ValueError("Failed to combine audio and video")
        
        download_link = upload_to_google_drive(final_video_path, folder_id=FOLDER_ID)
        if download_link:
            print(f"Video uploaded to Google Drive. Download link: {download_link}")
            return f"[Download video]({download_link})"
        else:
            raise ValueError("Error uploading video to Google Drive")
    except Exception as e:
        print(f"Error during processing: {e}")
        return None
    finally:
        cleanup_temp_files()

with gr.Blocks() as demo:
    gr.Markdown("# Text-to-Video Generator")
    with gr.Row():
        with gr.Column():
            text_input = gr.Textbox(label="Write your text here", lines=5)
            txt_file_input = gr.File(label="Or upload a .txt file", file_types=[".txt"])
            mp3_file_input = gr.File(label="Upload background music (.mp3)", file_types=[".mp3"])
            keyword_input = gr.Textbox(
                label="Enter keywords separated by commas",
                value="fear, religion, god, demons, aliens, possession, galaxy, mysterious, dystopian, astral, warfare, space, space, galaxy, moon, fear, astral, god, evil, mystery, cosmos, stars, paranormal, inexplicable, hidden, enigma, unknown, unusual, intriguing, curious, strange, supernatural, esoteric, arcane, occultism, supernatural, mystery, phenomenon, rare, unusual, enigmatic, sinister, gloomy, dark, shadowy, macabre, eerie, chilling, cursed, fantastic, unreal, unknown, mysterious, enigmatic, inexplicable, unusual, strange, unusual, arcane, esoteric, hidden, shadowy, dark, gloomy, sinister, macabre, eerie, chilling, cursed, fantastic, unreal, paranormal, supernatural, occultism, phenomenon, rare, intriguing, curious"
            )
            voices = asyncio.run(get_voices())
            voice_dropdown = gr.Dropdown(choices=list(voices.keys()), label="Select Voice")
            rate_slider = gr.Slider(minimum=-50, maximum=50, value=0, label="Speech Rate Adjustment (%)", step=1)
            pitch_slider = gr.Slider(minimum=-20, maximum=20, value=0, label="Pitch Adjustment (Hz)", step=1)
        with gr.Column():
            output_link = gr.Markdown("")

    btn = gr.Button("Generate Video")
    btn.click(
        process_input,
        inputs=[text_input, txt_file_input, mp3_file_input, voice_dropdown, rate_slider, pitch_slider, keyword_input],
        outputs=output_link
    )

port = int(os.getenv("PORT", 7860))
demo.launch(server_name="0.0.0.0", server_port=port, share=True, show_error=True)