app.py

import gradio as gr
from tts_module import get_voices, text_to_speech
from pexels_api import search_pexels
from moviepy.editor import (
    AudioFileClip, VideoFileClip, CompositeAudioClip,
    concatenate_audioclips, concatenate_videoclips, vfx, CompositeVideoClip
)
import asyncio
import os
import time
import requests
from googleapiclient.discovery import build
from googleapiclient.http import MediaFileUpload
import tempfile
import re
import random

# Define la carpeta de salida temporal
output_folder = "outputs"
os.makedirs(output_folder, exist_ok=True)

def clean_text_for_search(text):
    """Limpia el texto para hacer búsquedas válidas en Pexels"""
    # Eliminar caracteres especiales y limitar longitud
    text = re.sub(r'[^\w\s]', '', text).strip()
    return text

def resize_and_blur_video(clip, target_aspect_ratio=16/9):
    """
    Redimensiona y aplica desenfoque al fondo del video para mantener el aspecto 16:9.
    """
    try:
        w, h = clip.size
        current_aspect_ratio = w / h

        print(f"Procesando video: {w}x{h}, ratio: {current_aspect_ratio}")
        if abs(current_aspect_ratio - target_aspect_ratio) < 0.1:
            return clip

        if current_aspect_ratio < target_aspect_ratio:  # Video vertical
            target_w = int(h * target_aspect_ratio)
            target_h = h

            background = clip.resize(width=target_w)
            try:
                background = background.fx(vfx.blur, sigma=50)
            except Exception as e:
                print(f"Error al aplicar blur: {e}")

            foreground = clip.resize(height=target_h)
            x_center = (target_w - foreground.w) / 2
            return CompositeVideoClip(
                [background, foreground.set_position((x_center, 0))],
                size=(target_w, target_h)
            )
        else:  # Video horizontal
            return clip.resize(width=int(h * target_aspect_ratio), height=h)

    except Exception as e:
        print(f"Error en resize_and_blur_video: {e}")
        return clip

def concatenate_pexels_videos(keywords, num_videos_per_keyword=1):
    """
    Concatena videos de Pexels basados en palabras clave proporcionadas por el usuario.
    :param keywords: Palabras clave separadas por comas (ejemplo: "universo, galaxia, bosque, gato").
    """
    keyword_list = [keyword.strip() for keyword in keywords.split(",") if keyword.strip()]
    if not keyword_list:
        raise Exception("No se proporcionaron palabras clave válidas.")

    video_clips = []

    for keyword in keyword_list:
        try:
            print(f"Buscando videos para la palabra clave '{keyword}'...")
            links = search_pexels(keyword, num_results=num_videos_per_keyword)
            if not links:
                print(f"No se encontraron videos para la palabra clave '{keyword}'.")
                continue

            link = links[0]  # Usamos solo el primer video encontrado
            video_response = requests.get(link)
            if video_response.status_code != 200:
                print(f"Error al descargar video desde {link}: Código de estado {video_response.status_code}")
                continue

            with tempfile.NamedTemporaryFile(delete=False, suffix=".mp4") as tmp_video:
                tmp_video.write(video_response.content)
                clip = VideoFileClip(tmp_video.name)
                processed_clip = resize_and_blur_video(clip)
                video_clips.append(processed_clip)

        except Exception as e:
            print(f"Error procesando palabra clave '{keyword}': {e}")
            continue

    if not video_clips:
        raise Exception("No se pudieron obtener videos válidos.")

    # Aleatorizar el orden de los clips si es necesario
    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)
        if music.duration > video_duration:
            music = music.subclip(0, video_duration)
        music = music.volumex(0.2)
        return music
    except Exception as e:
        print(f"Error ajustando música: {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 + 5
        if video_clip.duration < total_duration:
            video_clip = video_clip.loop(duration=total_duration)
        video_clip = video_clip.set_duration(total_duration).fadeout(5)
        final_clip = video_clip.set_audio(audio_clip)

        if music_clip:
            if music_clip.duration < total_duration:
                repetitions = int(total_duration / music_clip.duration) + 1
                music_clips = [music_clip] * repetitions
                music_clip = concatenate_audioclips(music_clips)
            if music_clip.duration > total_duration:
                music_clip = music_clip.subclip(0, total_duration)
            music_clip = music_clip.audio_fadeout(5)
            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)
        return output_path

    except Exception as e:
        print(f"Error combinando audio y video: {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:
            return "No input provided"

        voices = asyncio.run(get_voices())
        if selected_voice not in voices:
            return f"La voz '{selected_voice}' no es válida. Por favor, seleccione una de las siguientes voces: {', '.join(voices.keys())}"

        try:
            audio_file = asyncio.run(text_to_speech(final_text, selected_voice, rate, pitch))
        except Exception as e:
            return f"Error generando audio: {e}"

        try:
            video_clip = concatenate_pexels_videos(keywords, num_videos_per_keyword=1)
        except Exception as e:
            return f"Error concatenando videos: {e}"

        if mp3_file is not None:
            music_clip = adjust_background_music(video_clip.duration, mp3_file.name)
        else:
            music_clip = None

        final_video_path = combine_audio_video(audio_file, video_clip, music_clip)
        upload_to_google_drive(final_video_path)
        return final_video_path

    except Exception as e:
        return f"Error durante el procesamiento: {e}"

def upload_to_google_drive(file_path):
    try:
        api_key = os.getenv("GOOGLE_API_KEY")
        if not api_key:
            print("Error: GOOGLE_API_KEY no está definida en las variables de entorno.")
            return None

        service = build("drive", "v3", developerKey=api_key)
        file_metadata = {"name": os.path.basename(file_path)}
        media = MediaFileUpload(file_path, resumable=True)
        file = service.files().create(body=file_metadata, media_body=media, fields="id").execute()
        print(f"Archivo subido exitosamente con ID: {file.get('id')}")
        return file.get("id")
    except Exception as e:
        print(f"Error subiendo a Google Drive: {e}")
        return None

# Interfaz Gradio
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 (e.g., universe, galaxy, forest, cat)")
            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_video = gr.File(label="Download Generated Video")

    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_video
    )

# Leer el puerto asignado por Hugging Face
port = int(os.getenv("PORT", 7860))

# Lanzar la aplicación
demo.launch(server_name="0.0.0.0", server_port=port, share=True)