app.py

import gradio as gr
from transformers import pipeline
from diffusers import StableDiffusionPipeline
import torch
from PIL import Image, ImageDraw, ImageFont
import scipy.io.wavfile
from TTS.api import TTS  # Coqui TTS (open source)
from moviepy.editor import CompositeVideoClip, ImageClip, AudioFileClip, concatenate_videoclips
import os
import subprocess
from moviepy.editor import VideoFileClip, AudioFileClip


# Initialize Clients
# Replace with your actual API keys or methods of getting them
# groq_client = Groq(api_key=os.environ.get("GROQ_API_KEY"))
# deepgram_client = Deepgram(api_key=os.environ.get("DEEGRAM_API_KEY"))

# Use DistilGPT-2 for text generation
script_generator = pipeline("text-generation", model="distilgpt2", truncation=True, max_length=100)

# Use Coqui TTS for text-to-speech
tts = TTS(model_name="tts_models/en/ljspeech/tacotron2-DDC", progress_bar=False, gpu=False)

# Use MusicGen for music generation
music_generator = pipeline("text-to-audio", model="facebook/musicgen-small", device="cpu")

# Use Fluently Anime (Stable Diffusion) for anime image generation
model_id = "fluently/Fluently-anime"
anime_image_generator = StableDiffusionPipeline.from_pretrained(model_id).to("cpu")


# Generate Comedy Script using DistilGPT-2
def generate_comedy_script(prompt):
    script = script_generator(prompt)[0]['generated_text']
    return script

# Convert Text to Speech using Coqui TTS
def text_to_speech(script):
    output_audio = 'output.wav'
    tts.tts_to_file(text=script, file_path=output_audio)
    return output_audio

# Create Anime Images Using Fluently Anime
def create_images_from_script(script):
    lines = script.split('. ')
    image_paths = []
    for i, line in enumerate(lines):
        img = anime_image_generator(line).images[0]
        img_path = f'/tmp/anime_image_{i}.png'
        img.save(img_path)
        image_paths.append(img_path)
    return image_paths

# Generate Fun Music Track using MusicGen
def generate_fun_music(prompt, output_music_file="fun_music.wav"):
    response = music_generator(prompt)
    audio_data = response["audio"]
    sampling_rate = response["sampling_rate"]
    scipy.io.wavfile.write(output_music_file, rate=sampling_rate, data=audio_data)
    return output_music_file

# Create Video from Generated Anime Images Using FFmpeg
def generate_text_video(script):
    image_paths = create_images_from_script(script)
    # Generate video using moviepy from the sequence of images
    video_clip = ImageSequenceClip(image_paths, fps=24)
    video_path = "/tmp/final_video.mp4"
    video_clip.write_videofile(video_path, codec='libx264')
    return video_path

def combine_audio_video(video_file, audio_file):
    # Load the video file
    video = VideoFileClip(video_file)
    # Load the audio file
    audio = AudioFileClip(audio_file)
    
    # Set the audio of the video clip
    final_video = video.set_audio(audio)

    # Return the final video clip
    return final_video

# Main Function to Generate Comedy Animation
def generate_comedy_and_animation(prompt):
    script = generate_comedy_script(prompt)
    audio_file = text_to_speech(script)
    video_file = generate_text_video(script)
    fun_music = generate_fun_music(prompt)
    final_video = combine_audio_video(video_file, fun_music)
    return script, audio_file, final_video

# Generate Kids Content
def generate_kids_content(theme):
    music_file = generate_fun_music(theme, output_music_file="kids_music.wav")
    clips = []
    for i in range(5):
        img = Image.new('RGB', (800, 400), color=(0, 0, 255))
        d = ImageDraw.Draw(img)
        fnt = ImageFont.load_default()
        d.text((10, 180), f"Kids Music: {theme}", font=fnt, fill=(255, 255, 0))
        frame_path = f'/tmp/kids_temp_{i}.png'
        img.save(frame_path)
        clips.append(ImageClip(frame_path).set_duration(1).set_position(('center', 'center')))
    final_video = concatenate_videoclips(clips, method="compose").set_audio(AudioFileClip(music_file))
    final_video.write_videofile("/tmp/kids_animation.mp4", fps=24)
    return music_file, "/tmp/kids_animation.mp4"

# Gradio Interface
with gr.Blocks() as app:
    gr.Markdown("## AI Comedy and Kids Content Generator")

    # Comedy Animation Tab
    with gr.Tab("Generate Comedy Animation"):
        prompt_input = gr.Textbox(label="Comedy Prompt")
        generate_btn = gr.Button("Generate Comedy Script and Animation")
        comedy_script = gr.Textbox(label="Generated Script")
        comedy_audio = gr.Audio(label="Generated Audio")
        comedy_video = gr.Video(label="Generated Animation")

        generate_btn.click(
            generate_comedy_and_animation,
            inputs=prompt_input,
            outputs=[comedy_script, comedy_audio, comedy_video]
        )

    # Kids Music Animation Tab
    with gr.Tab("Generate Kids Music Animation"):
        theme_input = gr.Textbox(label="Kids Music Theme")
        generate_music_btn = gr.Button("Generate Kids Music and Animation")
        kids_music_audio = gr.Audio(label="Generated Music")
        kids_music_video = gr.Video(label="Generated Kids Animation")

        generate_music_btn.click(
            generate_kids_content,
            inputs=theme_input,
            outputs=[kids_music_audio, kids_music_video]
        )

app.launch()