Upload 7 files

animate.py

@@ -0,0 +1,18 @@
+from moviepy.editor import ImageClip, AudioFileClip
+
+def animate_scene(image_path, audio_path, idx):
+    audio = AudioFileClip(audio_path)
+    duration = audio.duration
+    
+    image_clip = (
+        ImageClip(image_path)
+        .set_duration(duration)
+        .set_audio(audio)
+        .resize(height=720)
+        .fx(lambda clip: clip.crossfadein(1).crossfadeout(1))
+        .zoom_in(1.05)
+    )
+
+    out_path = f"assets/video/scene_{idx}.mp4"
+    image_clip.write_videofile(out_path, fps=24, audio_codec="aac", verbose=False, logger=None)
+    return out_path

app.py

@@ -0,0 +1,46 @@
+import gradio as gr
+from scene_planner import plan_scenes
+from generate_image import generate_scene_image
+from tts import generate_audio
+from animate import animate_scene
+from compose import compose_video
+import os
+import shutil
+import spaces
+
+LANGUAGES = ["Gujarati", "Hindi", "English"]
+
+@spaces.GPU
+def full_pipeline(script, language):
+    scenes = plan_scenes(script)
+
+    if os.path.exists("assets"):
+        shutil.rmtree("assets")
+    os.makedirs("assets/images")
+    os.makedirs("assets/audio")
+    os.makedirs("assets/video")
+
+    video_segments = []
+    for idx, scene in enumerate(scenes):
+        prompt = scene['prompt']
+        line = scene['dialogue']
+
+        image_path = generate_scene_image(prompt, idx)
+        audio_path = generate_audio(line, idx, language)
+        video_path = animate_scene(image_path, audio_path, idx)
+        video_segments.append(video_path)
+
+    final_path = compose_video(video_segments)
+    return final_path
+
+with gr.Blocks() as demo:
+    gr.Markdown("# 🧙‍♂️ Script to Cartoon Video Generator (Gujarati | Hindi | English)")
+    with gr.Row():
+        script_input = gr.Textbox(label="Enter Story or Script", lines=10)
+        lang_input = gr.Dropdown(choices=LANGUAGES, label="Select Narration Language")
+    gen_btn = gr.Button("Generate Cartoon Video")
+    output_video = gr.Video(label="Final Video")
+
+    gen_btn.click(full_pipeline, inputs=[script_input, lang_input], outputs=output_video)
+
+demo.launch()

compose.py

@@ -0,0 +1,14 @@
+from moviepy.editor import concatenate_videoclips
+import os
+
+def compose_video(video_paths):
+    clips = []
+    for path in video_paths:
+        if os.path.exists(path):
+            from moviepy.editor import VideoFileClip
+            clips.append(VideoFileClip(path))
+
+    final = concatenate_videoclips(clips, method="compose")
+    out_path = "assets/final_video.mp4"
+    final.write_videofile(out_path, fps=24, audio_codec="aac", verbose=False, logger=None)
+    return out_path

generate_image.py

@@ -0,0 +1,17 @@
+from diffusers import StableDiffusionXLPipeline
+import torch
+from PIL import Image
+import os
+
+pipe = StableDiffusionXLPipeline.from_pretrained(
+    "stabilityai/sdxl-turbo",
+    torch_dtype=torch.float16,
+    variant="fp16"
+).to("cuda" if torch.cuda.is_available() else "cpu")
+pipe.safety_checker = None
+
+def generate_scene_image(prompt, idx):
+    image = pipe(prompt).images[0]
+    out_path = f"assets/images/scene_{idx}.png"
+    image.save(out_path)
+    return out_path

requirements.txt

@@ -0,0 +1,11 @@
+gradio
+transformers
+accelerate
+torch
+pillow
+moviepy
+ffmpeg-python
+diffusers
+bark
+pyttsx3
+spaces

scene_planner.py

@@ -0,0 +1,20 @@
+from transformers import pipeline
+
+scene_splitter = pipeline("text2text-generation", model="mistralai/Mistral-7B-Instruct-v0.1")
+
+def plan_scenes(script):
+    prompt = (
+        "Split the following story into 10-15 scenes. For each scene, return a JSON with 'prompt' "
+        "(scene image description) and 'dialogue' (narration to speak). Make the prompt cartoon-friendly.\nStory:\n" + script
+    )
+
+    response = scene_splitter(prompt, max_new_tokens=1024, do_sample=False)[0]['generated_text']
+
+    try:
+        import json
+        scenes = json.loads(response)
+    except:
+        scenes = [
+            {"prompt": line.strip(), "dialogue": line.strip()} for line in script.split(".") if line.strip()
+        ]
+    return scenes[:15]

tts.py

@@ -0,0 +1,16 @@
+import os
+from bark import generate_audio as bark_tts
+
+def generate_audio(text, idx, language):
+    out_path = f"assets/audio/scene_{idx}.wav"
+    try:
+        audio = bark_tts(text, lang=language.lower())
+        with open(out_path, "wb") as f:
+            f.write(audio)
+    except:
+        import pyttsx3
+        engine = pyttsx3.init()
+        engine.setProperty('rate', 150)
+        engine.save_to_file(text, out_path)
+        engine.runAndWait()
+    return out_path