upload

app.py

@@ -1,13 +1,120 @@
 import gradio as gr
 from transformers import pipeline
+import requests
+import random
+import html
+import base64
+import os
+import json
+from datetime import datetime
+from moviepy.editor import VideoFileClip, AudioFileClip, CompositeVideoClip, TextClip
 
-chatbot = pipeline("text-generation", model="mistralai/Mistral-7B-Instruct-v0.2")
+chat_gen = pipeline("text-generation", model="mistralai/Mistral-7B-Instruct-v0.2")
 
-def generate_chat(prompt):
-    response = chatbot(f"Create a fake chat conversation based on this prompt: {prompt}", max_new_tokens=300)
-    return response[0]["generated_text"]
+# Static voice options
+voice_options = {
+    "Rachel": "EXAVITQu4vr4xnSDxMaL",
+    "Adam": "21m00Tcm4TlvDq8ikWAM",
+    "Elli": "AZnzlk1XvdvUeBnXmlld",
+    "Josh": "VR6AewLTigWG4xSOukaG"
+}
 
-iface = gr.Interface(fn=generate_chat, inputs="text", outputs="text", title="Fake Chat AI")
+library_file = "/tmp/library.json"
 
-if __name__ == "__main__":
-    iface.launch()
+def generate_with_voices(prompt, video, character_inputs):
+    try:
+        characters = [c.split(":")[0].strip() for c in character_inputs if ":" in c]
+        voices = {c.split(":")[0].strip(): c.split(":")[1].strip() for c in character_inputs if ":" in c}
+
+        system_prompt = (
+            f"Group chat between {', '.join(characters)}. Prompt: {prompt}. "
+            "Use casual format like: Name: message"
+        )
+        response = chat_gen(system_prompt, max_new_tokens=400)[0]["generated_text"]
+        chat_lines = [(line.split(':', 1)[0].strip(), line.split(':', 1)[1].strip()) for line in response.splitlines() if ':' in line]
+
+        audio_paths = []
+        chat_overlay_text = []
+
+        for i, (name, msg) in enumerate(chat_lines):
+            voice_id = voices.get(name, list(voice_options.values())[0])
+            eleven_response = requests.post(
+                f"https://api.elevenlabs.io/v1/text-to-speech/{voice_id}",
+                headers={
+                    "xi-api-key": os.environ.get("ELEVEN_API_KEY", "demo-key"),
+                    "Content-Type": "application/json"
+                },
+                json={ "text": msg, "model_id": "eleven_monolingual_v1" }
+            )
+            if eleven_response.status_code == 200:
+                file_path = f"/tmp/audio_{i}.mp3"
+                with open(file_path, 'wb') as f:
+                    f.write(eleven_response.content)
+                audio_paths.append(file_path)
+
+            chat_overlay_text.append(f"{name}: {msg}")
+
+        final_audio = AudioFileClip(audio_paths[0])
+        for a in audio_paths[1:]:
+            final_audio = final_audio.append(AudioFileClip(a))
+
+        video_clip = VideoFileClip(video.name).resize(height=480)
+        video_clip = video_clip.set_audio(final_audio)
+
+        txt = "\n".join(chat_overlay_text)
+        txt_clip = TextClip(txt, fontsize=24, color='black', bg_color='white', font="Courier").set_duration(video_clip.duration)
+        txt_clip = txt_clip.set_position((20, 20))
+
+        final = CompositeVideoClip([video_clip, txt_clip])
+        timestamp = datetime.now().strftime('%Y%m%d-%H%M%S')
+        out_path = f"/tmp/fakechat_{timestamp}.mp4"
+        final.write_videofile(out_path, codec="libx264", audio_codec="aac")
+
+        # Save to library
+        if os.path.exists(library_file):
+            with open(library_file, 'r') as f:
+                library = json.load(f)
+        else:
+            library = []
+
+        library.insert(0, {
+            "timestamp": timestamp,
+            "characters": characters,
+            "prompt": prompt,
+            "path": out_path
+        })
+
+        with open(library_file, 'w') as f:
+            json.dump(library, f, indent=2)
+
+        return out_path
+    except Exception as e:
+        return f"Error: {str(e)}"
+
+def get_library():
+    if not os.path.exists(library_file):
+        return "No stories yet."
+    with open(library_file, 'r') as f:
+        library = json.load(f)
+    html_items = [f"<li><a href='file={item['path']}' target='_blank'>{item['timestamp']} — {', '.join(item['characters'])}</a></li>" for item in library]
+    return f"<ul>{''.join(html_items)}</ul>"
+
+def build_voice_inputs():
+    return [gr.Textbox(label=f"Character {i+1} (Format: Name:Voice)", placeholder="e.g. Anna:Rachel") for i in range(4)]
+
+with gr.Blocks() as app:
+    with gr.Row():
+        with gr.Column(scale=3):
+            character_inputs = build_voice_inputs()
+            prompt = gr.Textbox(label="Scene Prompt")
+            video = gr.File(label="Upload Background Video (.mp4)", file_types=[".mp4"])
+            generate_btn = gr.Button("Generate Story")
+            output = gr.File(label="Download MP4 Story")
+        with gr.Column(scale=1):
+            library_html = gr.HTML(label="My Saved Stories")
+            refresh_btn = gr.Button("🔁 Refresh Library")
+
+    generate_btn.click(fn=generate_with_voices, inputs=[prompt, video, character_inputs], outputs=output)
+    refresh_btn.click(fn=get_library, inputs=[], outputs=library_html)
+
+app.launch()

requirements.txt

@@ -1,3 +1,5 @@
 transformers
 gradio
-torch
+torch
+requests
+moviepy