hello

app.py

@@ -1,17 +1,15 @@
 import gradio as gr
 from transformers import pipeline
 import requests
-import random
+import os
 import html
 import base64
-import os
+import random
 import json
 from datetime import datetime
-from moviepy.editor import VideoFileClip, AudioFileClip, CompositeVideoClip, TextClip
 
 chat_gen = pipeline("text-generation", model="mistralai/Mistral-7B-Instruct-v0.2")
 
-# Static voice options
 voice_options = {
     "Rachel": "EXAVITQu4vr4xnSDxMaL",
     "Adam": "21m00Tcm4TlvDq8ikWAM",
@@ -21,100 +19,93 @@ voice_options = {
 
 library_file = "/tmp/library.json"
 
-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"
+def generate_chat_bundle(prompt, character_inputs):
+    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 texting style: 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]
+
+    html_chat = ""
+    file_paths = []
+
+    for i, (name, msg) in enumerate(chat_lines):
+        color = random.choice(["#DCF8C6", "#FEE2E2", "#EDEDED", "#C7E2FF"])
+        html_chat += f"""<div style='margin: 10px; text-align: {'left' if i % 2 == 0 else 'right'}'>
+        <div style='display:inline-block; padding: 10px 15px; background: {color}; border-radius: 15px; max-width: 70%;'>
+        <strong>{html.escape(name)}:</strong> {html.escape(msg)}
+        </div></div>"""
+
+        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"}
         )
-        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)}"
+
+        if eleven_response.status_code == 200:
+            file_path = f"/tmp/audio_{i}_{name}.mp3"
+            with open(file_path, 'wb') as f:
+                f.write(eleven_response.content)
+            file_paths.append(file_path)
+
+    # Save HTML for download
+    timestamp = datetime.now().strftime('%Y%m%d-%H%M%S')
+    html_path = f"/tmp/fakechat_{timestamp}.html"
+    with open(html_path, "w") as f:
+        f.write(f"<html><body>{html_chat}</body></html>")
+    file_paths.append(html_path)
+
+    # Save metadata
+    if os.path.exists(library_file):
+        with open(library_file, "r") as f:
+            lib = json.load(f)
+    else:
+        lib = []
+
+    lib.insert(0, {
+        "timestamp": timestamp,
+        "characters": characters,
+        "prompt": prompt,
+        "html": html_path,
+        "audio_files": file_paths[:-1]
+    })
+
+    with open(library_file, "w") as f:
+        json.dump(lib, f, indent=2)
+
+    return file_paths
 
 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>"
+        return "No stories saved yet."
+    with open(library_file, "r") as f:
+        lib = json.load(f)
+    html_links = [f"<li><b>{item['timestamp']}</b>: {', '.join(item['characters'])} — <a href='file={item['html']}' target='_blank'>Chat</a></li>" for item in lib]
+    return "<ul>" + "".join(html_links) + "</ul>"
 
-def build_voice_inputs():
+def build_character_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()
+            characters = build_character_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")
+            generate_btn = gr.Button("🎤 Generate Chat + Audio")
+            file_output = gr.File(file_types=[".html", ".mp3"], label="Download Files")
         with gr.Column(scale=1):
-            library_html = gr.HTML(label="My Saved Stories")
+            lib_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)
+    generate_btn.click(fn=generate_chat_bundle, inputs=[prompt, characters], outputs=file_output)
+    refresh_btn.click(fn=get_library, outputs=lib_html)
 
 app.launch()

requirements.txt

@@ -1,5 +1,4 @@
 transformers
 gradio
 torch
-requests
-moviepy
+requests