Update app.py

app.py

@@ -11,6 +11,7 @@ import markdown
 import threading
 from queue import Queue
 import time
+import uuid
 
 app = Flask(__name__)
 
@@ -54,52 +55,48 @@ def get_ollama_models(base_host=DEFAULT_BASE_HOST):
     except:
         return []
 
-def load_chat_history(chat_file):
+def load_chat_history(session_id):
     try:
-        with open(os.path.join(chats_folder, chat_file), 'r', encoding='utf-8') as f:
+        with open(os.path.join(chats_folder, f'{session_id}.json'), 'r', encoding='utf-8') as f:
             return json.load(f).get('messages', [])
     except:
         return []
 
-def save_chat(messages, chat_file=None):
+def save_chat(session_id, messages):
     if not messages:
         return None
-    
-    if not chat_file:
-        timestamp = datetime.now().strftime('%Y%m%d-%H%M%S')
-        chat_file = f'chat-{timestamp}.json'
-    
-    filepath = os.path.join(chats_folder, chat_file)
+
+    filepath = os.path.join(chats_folder, f'{session_id}.json')
     with open(filepath, 'w', encoding='utf-8') as f:
         json.dump({
             'timestamp': datetime.now().strftime('%Y-%m-%d %H:%M:%S'),
             'messages': messages
         }, f, ensure_ascii=False, indent=2)
-    
-    return chat_file
+
+    return session_id
 
 def remove_markdown(text):
     html_content = markdown.markdown(text)
     soup = BeautifulSoup(html_content, 'html.parser')
     return soup.get_text().strip()
-    
+
 def convert_to_speech(text, model_name, remove_md=False):
     if model_name not in get_available_models():
         return None
-        
+
     if remove_md:
         text = remove_markdown(text)
-    
+
     random_name = ''.join(random.choices(string.ascii_letters + string.digits, k=8)) + '.wav'
     output_file = os.path.join(temp_audio_folder, random_name)
-    
+
     # Clean old audio files
     for file in os.listdir(temp_audio_folder):
         if file.endswith('.wav'):
             os.remove(os.path.join(temp_audio_folder, file))
-    
+
     model_path = os.path.join(model_folder, model_name + '.onnx')
-    
+
     try:
         # Use the text directly in the command
         command = (
@@ -108,69 +105,69 @@ def convert_to_speech(text, model_name, remove_md=False):
             f'--length_scale {SETTINGS["length_scale"]} --noise_w {SETTINGS["noise_w"]} '
             f'--sentence_silence {SETTINGS["sentence_silence"]}'
         )
-        
+
         # Pass the text as input to the command
         result = subprocess.run(command, input=text.encode('utf-8'), shell=True, check=True)
-        
+
         if os.path.exists(output_file):
             return output_file
     except Exception as e:
         print(f"Error during text-to-speech conversion: {e}")
-    
+
     return None
 
 def set_default_models():
     tts_models = get_available_models()
     ollama_models = get_ollama_models()
-    
+
     default_tts_model = "RecomendacionesConMiau" if "RecomendacionesConMiau" in tts_models else None
     default_ollama_model = "llama3.2:1b" if "llama3.2:1b" in ollama_models else None
-    
+
     return default_tts_model, default_ollama_model
 
 @app.route('/')
 def index():
     tts_models = get_available_models()
-    chat_files = sorted([f for f in os.listdir(chats_folder) if f.endswith('.json')], reverse=True)
+    session_id = str(uuid.uuid4())
     default_tts_model, default_ollama_model = set_default_models()
-    return render_template('index.html', tts_models=tts_models, chat_files=chat_files, default_tts_model=default_tts_model, default_ollama_model=default_ollama_model)
+    return render_template('index.html', tts_models=tts_models, session_id=session_id, default_tts_model=default_tts_model, default_ollama_model=default_ollama_model)
 
 @app.route('/api/list_ollama_models')
 def list_ollama_models():
     base_host = request.args.get('base_host', DEFAULT_BASE_HOST)
     return jsonify(models=get_ollama_models(base_host))
 
-@app.route('/api/load_chat/<chat_file>')
-def load_chat(chat_file):
-    messages = load_chat_history(chat_file)
+@app.route('/api/load_chat/<session_id>')
+def load_chat(session_id):
+    messages = load_chat_history(session_id)
     return jsonify(messages=messages)
 
 @app.route('/api/update_message', methods=['POST'])
 def update_message():
     data = request.json
-    chat_file = data.get('chat_file')
+    session_id = data.get('session_id')
     message_index = data.get('message_index')
     new_content = data.get('content')
     is_user = data.get('is_user', False)
-    
-    if not chat_file or message_index is None or not new_content:
+
+    if not session_id or message_index is None or not new_content:
         return jsonify(error="Missing required parameters"), 400
-    
-    messages = load_chat_history(chat_file)
+
+    messages = load_chat_history(session_id)
     if message_index >= len(messages):
         return jsonify(error="Invalid message index"), 400
-    
+
     # Update the message content
     messages[message_index]['content'] = new_content
-    
+
     # If it's a user message, regenerate all subsequent responses
     if is_user:
         # Keep messages up to and including the edited message
         messages = messages[:message_index + 1]
-    
+
     # Save the updated chat
-    save_chat(messages, chat_file)
-    
+    save_chat(session_id, messages)
+
     return jsonify(success=True, messages=messages)
 
 @app.route('/api/chat', methods=['POST'])
@@ -179,8 +176,8 @@ def chat():
     base_host = data.get('base_host', DEFAULT_BASE_HOST)
     model = data.get('model')
     messages = data.get('messages', [])
-    chat_file = data.get('chat_file')
-    
+    session_id = data.get('session_id')
+
     def generate():
         queue = Queue()
         thread = threading.Thread(
@@ -188,7 +185,7 @@ def chat():
             args=(base_host, model, messages, queue)
         )
         thread.start()
-        
+
         complete_response = ""
         while True:
             msg_type, content = queue.get()
@@ -201,10 +198,10 @@ def chat():
             elif msg_type == "done":
                 # Save chat history
                 messages.append({"role": "assistant", "content": complete_response})
-                save_chat(messages, chat_file)
+                save_chat(session_id, messages)
                 yield f"data: {json.dumps({'done': complete_response})}\n\n"
                 break
-    
+
     return Response(stream_with_context(generate()), mimetype='text/event-stream')
 
 def stream_ollama_response(base_host, model, messages, queue):
@@ -214,7 +211,7 @@ def stream_ollama_response(base_host, model, messages, queue):
         "messages": messages,
         "stream": True
     }
-    
+
     try:
         with requests.post(url, json=data, stream=True) as response:
             if response.status_code == 200:
@@ -241,14 +238,14 @@ def text_to_speech():
     text = data.get('text', '')
     model = data.get('model')
     remove_md = data.get('remove_markdown', False)
-    
+
     if not text or not model:
         return jsonify(error="Missing text or model"), 400
-    
+
     audio_file = convert_to_speech(text, model, remove_md)
     if not audio_file:
         return jsonify(error="Failed to convert text to speech"), 500
-    
+
     return jsonify(audio_file=os.path.basename(audio_file))
 
 @app.route('/audio/<filename>')
@@ -256,4 +253,4 @@ def serve_audio(filename):
     return send_file(os.path.join(temp_audio_folder, filename))
 
 if __name__ == '__main__':
-    app.run(debug=True, port=7860, host='0.0.0.0')
+    app.run(debug=True, port=7860, host='0.0.0.0')