Update app.py

app.py

@@ -1,608 +1,3 @@
-# import subprocess
-# import sys
-
-# def install_parler_tts():
-#     subprocess.check_call([sys.executable, "-m", "pip", "install", "git+https://github.com/huggingface/parler-tts.git"])
-
-# # Call the function to install parler-tts
-# install_parler_tts()
-
-
-# import gradio as gr
-# import requests
-# import os
-# import time
-# import re
-# import logging
-# import tempfile
-# import folium
-# import concurrent.futures
-# import torch
-# from PIL import Image
-# from datetime import datetime
-# from transformers import pipeline, AutoModelForSpeechSeq2Seq, AutoProcessor
-# from googlemaps import Client as GoogleMapsClient
-# from gtts import gTTS
-# from diffusers import StableDiffusion3Pipeline
-# import soundfile as sf
-
-# from langchain_openai import OpenAIEmbeddings, ChatOpenAI
-# from langchain_pinecone import PineconeVectorStore
-# from langchain.prompts import PromptTemplate
-# from langchain.chains import RetrievalQA
-# from langchain.chains.conversation.memory import ConversationBufferWindowMemory
-# from langchain.agents import Tool, initialize_agent
-# from huggingface_hub import login
-
-# # Check if the token is already set in the environment variables
-# hf_token = os.getenv("HF_TOKEN")
-
-# if hf_token is None:
-#     # If the token is not set, prompt for it (this should be done securely)
-#     print("Please set your Hugging Face token in the environment variables.")
-# else:
-#     # Login using the token
-#     login(token=hf_token)
-
-# # Your application logic goes here
-# print("Logged in successfully to Hugging Face Hub!")
-
-# # Set up logging
-# logging.basicConfig(level=logging.DEBUG)
-
-# # Initialize OpenAI embeddings
-# embeddings = OpenAIEmbeddings(api_key=os.environ['OPENAI_API_KEY'])
-
-# # Initialize Pinecone
-# from pinecone import Pinecone
-# pc = Pinecone(api_key=os.environ['PINECONE_API_KEY'])
-
-# index_name = "birmingham-dataset"
-# vectorstore = PineconeVectorStore(index_name=index_name, embedding=embeddings)
-# retriever = vectorstore.as_retriever(search_kwargs={'k': 5})
-
-# # Initialize ChatOpenAI model
-# chat_model = ChatOpenAI(api_key=os.environ['OPENAI_API_KEY'],
-#                         temperature=0, model='gpt-4o')
-
-# conversational_memory = ConversationBufferWindowMemory(
-#     memory_key='chat_history',
-#     k=10,
-#     return_messages=True
-# )
-
-# def get_current_time_and_date():
-#     now = datetime.now()
-#     return now.strftime("%Y-%m-%d %H:%M:%S")
-
-# # Example usage
-# current_time_and_date = get_current_time_and_date()
-
-# def fetch_local_events():
-#     api_key = os.environ['SERP_API']
-#     url = f'https://serpapi.com/search.json?engine=google_events&q=Events+in+Birmingham&hl=en&gl=us&api_key={api_key}'
-
-#     response = requests.get(url)
-#     if response.status_code == 200:
-#         events_results = response.json().get("events_results", [])
-#         events_html = """
-#         <h2 style="font-family: 'Georgia', serif; color: #ff0000; background-color: #f8f8f8; padding: 10px; border-radius: 10px;">Local Events</h2>
-#         <style>
-#             .event-item {
-#                 font-family: 'Verdana', sans-serif;
-#                 color: #333;
-#                 margin-bottom: 15px;
-#                 padding: 10px;
-#                 font-weight: bold;
-#             }
-#             .event-item a {
-#                 color: #1E90FF;
-#                 text-decoration: none;
-#             }
-#             .event-item a:hover {
-#                 text-decoration: underline;
-#             }
-#         </style>
-#         """
-#         for index, event in enumerate(events_results):
-#             title = event.get("title", "No title")
-#             date = event.get("date", "No date")
-#             location = event.get("address", "No location")
-#             link = event.get("link", "#")
-#             events_html += f"""
-#             <div class="event-item">
-#                 <a href='{link}' target='_blank'>{index + 1}. {title}</a>
-#                 <p>Date: {date}<br>Location: {location}</p>
-#             </div>
-#             """
-#         return events_html
-#     else:
-#         return "<p>Failed to fetch local events</p>"
-
-# def fetch_local_weather():
-#     try:
-#         api_key = os.environ['WEATHER_API']
-#         url = f'https://weather.visualcrossing.com/VisualCrossingWebServices/rest/services/timeline/birmingham?unitGroup=metric&include=events%2Calerts%2Chours%2Cdays%2Ccurrent&key={api_key}'
-#         response = requests.get(url)
-#         response.raise_for_status()
-#         jsonData = response.json()
-        
-#         current_conditions = jsonData.get("currentConditions", {})
-#         temp_celsius = current_conditions.get("temp", "N/A")
-        
-#         if temp_celsius != "N/A":
-#             temp_fahrenheit = int((temp_celsius * 9/5) + 32)
-#         else:
-#             temp_fahrenheit = "N/A"
-            
-#         condition = current_conditions.get("conditions", "N/A")
-#         humidity = current_conditions.get("humidity", "N/A")
-
-#         weather_html = f"""
-#         <div class="weather-theme">
-#             <h2 style="font-family: 'Georgia', serif; color: #ff0000; background-color: #f8f8f8; padding: 10px; border-radius: 10px;">Local Weather</h2>
-#             <div class="weather-content">
-#                 <div class="weather-icon">
-#                     <img src="https://www.weatherbit.io/static/img/icons/{get_weather_icon(condition)}.png" alt="{condition}" style="width: 100px; height: 100px;">
-#                 </div>
-#                 <div class="weather-details">
-#                     <p style="font-family: 'Verdana', sans-serif; color: #333; font-size: 1.2em;">Temperature: {temp_fahrenheit}°F</p>
-#                     <p style="font-family: 'Verdana', sans-serif; color: #333; font-size: 1.2em;">Condition: {condition}</p>
-#                     <p style="font-family: 'Verdana', sans-serif; color: #333; font-size: 1.2em;">Humidity: {humidity}%</p>
-#                 </div>
-#             </div>
-#         </div>
-#         <style>
-#             .weather-theme {{
-#                 animation: backgroundAnimation 10s infinite alternate;
-#                 border-radius: 10px;
-#                 padding: 10px;
-#                 margin-bottom: 15px;
-#                 background: linear-gradient(45deg, #ffcc33, #ff6666, #ffcc33, #ff6666);
-#                 background-size: 400% 400%;
-#                 box-shadow: 0 4px 8px rgba(0, 0, 0, 0.1);
-#                 transition: box-shadow 0.3s ease, background-color 0.3s ease;
-#             }}
-#             .weather-theme:hover {{
-#                 box-shadow: 0 8px 16px rgba(0, 0, 0, 0.2);
-#                 background-position: 100% 100%;
-#             }}
-#             @keyframes backgroundAnimation {{
-#                 0% {{ background-position: 0% 50%; }}
-#                 100% {{ background-position: 100% 50%; }}
-#             }}
-#             .weather-content {{
-#                 display: flex;
-#                 align-items: center;
-#             }}
-#             .weather-icon {{
-#                 flex: 1;
-#             }}
-#             .weather-details {{
-#                 flex: 3;
-#             }}
-#         </style>
-#         """
-#         return weather_html
-#     except requests.exceptions.RequestException as e:
-#         return f"<p>Failed to fetch local weather: {e}</p>"
-
-# def get_weather_icon(condition):
-#     condition_map = {
-#         "Clear": "c01d",
-#         "Partly Cloudy": "c02d",
-#         "Cloudy": "c03d",
-#         "Overcast": "c04d",
-#         "Mist": "a01d",
-#         "Patchy rain possible": "r01d",
-#         "Light rain": "r02d",
-#         "Moderate rain": "r03d",
-#         "Heavy rain": "r04d",
-#         "Snow": "s01d",
-#         "Thunderstorm": "t01d",
-#         "Fog": "a05d",
-#     }
-#     return condition_map.get(condition, "c04d")
-
-# # Update prompt templates to include fetched details
-
-# current_time_and_date = get_current_time_and_date()
-
-# # Define prompt templates
-# template1 = """You are an expert concierge who is helpful and a renowned guide for Birmingham,Alabama. Based on weather being a sunny bright day and the today's date is 1st july 2024, use the following pieces of context, 
-# memory, and message history, along with your knowledge of perennial events in Birmingham,Alabama, to answer the question at the end. If you don't know the answer, just say "Homie, I need to get more data for this," and don't try to make up an answer. 
-# Use fifteen sentences maximum. Keep the answer as detailed as possible. Always include the address, time, date, and
-# event type and description. Always say "It was my pleasure!" at the end of the answer.
-# {context}
-# Question: {question}
-# Helpful Answer:"""
-
-# template2 = """You are an expert concierge who is helpful and a renowned guide for Birmingham,Alabama. Based on today's weather being a sunny bright day and today's date is 1st july 2024, take the location or address but don't show the location or address on the output prompts. Use the following pieces of context, 
-# memory, and message history, along with your knowledge of perennial events in Birmingham,Alabama, to answer the question at the end. If you don't know the answer, just say "Homie, I need to get more data for this," and don't try to make up an answer. 
-# Keep the answer short and sweet and crisp. Always say "It was my pleasure!" at the end of the answer.
-# {context}
-# Question: {question}
-# Helpful Answer:"""
-
-# QA_CHAIN_PROMPT_1 = PromptTemplate(input_variables=["context", "question"], template=template1)
-# QA_CHAIN_PROMPT_2 = PromptTemplate(input_variables=["context", "question"], template=template2)
-
-# # Define the retrieval QA chain
-# def build_qa_chain(prompt_template):
-#     qa_chain = RetrievalQA.from_chain_type(
-#         llm=chat_model,
-#         chain_type="stuff",
-#         retriever=retriever,
-#         chain_type_kwargs={"prompt": prompt_template}
-#     )
-#     tools = [
-#         Tool(
-#             name='Knowledge Base',
-#             func=qa_chain,
-#             description='Use this tool when answering general knowledge queries to get more information about the topic'
-#         )
-#     ]
-#     return qa_chain, tools
-
-# # Define the agent initializer
-# def initialize_agent_with_prompt(prompt_template):
-#     qa_chain, tools = build_qa_chain(prompt_template)
-#     agent = initialize_agent(
-#         agent='chat-conversational-react-description',
-#         tools=tools,
-#         llm=chat_model,
-#         verbose=False,
-#         max_iteration=5,
-#         early_stopping_method='generate',
-#         memory=conversational_memory
-#     )
-#     return agent
-
-# # Define the function to generate answers
-# def generate_answer(message, choice):
-#     logging.debug(f"generate_answer called with prompt_choice: {choice}")
-    
-#     if choice == "Details":
-#         agent = initialize_agent_with_prompt(QA_CHAIN_PROMPT_1)
-#     elif choice == "Conversational":
-#         agent = initialize_agent_with_prompt(QA_CHAIN_PROMPT_2)
-#     else:
-#         logging.error(f"Invalid prompt_choice: {choice}. Defaulting to 'Conversational'")
-#         agent = initialize_agent_with_prompt(QA_CHAIN_PROMPT_2)
-#     response = agent(message)
-
-#     # Extract addresses for mapping regardless of the choice
-#     addresses = extract_addresses(response['output'])
-#     return response['output'], addresses
-
-# def bot(history, choice, tts_model):
-#     if not history:
-#         return history
-#     response, addresses = generate_answer(history[-1][0], choice)
-#     history[-1][1] = ""
-    
-#     # Generate audio for the entire response in a separate thread
-#     with concurrent.futures.ThreadPoolExecutor() as executor:
-#         if tts_model == "ElevenLabs":
-#             audio_future = executor.submit(generate_audio_elevenlabs, response)
-#         else:
-#             audio_future = executor.submit(generate_audio_parler_tts, response)
-        
-#         for character in response:
-#             history[-1][1] += character
-#             time.sleep(0.05)  # Adjust the speed of text appearance
-#             yield history, None
-        
-#         audio_path = audio_future.result()
-#         yield history, audio_path
-
-
-# def add_message(history, message):
-#     history.append((message, None))
-#     return history, gr.Textbox(value="", interactive=True, placeholder="Enter message or upload file...", show_label=False)
-
-# def print_like_dislike(x: gr.LikeData):
-#     print(x.index, x.value, x.liked)
-
-# def extract_addresses(response):
-#     if not isinstance(response, str):
-#         response = str(response)
-#     address_patterns = [
-#         r'([A-Z].*,\sBirmingham,\sAL\s\d{5})',
-#         r'(\d{4}\s.*,\sBirmingham,\sAL\s\d{5})',
-#         r'([A-Z].*,\sAL\s\d{5})',
-#         r'([A-Z].*,.*\sSt,\sBirmingham,\sAL\s\d{5})',
-#         r'([A-Z].*,.*\sStreets,\sBirmingham,\sAL\s\d{5})',
-#         r'(\d{2}.*\sStreets)',
-#         r'([A-Z].*\s\d{2},\sBirmingham,\sAL\s\d{5})'
-#         r'([a-zA-Z]\s Birmingham)'
-#     ]
-#     addresses = []
-#     for pattern in address_patterns:
-#         addresses.extend(re.findall(pattern, response))
-#     return addresses
-
-# all_addresses = []
-
-# def generate_map(location_names):
-#     global all_addresses
-#     all_addresses.extend(location_names)
-    
-#     api_key = os.environ['GOOGLEMAPS_API_KEY']
-#     gmaps = GoogleMapsClient(key=api_key)
-    
-#     m = folium.Map(location=[33.5175,-86.809444], zoom_start=16)
-    
-#     for location_name in all_addresses:
-#         geocode_result = gmaps.geocode(location_name)
-#         if geocode_result:
-#             location = geocode_result[0]['geometry']['location']
-#             folium.Marker(
-#                 [location['lat'], location['lng']],
-#                 tooltip=f"{geocode_result[0]['formatted_address']}"
-#             ).add_to(m)
-    
-#     map_html = m._repr_html_()
-#     return map_html
-
-# def fetch_local_news():
-#     api_key = os.environ['SERP_API']
-#     url = f'https://serpapi.com/search.json?engine=google_news&q=birmingham headline&api_key={api_key}'
-#     response = requests.get(url)
-#     if response.status_code == 200:
-#         results = response.json().get("news_results", [])
-#         news_html = """
-#         <h2 style="font-family: 'Georgia', serif; color: #ff0000; background-color: #f8f8f8; padding: 10px; border-radius: 10px;">Birmingham Today</h2>
-#         <style>
-#             .news-item {
-#                 font-family: 'Verdana', sans-serif;
-#                 color: #333;
-#                 background-color: #f0f8ff;
-#                 margin-bottom: 15px;
-#                 padding: 10px;
-#                 border-radius: 5px;
-#                 transition: box-shadow 0.3s ease, background-color 0.3s ease;
-#                 font-weight: bold;
-#             }
-#             .news-item:hover {
-#                 box-shadow: 0 4px 8px rgba(0, 0, 0, 0.1);
-#                 background-color: #e6f7ff;
-#             }
-#             .news-item a {
-#                 color: #1E90FF;
-#                 text-decoration: none;
-#                 font-weight: bold;
-#             }
-#             .news-item a:hover {
-#                 text-decoration: underline;
-#             }
-#             .news-preview {
-#                 position: absolute;
-#                 display: none;
-#                 border: 1px solid #ccc;
-#                 border-radius: 5px;
-#                 box-shadow: 0 2px 4px rgba(0, 0, 0, 0.2);
-#                 background-color: white;
-#                 z-index: 1000;
-#                 max-width: 300px;
-#                 padding: 10px;
-#                 font-family: 'Verdana', sans-serif;
-#                 color: #333;
-#             }
-#         </style>
-#         <script>
-#             function showPreview(event, previewContent) {
-#                 var previewBox = document.getElementById('news-preview');
-#                 previewBox.innerHTML = previewContent;
-#                 previewBox.style.left = event.pageX + 'px';
-#                 previewBox.style.top = event.pageY + 'px';
-#                 previewBox.style.display = 'block';
-#             }
-#             function hidePreview() {
-#                 var previewBox = document.getElementById('news-preview');
-#                 previewBox.style.display = 'none';
-#             }
-#         </script>
-#         <div id="news-preview" class="news-preview"></div>
-#         """
-#         for index, result in enumerate(results[:7]):
-#             title = result.get("title", "No title")
-#             link = result.get("link", "#")
-#             snippet = result.get("snippet", "")
-#             news_html += f"""
-#             <div class="news-item" onmouseover="showPreview(event, '{snippet}')" onmouseout="hidePreview()">
-#                 <a href='{link}' target='_blank'>{index + 1}. {title}</a>
-#                 <p>{snippet}</p>
-#             </div>
-#             """
-#         return news_html
-#     else:
-#         return "<p>Failed to fetch local news</p>"
-
-# # Voice Control
-# import numpy as np
-# import torch
-# from transformers import pipeline, AutoModelForSpeechSeq2Seq, AutoProcessor
-# from parler_tts import ParlerTTSForConditionalGeneration
-# from transformers import AutoTokenizer
-
-# model_id = 'openai/whisper-large-v3'
-# device = "cuda:0" if torch.cuda.is_available() else "cpu"
-# torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
-# model = AutoModelForSpeechSeq2Seq.from_pretrained(model_id, torch_dtype=torch_dtype,
-#                                                   #low_cpu_mem_usage=True,
-#                                                   use_safetensors=True).to(device)
-# processor = AutoProcessor.from_pretrained(model_id)
-
-# # Optimized ASR pipeline
-# pipe_asr = pipeline("automatic-speech-recognition", model=model, tokenizer=processor.tokenizer, feature_extractor=processor.feature_extractor, max_new_tokens=128, chunk_length_s=15, batch_size=16, torch_dtype=torch_dtype, device=device, return_timestamps=True)
-
-# base_audio_drive = "/data/audio"
-
-# import numpy as np
-
-# def transcribe_function(stream, new_chunk):
-#     try:
-#         sr, y = new_chunk[0], new_chunk[1]
-#     except TypeError:
-#         print(f"Error chunk structure: {type(new_chunk)}, content: {new_chunk}")
-#         return stream, "", None
-
-#     y = y.astype(np.float32) / np.max(np.abs(y))
-
-#     if stream is not None:
-#         stream = np.concatenate([stream, y])
-#     else:
-#         stream = y
-
-#     result = pipe_asr({"array": stream, "sampling_rate": sr}, return_timestamps=False)
-
-#     full_text = result.get("text", "")
-    
-#     return stream, full_text, result
-
-# def update_map_with_response(history):
-#     if not history:
-#         return ""
-#     response = history[-1][1]
-#     addresses = extract_addresses(response)
-#     return generate_map(addresses)
-
-# def clear_textbox():
-#     return "" 
-
-# def show_map_if_details(history,choice):
-#     if choice in ["Details", "Conversational"]:
-#         return gr.update(visible=True), update_map_with_response(history)
-#     else:
-#         return gr.update(visible=False), ""
-
-# def generate_audio_elevenlabs(text):
-#     XI_API_KEY = os.environ['ELEVENLABS_API']
-#     VOICE_ID = 'd9MIrwLnvDeH7aZb61E9'  # Replace with your voice ID
-#     tts_url = f"https://api.elevenlabs.io/v1/text-to-speech/{VOICE_ID}/stream"
-#     headers = {
-#         "Accept": "application/json",
-#         "xi-api-key": XI_API_KEY
-#     }
-#     data = {
-#         "text": str(text),
-#         "model_id": "eleven_multilingual_v2",
-#         "voice_settings": {
-#             "stability": 1.0,
-#             "similarity_boost": 0.0,
-#             "style": 0.60,  # Adjust style for more romantic tone
-#             "use_speaker_boost": False
-#         }
-#     }
-#     response = requests.post(tts_url, headers=headers, json=data, stream=True)
-#     if response.ok:
-#         with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as f:
-#             for chunk in response.iter_content(chunk_size=1024):
-#                 f.write(chunk)
-#             temp_audio_path = f.name
-#         logging.debug(f"Audio saved to {temp_audio_path}")
-#         return temp_audio_path
-#     else:
-#         logging.error(f"Error generating audio: {response.text}")
-#         return None
-
-# def generate_audio_parler_tts(text):
-#     model_id = 'parler-tts/parler_tts_mini_v0.1'
-#     device = "cuda:0" if torch.cuda.is_available() else "cpu"
-#     try:
-#         model = ParlerTTSForConditionalGeneration.from_pretrained(model_id).to(device)
-#     except torch.cuda.OutOfMemoryError:
-#         print("CUDA out of memory. Switching to CPU.")
-#         device = "cpu"
-#         model = ParlerTTSForConditionalGeneration.from_pretrained(model_id).to(device)
-#     tokenizer = AutoTokenizer.from_pretrained(model_id)
-
-#     description = "A female speaker with a slightly low-pitched voice delivers her words quite expressively, in a very confined sounding environment with clear audio quality. She speaks very fast."
-
-#     input_ids = tokenizer(description, return_tensors="pt").input_ids.to(device)
-#     prompt_input_ids = tokenizer(text, return_tensors="pt").input_ids.to(device)
-
-#     generation = model.generate(input_ids=input_ids, prompt_input_ids=prompt_input_ids)
-#     audio_arr = generation.cpu().numpy().squeeze()
-    
-#     with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as f:
-#         sf.write(f.name, audio_arr, model.config.sampling_rate)
-#         temp_audio_path = f.name
-
-#     logging.debug(f"Audio saved to {temp_audio_path}")
-#     return temp_audio_path
-
-# # Stable Diffusion setup
-# pipe = StableDiffusion3Pipeline.from_pretrained("stabilityai/stable-diffusion-3-medium-diffusers", torch_dtype=torch.float16)
-# pipe = pipe.to("cuda")
-
-# def generate_image(prompt):
-#     image = pipe(
-#         prompt,
-#         negative_prompt="",
-#         num_inference_steps=28,
-#         guidance_scale=3.0,
-#     ).images[0]
-#     return image
-
-# # Hardcoded prompt for image generation
-# hardcoded_prompt_1="Give a high quality photograph of a great looking red 2026 Bentley coupe against a skyline setting in th night, michael mann style in omaha enticing the consumer to buy this product"
-# hardcoded_prompt_2="A vibrant and dynamic football game scene in the style of Peter Paul Rubens, showcasing the intense match between Alabama and Nebraska. The players are depicted with the dramatic, muscular physiques and expressive faces typical of Rubens' style. The Alabama team is wearing their iconic crimson and white uniforms, while the Nebraska team is in their classic red and white attire. The scene is filled with action, with players in mid-motion, tackling, running, and catching the ball. The background features a grand stadium filled with cheering fans, banners, and the natural landscape in the distance. The colors are rich and vibrant, with a strong use of light and shadow to create depth and drama. The overall atmosphere captures the intensity and excitement of the game, infused with the grandeur and dynamism characteristic of Rubens' work."
-# hardcoded_prompt_3 = "Create a high-energy scene of a DJ performing on a large stage with vibrant lights, colorful lasers, a lively dancing crowd, and various electronic equipment in the background."
-
-# def update_images():
-#     image_1 = generate_image(hardcoded_prompt_1)
-#     image_2 = generate_image(hardcoded_prompt_2)
-#     image_3 = generate_image(hardcoded_prompt_3)
-#     return image_1, image_2, image_3
-
-# with gr.Blocks(theme='Pijush2023/scikit-learn-pijush') as demo:
-    
-#     with gr.Row():
-#         with gr.Column():
-#             state = gr.State()
-            
-#             chatbot = gr.Chatbot([], elem_id="RADAR:Channel 94.1", bubble_full_width=False)
-#             choice = gr.Radio(label="Select Style", choices=["Details", "Conversational"], value="Conversational")
-#             tts_choice = gr.Radio(label="Select TTS Model", choices=["ElevenLabs", "Parler TTS"], value="Parler TTS")
-            
-#             gr.Markdown("<h1 style='color: red;'>Talk to RADAR</h1>", elem_id="voice-markdown")
-#             chat_input = gr.Textbox(show_copy_button=True, interactive=True, show_label=False, label="ASK Radar !!!")
-#             chat_msg = chat_input.submit(add_message, [chatbot, chat_input], [chatbot, chat_input])
-#             bot_msg = chat_msg.then(bot, [chatbot, choice, tts_choice], [chatbot, gr.Audio(interactive=False, autoplay=True)])
-#             bot_msg.then(lambda: gr.Textbox(value="", interactive=True, placeholder="Ask Radar!!!...", show_label=False), None, [chat_input])
-#             chatbot.like(print_like_dislike, None, None)
-#             clear_button = gr.Button("Clear")
-#             clear_button.click(fn=clear_textbox, inputs=None, outputs=chat_input)
-           
-            
-#             audio_input = gr.Audio(sources=["microphone"], streaming=True, type='numpy')
-#             audio_input.stream(transcribe_function, inputs=[state, audio_input], outputs=[state, chat_input], api_name="SAMLOne_real_time")
-
-#             # gr.Markdown("<h1 style='color: red;'>Map</h1>", elem_id="location-markdown")
-#             # location_output = gr.HTML()
-#             # bot_msg.then(show_map_if_details, [chatbot, choice], [location_output, location_output])
-        
-#         # with gr.Column():
-#         #     weather_output = gr.HTML(value=fetch_local_weather())
-#         #     news_output = gr.HTML(value=fetch_local_news())
-#         #     news_output = gr.HTML(value=fetch_local_events())
-            
-#         with gr.Column():
-            
-#             image_output_1 = gr.Image(value=generate_image(hardcoded_prompt_1), width=400, height=400)
-#             image_output_2 = gr.Image(value=generate_image(hardcoded_prompt_2), width=400, height=400)
-#             image_output_3 = gr.Image(value=generate_image(hardcoded_prompt_3), width=400, height=400)
-
-
-#             refresh_button = gr.Button("Refresh Images")
-#             refresh_button.click(fn=update_images, inputs=None, outputs=[image_output_1, image_output_2, image_output_3])
-     
-# demo.queue()
-# demo.launch(share=True)
-
-
 import subprocess
 import sys
 
@@ -612,6 +7,7 @@ def install_parler_tts():
 # Call the function to install parler-tts
 install_parler_tts()
 
+
 import gradio as gr
 import requests
 import os
@@ -637,15 +33,6 @@ from langchain.chains import RetrievalQA
 from langchain.chains.conversation.memory import ConversationBufferWindowMemory
 from langchain.agents import Tool, initialize_agent
 from huggingface_hub import login
-from typing import Optional
-
-from pydub import AudioSegment
-import io
-import math
-from threading import Thread
-from queue import Queue
-from transformers.generation.streamers import BaseStreamer
-import numpy as np
 
 # Check if the token is already set in the environment variables
 hf_token = os.getenv("HF_TOKEN")
@@ -822,17 +209,17 @@ def get_weather_icon(condition):
 current_time_and_date = get_current_time_and_date()
 
 # Define prompt templates
-template1 = """You are an expert concierge who is helpful and a renowned guide for Birmingham, Alabama. Based on weather being a sunny bright day and today's date being 1st July 2024, use the following pieces of context, 
-memory, and message history, along with your knowledge of perennial events in Birmingham, Alabama, to answer the question at the end. If you don't know the answer, just say "Homie, I need to get more data for this," and don't try to make up an answer. 
+template1 = """You are an expert concierge who is helpful and a renowned guide for Birmingham,Alabama. Based on weather being a sunny bright day and the today's date is 1st july 2024, use the following pieces of context, 
+memory, and message history, along with your knowledge of perennial events in Birmingham,Alabama, to answer the question at the end. If you don't know the answer, just say "Homie, I need to get more data for this," and don't try to make up an answer. 
 Use fifteen sentences maximum. Keep the answer as detailed as possible. Always include the address, time, date, and
 event type and description. Always say "It was my pleasure!" at the end of the answer.
 {context}
 Question: {question}
 Helpful Answer:"""
 
-template2 = """You are an expert concierge who is helpful and a renowned guide for Birmingham, Alabama. Based on today's weather being a sunny bright day and today's date being 1st July 2024, take the location or address but don't show the location or address on the output prompts. Use the following pieces of context, 
-memory, and message history, along with your knowledge of perennial events in Birmingham, Alabama, to answer the question at the end. If you don't know the answer, just say "Homie, I need to get more data for this," and don't try to make up an answer. 
-Keep the answer short, sweet, and crisp. Always say "It was my pleasure!" at the end of the answer.
+template2 = """You are an expert concierge who is helpful and a renowned guide for Birmingham,Alabama. Based on today's weather being a sunny bright day and today's date is 1st july 2024, take the location or address but don't show the location or address on the output prompts. Use the following pieces of context, 
+memory, and message history, along with your knowledge of perennial events in Birmingham,Alabama, to answer the question at the end. If you don't know the answer, just say "Homie, I need to get more data for this," and don't try to make up an answer. 
+Keep the answer short and sweet and crisp. Always say "It was my pleasure!" at the end of the answer.
 {context}
 Question: {question}
 Helpful Answer:"""
@@ -892,7 +279,7 @@ def bot(history, choice, tts_model):
     if not history:
         return history
     response, addresses = generate_answer(history[-1][0], choice)
-    history[-1][1] = response
+    history[-1][1] = ""
     
     # Generate audio for the entire response in a separate thread
     with concurrent.futures.ThreadPoolExecutor() as executor:
@@ -909,6 +296,7 @@ def bot(history, choice, tts_model):
         audio_path = audio_future.result()
         yield history, audio_path
 
+
 def add_message(history, message):
     history.append((message, None))
     return history, gr.Textbox(value="", interactive=True, placeholder="Enter message or upload file...", show_label=False)
@@ -1119,175 +507,31 @@ def generate_audio_elevenlabs(text):
         logging.error(f"Error generating audio: {response.text}")
         return None
 
-# Changes start here
-class ParlerTTSStreamer(BaseStreamer):
-    def __init__(
-        self,
-        model: ParlerTTSForConditionalGeneration,
-        device: Optional[str] = None,
-        play_steps: Optional[int] = 10,
-        stride: Optional[int] = None,
-        timeout: Optional[float] = None,
-    ):
-        self.decoder = model.decoder
-        self.audio_encoder = model.audio_encoder
-        self.generation_config = model.generation_config
-        self.device = device if device is not None else model.device
-
-        self.play_steps = play_steps
-        if stride is not None:
-            self.stride = stride
-        else:
-            hop_length = math.floor(self.audio_encoder.config.sampling_rate / self.audio_encoder.config.frame_rate)
-            self.stride = hop_length * (play_steps - self.decoder.num_codebooks) // 6
-        self.token_cache = None
-        self.to_yield = 0
-
-        self.audio_queue = Queue()
-        self.stop_signal = None
-        self.timeout = timeout
-
-    def apply_delay_pattern_mask(self, input_ids):
-        _, delay_pattern_mask = self.decoder.build_delay_pattern_mask(
-            input_ids[:, :1],
-            bos_token_id=self.generation_config.bos_token_id,
-            pad_token_id=self.generation_config.decoder_start_token_id,
-            max_length=input_ids.shape[-1],
-        )
-        input_ids = self.decoder.apply_delay_pattern_mask(input_ids, delay_pattern_mask)
-
-        mask = (delay_pattern_mask != self.generation_config.bos_token_id) & (delay_pattern_mask != self.generation_config.pad_token_id)
-        input_ids = input_ids[mask].reshape(1, self.decoder.num_codebooks, -1)
-        input_ids = input_ids[None, ...]
-
-        input_ids = input_ids.to(self.audio_encoder.device)
-
-        decode_sequentially = (
-            self.generation_config.bos_token_id in input_ids
-            or self.generation_config.pad_token_id in input_ids
-            or self.generation_config.eos_token_id in input_ids
-        )
-        if not decode_sequentially:
-            output_values = self.audio_encoder.decode(
-                input_ids,
-                audio_scales=[None],
-            )
-        else:
-            sample = input_ids[:, 0]
-            sample_mask = (sample >= self.audio_encoder.config.codebook_size).sum(dim=(0, 1)) == 0
-            sample = sample[:, :, sample_mask]
-            output_values = self.audio_encoder.decode(sample[None, ...], [None])
-
-        audio_values = output_values.audio_values[0, 0]
-        return audio_values.cpu().float().numpy()
-
-    def put(self, value):
-        batch_size = value.shape[0] // self.decoder.num_codebooks
-        if batch_size > 1:
-            raise ValueError("ParlerTTSStreamer only supports batch size 1")
-
-        if self.token_cache is None:
-            self.token_cache = value
-        else:
-            self.token_cache = torch.concatenate([self.token_cache, value[:, None]], dim=-1)
-
-        if self.token_cache.shape[-1] % self.play_steps == 0:
-            audio_values = self.apply_delay_pattern_mask(self.token_cache)
-            self.on_finalized_audio(audio_values[self.to_yield : -self.stride])
-            self.to_yield += len(audio_values) - self.to_yield - self.stride
-
-    def end(self):
-        if self.token_cache is not None:
-            audio_values = self.apply_delay_pattern_mask(self.token_cache)
-        else:
-            audio_values = np.zeros(self.to_yield)
-
-        self.on_finalized_audio(audio_values[self.to_yield :], stream_end=True)
-
-    def on_finalized_audio(self, audio: np.ndarray, stream_end: bool = False):
-        self.audio_queue.put(audio, timeout=self.timeout)
-        if stream_end:
-            self.audio_queue.put(self.stop_signal, timeout=self.timeout)
-
-    def __iter__(self):
-        return self
-
-    def __next__(self):
-        value = self.audio_queue.get(timeout=self.timeout)
-        if not isinstance(value, np.ndarray) and value == self.stop_signal:
-            raise StopIteration()
-        else:
-            return value
-
-def numpy_to_mp3(audio_array, sampling_rate):
-    if np.issubdtype(audio_array.dtype, np.floating):
-        max_val = np.max(np.abs(audio_array))
-        audio_array = (audio_array / max_val) * 32767
-        audio_array = audio_array.astype(np.int16)
-
-    audio_segment = AudioSegment(
-        audio_array.tobytes(),
-        frame_rate=sampling_rate,
-        sample_width=audio_array.dtype.itemsize,
-        channels=1
-    )
-
-    mp3_io = io.BytesIO()
-    audio_segment.export(mp3_io, format="mp3", bitrate="320k")
-
-    mp3_bytes = mp3_io.getvalue()
-    mp3_io.close()
-
-    return mp3_bytes
-
 def generate_audio_parler_tts(text):
     model_id = 'parler-tts/parler_tts_mini_v0.1'
     device = "cuda:0" if torch.cuda.is_available() else "cpu"
-
     try:
         model = ParlerTTSForConditionalGeneration.from_pretrained(model_id).to(device)
     except torch.cuda.OutOfMemoryError:
         print("CUDA out of memory. Switching to CPU.")
         device = "cpu"
         model = ParlerTTSForConditionalGeneration.from_pretrained(model_id).to(device)
-
     tokenizer = AutoTokenizer.from_pretrained(model_id)
-    sampling_rate = model.audio_encoder.config.sampling_rate
-    frame_rate = model.audio_encoder.config.frame_rate
 
     description = "A female speaker with a slightly low-pitched voice delivers her words quite expressively, in a very confined sounding environment with clear audio quality. She speaks very fast."
-    play_steps_in_s = 2.0
-    play_steps = int(frame_rate * play_steps_in_s)
-
-    streamer = ParlerTTSStreamer(model, device=device, play_steps=play_steps)
-
-    inputs = tokenizer(description, return_tensors="pt").to(device)
-    prompt = tokenizer(text, return_tensors="pt").to(device)
 
-    generation_kwargs = dict(
-        input_ids=inputs.input_ids,
-        prompt_input_ids=prompt.input_ids,
-        streamer=streamer,
-        do_sample=True,
-        temperature=1.0,
-        min_new_tokens=10,
-    )
-
-    thread = Thread(target=model.generate, kwargs=generation_kwargs)
-    thread.start()
-
-    combined_audio_path = tempfile.NamedTemporaryFile(delete=False, suffix=".wav").name
-    combined_audio = []
-
-    for new_audio in streamer:
-        print(f"Sample of length: {round(new_audio.shape[0] / sampling_rate, 2)} seconds")
-        combined_audio.extend(new_audio)
+    input_ids = tokenizer(description, return_tensors="pt").input_ids.to(device)
+    prompt_input_ids = tokenizer(text, return_tensors="pt").input_ids.to(device)
 
-    sf.write(combined_audio_path, combined_audio, sampling_rate)
+    generation = model.generate(input_ids=input_ids, prompt_input_ids=prompt_input_ids)
+    audio_arr = generation.cpu().numpy().squeeze()
+    
+    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as f:
+        sf.write(f.name, audio_arr, model.config.sampling_rate)
+        temp_audio_path = f.name
 
-    logging.debug(f"Combined audio saved to {combined_audio_path}")
-    return combined_audio_path
-# Changes end here
+    logging.debug(f"Audio saved to {temp_audio_path}")
+    return temp_audio_path
 
 # Stable Diffusion setup
 pipe = StableDiffusion3Pipeline.from_pretrained("stabilityai/stable-diffusion-3-medium-diffusers", torch_dtype=torch.float16)
@@ -1358,3 +602,6 @@ with gr.Blocks(theme='Pijush2023/scikit-learn-pijush') as demo:
 demo.queue()
 demo.launch(share=True)
 
+
+
+