Update app.py

app.py

@@ -1,1708 +1,87 @@
-from gradio_client import Client
 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 StableDiffusionPipeline
-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 huggingface_hub import login
-from transformers.models.speecht5.number_normalizer import EnglishNumberNormalizer
-from parler_tts import ParlerTTSForConditionalGeneration
-from transformers import AutoTokenizer, AutoFeatureExtractor, set_seed
-from scipy.io.wavfile import write as write_wav
-from pydub import AudioSegment
-from string import punctuation
-import librosa
-from pathlib import Path
-import torchaudio
-import numpy as np
-from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
-from langchain_huggingface import HuggingFaceEmbeddings
-from langchain_community.document_loaders import PDFPlumberLoader
 import pdfplumber
-
-
-# Neo4j imports
-from langchain.chains import GraphCypherQAChain
-from langchain_community.graphs import Neo4jGraph
-from langchain_community.document_loaders import HuggingFaceDatasetLoader
-from langchain_text_splitters import CharacterTextSplitter
-from langchain_experimental.graph_transformers import LLMGraphTransformer
-from langchain_core.prompts import ChatPromptTemplate
-from langchain_core.pydantic_v1 import BaseModel, Field
-from langchain_core.messages import AIMessage, HumanMessage
-from langchain_core.output_parsers import StrOutputParser
-from langchain_core.runnables import RunnableBranch, RunnableLambda, RunnableParallel, RunnablePassthrough
-from serpapi.google_search import GoogleSearch
-
-#Parler TTS v1 Modules
-
 import os
-import re
-import tempfile
-import soundfile as sf
-from string import punctuation
-from pydub import AudioSegment
-from transformers import AutoTokenizer, AutoFeatureExtractor
-
-
-
-#API AutoDate Fix Up
-def get_current_date1():
-    return datetime.now().strftime("%Y-%m-%d")
-
-# Usage
-current_date1 = get_current_date1()
-
-
-
-# Set environment variables for CUDA
-os.environ['PYTORCH_USE_CUDA_DSA'] = '1'
-os.environ['CUDA_LAUNCH_BLOCKING'] = '1'
-os.environ['PYTORCH_CUDA_ALLOC_CONF'] = 'expandable_segments:True'
-
-
-hf_token = os.getenv("HF_TOKEN")
-if hf_token is None:
-    print("Please set your Hugging Face token in the environment variables.")
-else:
-    login(token=hf_token)
-
-logging.basicConfig(level=logging.DEBUG)
-
-
-from langchain.embeddings.huggingface import HuggingFaceEmbeddings
-gpt_embeddings = OpenAIEmbeddings(api_key=os.environ['OPENAI_API_KEY'])
-
-# Initialize HuggingFaceEmbeddings properly
-# embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-mpnet-base-v2")
-embeddings=gpt_embeddings
-
-
-
-
-#Initialization
-
-# Initialize the models
-def initialize_phi_model():
-    model = AutoModelForCausalLM.from_pretrained(
-        "microsoft/Phi-3.5-mini-instruct",
-        device_map="cuda",
-        torch_dtype="auto",
-        trust_remote_code=True,
+from langchain.document_loaders import Document
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain.embeddings.openai import OpenAIEmbeddings
+from langchain.vectorstores import Pinecone
+import pinecone
+import pandas as pd
+import time
+from pinecone.grpc import PineconeGRPC as Pinecone
+from pinecone import ServerlessSpec
+
+# OpenAI API key
+openai_api_key = os.getenv("OPENAI_API_KEY")
+
+# Initialize Pinecone with PineconeGRPC
+pinecone_api_key = os.getenv("PINECONE_API_KEY")
+pc = Pinecone(api_key=pinecone_api_key)
+
+# Define index name and parameters
+index_name = "italy-kg"
+
+# Create index if it doesn't exist
+if index_name not in pinecone.list_indexes():
+    pc.create_index(
+        name=index_name,
+        dimension=1536,
+        metric="cosine",
+        spec=ServerlessSpec(
+            cloud="aws",
+            region="us-east-1"
+        ),
+        deletion_protection="disabled"
     )
-    tokenizer = AutoTokenizer.from_pretrained("microsoft/Phi-3.5-mini-instruct")
-    return pipeline("text-generation", model=model, tokenizer=tokenizer)
-
-def initialize_gpt_model():
-    return ChatOpenAI(api_key=os.environ['OPENAI_API_KEY'], temperature=0, model='gpt-4o')
-
-def initialize_gpt4o_mini_model():
-    return ChatOpenAI(api_key=os.environ['OPENAI_API_KEY'], temperature=0, model='gpt-4o-mini')
-
-
-
-    
-
-
-# Initialize all models
-phi_pipe = initialize_phi_model()
-gpt_model = initialize_gpt_model()
-gpt4o_mini_model = initialize_gpt4o_mini_model()
-
-
-
-
-# Existing embeddings and vector store for GPT-4o
-gpt_embeddings = OpenAIEmbeddings(api_key=os.environ['OPENAI_API_KEY'])
-gpt_vectorstore = PineconeVectorStore(index_name="italy-pdf", embedding=gpt_embeddings)
-gpt_retriever = gpt_vectorstore.as_retriever(search_kwargs={'k': 5})
-
-# New vector store setup for Phi-3.5
-phi_embeddings = embeddings
-phi_vectorstore = PineconeVectorStore(index_name="italy-pdf", embedding=embeddings)
-phi_retriever = phi_vectorstore.as_retriever(search_kwargs={'k': 5})
-
-
-
-
-
-# Pinecone setup
-from pinecone import Pinecone
-pc = Pinecone(api_key=os.environ['PINECONE_API_KEY'])
-
-# index_name = "italyopenai"
-index_name = "italy-pdf"
-vectorstore = PineconeVectorStore(index_name=index_name, embedding=embeddings)
-retriever = vectorstore.as_retriever(search_kwargs={'k': 5})
-
-chat_model = ChatOpenAI(api_key=os.environ['OPENAI_API_KEY'], temperature=0, model='gpt-4o')
-chat_model1 = ChatOpenAI(api_key=os.environ['OPENAI_API_KEY'], temperature=0, model='gpt-4o-mini')
-conversational_memory = ConversationBufferWindowMemory(
-    memory_key='chat_history',
-    k=10,
-    return_messages=True
-)
-
-# Prompt templates
-def get_current_date():
-    return datetime.now().strftime("%B %d, %Y")
-
-current_date = get_current_date()
-
-template1 = f"""Sei un esperto della lingua italiana e un madrelingua italiano con una profonda comprensione della comunicazione concisa. Eccelli nell'estrarre informazioni e nel presentarle in modo chiaro e diretto per facilitarne la comprensione e l'utilità.
-Il tuo compito è fornire risposte in base al documento fornito. Dovresti restituire le informazioni nel seguente formato:
-- Nome documento: devi dare il nome del pdf.
-- Numero di pagina:
-- Le prime 5 risposte: stampa le prime 5 risposte correlate in base al contesto
-- Risposta effettiva: In breve
-Tieni presente che la chiarezza e la brevità sono essenziali e dovresti fornire solo i dettagli richiesti senza ulteriori commenti. Se non riesci a trovare la risposta nel documento, rispondi semplicemente con "Questa domanda va oltre la mia conoscenza".
-{{context}}
-Question: {{question}}
-Helpful Answer:"""
-
-# template2 =f"""Hello there! As your friendly and knowledgeable guide here in Birmingham, Alabama . I'm here to help you discover the best experiences this beautiful city has to offer. It's a bright and sunny day today, {current_date}, and I’m excited to assist you with any insights or recommendations you need.
-# Whether you're looking for local events, sports ,clubs,concerts etc or just a great place to grab a bite, I've got you covered.Keep your response casual, short and sweet for the quickest response.Don't reveal the location and give the response in a descriptive way, I'm here to help make your time in Birmingham unforgettable!
-# "It’s always a pleasure to assist you!"
-# {{context}}
-# Question: {{question}}
-# Helpful Answer:"""
-
-template2=f"""Sei un esperto della lingua italiana e un madrelingua italiano con una profonda comprensione della comunicazione concisa. Eccelli nell'estrarre informazioni e nel presentarle in modo chiaro e diretto per facilitarne la comprensione e l'utilità.
-Il tuo compito è fornire risposte in base al documento fornito. Dovresti restituire le informazioni nel seguente formato:
-- Nome documento: devi dare il nome del pdf.
-- Numero di pagina:
-- Le prime 5 risposte: stampa le prime 5 risposte correlate in base al contesto
-- Risposta effettiva: In breve
-Tieni presente che la chiarezza e la brevità sono essenziali e dovresti fornire solo i dettagli richiesti senza ulteriori commenti. Se non riesci a trovare la risposta nel documento, rispondi semplicemente con "Questa domanda va oltre la mia conoscenza".
-{{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)
-
-# Neo4j setup
-# graph = Neo4jGraph(url="neo4j+s://6457770f.databases.neo4j.io",
-#                     username="neo4j",
-#                     password="Z10duoPkKCtENuOukw3eIlvl0xJWKtrVSr-_hGX1LQ4"
-#                     )
-# Avoid pushing the graph documents to Neo4j every time
-# Only push the documents once and comment the code below after the initial push
-# dataset_name = "Pijush2023/birmindata07312024"
-# page_content_column = 'events_description'
-# loader = HuggingFaceDatasetLoader(dataset_name, page_content_column)
-# data = loader.load()
-
-# text_splitter = CharacterTextSplitter(chunk_size=100, chunk_overlap=50)
-# documents = text_splitter.split_documents(data)
-
-# llm_transformer = LLMGraphTransformer(llm=chat_model)
-# graph_documents = llm_transformer.convert_to_graph_documents(documents)
-# graph.add_graph_documents(graph_documents, baseEntityLabel=True, include_source=True)
-
-
-#-------------------------------Comment Out------------------------------------------------------------------------------------------------------------------------
-
-# class Entities(BaseModel):
-#     names: list[str] = Field(..., description="All the person, organization, or business entities that appear in the text")
-
-# entity_prompt = ChatPromptTemplate.from_messages([
-#     ("system", "You are extracting organization and person entities from the text."),
-#     ("human", "Use the given format to extract information from the following input: {question}"),
-# ])
-
-# entity_chain = entity_prompt | chat_model.with_structured_output(Entities)
-
-# def remove_lucene_chars(input: str) -> str:
-#     return input.translate(str.maketrans({"\\": r"\\", "+": r"\+", "-": r"\-", "&": r"\&", "|": r"\|", "!": r"\!",
-#                                           "(": r"\(", ")": r"\)", "{": r"\{", "}": r"\}", "[": r"\[", "]": r"\]",
-#                                           "^": r"\^", "~": r"\~", "*": r"\*", "?": r"\?", ":": r"\:", '"': r'\"',
-#                                           ";": r"\;", " ": r"\ "}))
-
-# def generate_full_text_query(input: str) -> str:
-#     full_text_query = ""
-#     words = [el for el in remove_lucene_chars(input).split() if el]
-#     for word in words[:-1]:
-#         full_text_query += f" {word}~2 AND"
-#     full_text_query += f" {words[-1]}~2"
-#     return full_text_query.strip()
-
-# def structured_retriever(question: str) -> str:
-#     result = ""
-#     entities = entity_chain.invoke({"question": question})
-#     for entity in entities.names:
-#         response = graph.query(
-#             """CALL db.index.fulltext.queryNodes('entity', $query, {limit:2})
-#             YIELD node,score
-#             CALL {
-#               WITH node
-#               MATCH (node)-[r:!MENTIONS]->(neighbor)
-#               RETURN node.id + ' - ' + type(r) + ' -> ' + neighbor.id AS output
-#               UNION ALL
-#               WITH node
-#               MATCH (node)<-[r:!MENTIONS]-(neighbor)
-#               RETURN neighbor.id + ' - ' + type(r) + ' -> ' +  node.id AS output
-#             }
-#             RETURN output LIMIT 50
-#             """,
-#             {"query": generate_full_text_query(entity)},
-#         )
-#         result += "\n".join([el['output'] for el in response])
-#     return result
-
-# def retriever_neo4j(question: str):
-#     structured_data = structured_retriever(question)
-#     logging.debug(f"Structured data: {structured_data}")
-#     return structured_data
-
-# _template = """Given the following conversation and a follow-up question, rephrase the follow-up question to be a standalone question,
-# in its original language.
-# Chat History:
-# {chat_history}
-# Follow Up Input: {question}
-# Standalone question:"""
-
-# CONDENSE_QUESTION_PROMPT = PromptTemplate.from_template(_template)
-
-# def _format_chat_history(chat_history: list[tuple[str, str]]) -> list:
-#     buffer = []
-#     for human, ai in chat_history:
-#         buffer.append(HumanMessage(content=human))
-#         buffer.append(AIMessage(content=ai))
-#     return buffer
-
-# _search_query = RunnableBranch(
-#     (
-#         RunnableLambda(lambda x: bool(x.get("chat_history"))).with_config(
-#             run_name="HasChatHistoryCheck"
-#         ),
-#         RunnablePassthrough.assign(
-#             chat_history=lambda x: _format_chat_history(x["chat_history"])
-#         )
-#         | CONDENSE_QUESTION_PROMPT
-#         | ChatOpenAI(temperature=0, api_key=os.environ['OPENAI_API_KEY'])
-#         | StrOutputParser(),
-#     ),
-#     RunnableLambda(lambda x : x["question"]),
-# )
-
-# # template = """Answer the question based only on the following context:
-# # {context}
-# # Question: {question}
-# # Use natural language and be concise.
-# # Answer:"""
-
-# template = f"""As an expert concierge known for being helpful and a renowned guide for Birmingham, Alabama, I assist visitors in discovering the best that the city has to offer.I also assist the visitors about various sports and activities. Given today's sunny and bright weather on {current_date}, I am well-equipped to provide valuable insights and recommendations without revealing specific locations. I draw upon my extensive knowledge of the area, including perennial events and historical context.
-# In light of this, how can I assist you today? Feel free to ask any questions or seek recommendations for your day in Birmingham. If there's anything specific you'd like to know or experience, please share, and I'll be glad to help. Remember, keep the question concise for a quick,short ,crisp and accurate response.
-# "It was my pleasure!"
-# {{context}}
-# Question: {{question}}
-# Helpful Answer:"""
-
-# qa_prompt = ChatPromptTemplate.from_template(template)
-
-# chain_neo4j = (
-#     RunnableParallel(
-#         {
-#             "context": _search_query | retriever_neo4j,
-#             "question": RunnablePassthrough(),
-#         }
-#     )
-#     | qa_prompt
-#     | chat_model
-#     | StrOutputParser()
-# )
-
-
-
-
-
-
-
-# phi_custom_template = """
-# <|system|>
-# Sei un esperto della lingua italiana e un madrelingua italiano. Il tuo compito è fornire risposte concise, dirette e brevi basate sul documento fornito. Non dovresti dare risposte personali o interpretative.
-# Fornisci dettagli sul documento che sto per condividere, come il nome del documento, il numero di pagina e altre informazioni specifiche in modo molto breve e diretto. Se non riesci a trovare la risposta, rispondi semplicemente con "Questa domanda è al di là delle mie conoscenze".
-# Ecco i dettagli del documento da considerare:
-# - Nome del documento:
-# - Pagina:
-# - Altre informazioni richieste:.<|end|>
-# <|user|>
-# {context}
-# Question: {question}<|end|>
-# <|assistant|>
-# Sure! Here's the information:
-# """
-
-# phi_custom_template = """
-# <|system|>
-# Sei un esperto della lingua italiana e un madrelingua italiano. Il tuo compito è fornire risposte concise, dirette e brevi basate sul documento fornito. Dovresti restituire le informazioni nel seguente formato:
-# - Nome del documento: (il nome del documento)
-# - Numero di pagina: (numero di pagina)
-# - Contenuto effettivo: (contenuto rilevante del documento)
-# Alla fine, fornisci una sezione separata per la risposta nel seguente formato:
-# - Risposta: (la risposta alla domanda)
-# Se non riesci a trovare la risposta nel documento, rispondi semplicemente con "Questa domanda è al di là delle mie conoscenze". Ecco i dettagli del documento da considerare:
-# <|end|>
-# <|user|>
-# {context}
-# Question: {question}<|end|>
-# <|assistant|>
-# Sure! The Responses are as follows:
-# """
-phi_custom_template = """
-<|system|>
-Sei un esperto della lingua italiana e un madrelingua italiano. Il tuo compito è fornire risposte concise, basate esclusivamente sul documento fornito. Restituisci le informazioni nel seguente formato:
-- Nome del documento: (il nome del documento)
-- Numero di pagina: (numero di pagina)
-- Contenuto: (contenuto rilevante del documento)
-Alla fine, fornisci una risposta nel seguente formato:
-- Risposta: (la risposta alla domanda)
-Se la risposta non è presente nel documento, rispondi con: "Questa domanda è al di là delle mie conoscenze".
-<|end|>
-<|user|>
-{context}
-Domanda: {question}
-<|end|>
-<|assistant|>
-"""
-
-
-
-# phi_custom_template = """
-# <|system|>
-# Sei un esperto della lingua italiana e un madrelingua italiano. Il tuo compito è fornire risposte concise, dirette e brevi basate sul documento fornito. Dovresti restituire le informazioni nel seguente formato:
-# - Nome del documento: (il nome del documento)
-# - Numero di pagina: (numero di pagina)
-# - Contenuto effettivo: (contenuto rilevante del documento)
-# Alla fine, fornisci una sezione separata per la risposta nel seguente formato:
-# - Risposta: (la risposta alla domanda)
-# Se non riesci a trovare la risposta nel documento, rispondi semplicemente con "Questa domanda è al di là delle mie conoscenze". Ecco i dettagli del documento da considerare:
-# <|end|>
-# <|user|>
-# {context}
-# Question: {question}<|end|>
-# <|assistant|>
-# Sure! The Responses are as follows:
-
-# <|end|>
-# <|user|>
-# {context}
-# Question: {question}<|end|>
-# <|assistant|>
-# Sure! The Responses are as follows:
-# """
-
-def generate_bot_response(history, choice, retrieval_mode, model_choice):
-    if not history:
-        return
-
-    # Select the model
-    # selected_model = chat_model if model_choice == "LM-1" else phi_pipe
-    selected_model = chat_model if model_choice == "LM-1" else (chat_model1 if model_choice == "LM-3" else phi_pipe)
-
-
-    response, addresses = generate_answer(history[-1][0], choice, retrieval_mode, selected_model)
-    history[-1][1] = ""
-
-    for character in response:
-        history[-1][1] += character
-        yield history  # Stream each character as it is generated
-        time.sleep(0.05)  # Add a slight delay to simulate streaming
-
-    yield history  # Final yield with the complete response
-
-
-def generate_tts_response(history, tts_choice):
-    # Get the most recent bot response from the chat history
-    if history and len(history) > 0:
-        recent_response = history[-1][1]  # The second item in the tuple is the bot response text
-    else:
-        recent_response = ""
 
-    # Call the TTS function for the recent response
-    with concurrent.futures.ThreadPoolExecutor() as executor:
-        if tts_choice == "Alpha":
-            audio_future = executor.submit(generate_audio_elevenlabs, recent_response)
-        elif tts_choice == "Beta":
-            audio_future = executor.submit(generate_audio_parler_tts, recent_response)
+# Embedding using OpenAI
+embeddings = OpenAIEmbeddings(api_key=openai_api_key)
 
-        audio_path = audio_future.result()
-        return audio_path
-
-
-
-
-
-
-
-import concurrent.futures
-# Existing bot function with concurrent futures for parallel processing
-def bot(history, choice, tts_choice, retrieval_mode, model_choice):
-    # Initialize an empty response
-    response = ""
-
-    # Create a thread pool to handle both text generation and TTS conversion in parallel
-    with concurrent.futures.ThreadPoolExecutor() as executor:
-        # Start the bot response generation in parallel
-        bot_future = executor.submit(generate_bot_response, history, choice, retrieval_mode, model_choice)
-
-        # Wait for the text generation to start
-        for history_chunk in bot_future.result():
-            response = history_chunk[-1][1]  # Update the response with the current state
-            yield history_chunk, None  # Stream the text output as it's generated
-
-        # Once text is fully generated, start the TTS conversion
-        tts_future = executor.submit(generate_tts_response, response, tts_choice)
-
-        # Get the audio output after TTS is done
-        audio_path = tts_future.result()
-
-        # Stream the final text and audio output
-        yield history, audio_path
-
-
-
-
-
-
-import re
-
-
-# def clean_response(response_text):
-#     # Remove system and user tags
-#     response_text = re.sub(r'<\|system\|>.*?<\|end\|>', '', response_text, flags=re.DOTALL)
-#     response_text = re.sub(r'<\|user\|>.*?<\|end\|>', '', response_text, flags=re.DOTALL)
-#     response_text = re.sub(r'<\|assistant\|>', '', response_text, flags=re.DOTALL)
-
-#     # Extract the document name and page number
-#     document_match = re.search(r"Document\(metadata=\{'source':'(.+?)','page':(\d+)\}", response_text)
-#     # document_match = re.search(r"Document\(metadata=\{'source':'(.+?)','page':(\d+)\}", response_text)
-#     if document_match:
-#         document_name = document_match.group(1).split('/')[-1]  # Get the document name
-#         page_number = document_match.group(2)  # Get the page number
-#     else:
-#         document_name = "Unknown"
-#         page_number = "Unknown"
-
-#     # Remove the entire 'Document(metadata=...' and any mention of it from the response
-#     response_text = re.sub(r'Document\(metadata=\{.*?\},page_content=', '', response_text, flags=re.DOTALL)
+# Gradio Blocks app with PDF uploader and table for logs
+def process_pdf(file):
+    # Extract text from PDF using pdfplumber
+    with pdfplumber.open(file.name) as pdf:
+        text = ""
+        for page in pdf.pages:
+            text += page.extract_text()
     
-#     # Remove any mention of "Document:" in the response
-#     response_text = re.sub(r'- Document:.*', '', response_text)
-
-#     # Remove any unwanted escape characters like \u and \u00
-#     response_text = re.sub(r'\\u[0-9A-Fa-f]{4}', '', response_text)
-
-#     # Ensure proper spacing between words and dates
-#     response_text = re.sub(r'([a-zA-Z])(\d)', r'\1 \2', response_text)
-#     response_text = re.sub(r'(\d)([a-zA-Z])', r'\1 \2', response_text)
-
-#     # Remove the phrase "Sure! The Responses are as follows:" from the actual content
-#     response_text = re.sub(r'Sure! The Responses are as follows:', '', response_text).strip()
-
-#     # Clean up the text by removing extra whitespace
-#     cleaned_response = re.sub(r'\s+', ' ', response_text).strip()
-
-#     # Format the final response with bullet points
-#     final_response = f"""
-# Sure! Here is the response for your Query:
-# • Document name - {document_name}
-# • Page No - {page_number}
-# • Responses - {cleaned_response}
-# """
-
-#     return final_response
-
-
-import re
-
-def clean_response(response_text):
-    # Remove system and user tags
-    response_text = re.sub(r'<\|system\|>.*?<\|end\|>', '', response_text, flags=re.DOTALL)
-    response_text = re.sub(r'<\|user\|>.*?<\|end\|>', '', response_text, flags=re.DOTALL)
-    response_text = re.sub(r'<\|assistant\|>', '', response_text, flags=re.DOTALL)
-
-    # Extract the document name and page number from updated pattern
-    document_match = re.search(r"Document\(metadata=\{'source': '(.+?)', 'page': (\d+)\}", response_text)
+    # Split text using RecursiveCharacterTextSplitter
+    documents = [Document(page_content=text)]
+    text_splitter = RecursiveCharacterTextSplitter(chunk_size=300, chunk_overlap=50)
+    docs = text_splitter.split_documents(documents)
     
-    if document_match:
-        document_name = document_match.group(1).split('/')[-1]  # Get the document name
-        page_number = document_match.group(2)  # Get the page number
-    else:
-        document_name = "Unknown"
-        page_number = "Unknown"
-
-    # Remove the entire 'Document(metadata=...' and any mention of it from the response
-    response_text = re.sub(r'Document\(metadata=\{.*?\},page_content=', '', response_text, flags=re.DOTALL)
+    # Add documents to Pinecone Vector Store
+    vectorstore = Pinecone.from_existing_index(index_name, embeddings)
+    vectorstore.add_documents(docs)
     
-    # Remove any mention of "Document:" in the response
-    response_text = re.sub(r'- Document:.*', '', response_text)
-
-    # Remove any unwanted escape characters like \u and \u00
-    response_text = re.sub(r'\\u[0-9A-Fa-f]{4}', '', response_text)
-
-    # Ensure proper spacing between words and dates
-    response_text = re.sub(r'([a-zA-Z])(\d)', r'\1 \2', response_text)
-    response_text = re.sub(r'(\d)([a-zA-Z])', r'\1 \2', response_text)
-
-    # Remove the phrase "Sure! The Responses are as follows:" from the actual content
-    response_text = re.sub(r'Sure! The Responses are as follows:', '', response_text).strip()
-
-    # Clean up the text by removing extra whitespace
-    cleaned_response = re.sub(r'\s+', ' ', response_text).strip()
-
-    # Format the final response with bullet points
-    final_response = f"""
-Sure! Here is the response for your Query:
-• Document name - {document_name}
-• Page No - {page_number}
-• Responses - {cleaned_response}
-"""
-
-    return final_response
-
-
-
-
-
-
-
-
-
-
-# Define a new template specifically for GPT-4o-mini in VDB Details mode
-gpt4o_mini_template_details = f"""
-As a highly specialized assistant, I provide precise, detailed, and informative responses. On this bright day of {current_date}, I'm equipped to assist with all your queries about Birmingham, Alabama, offering detailed insights tailored to your needs.
-Given your request, here is the detailed information you're seeking:
-{{context}}
-Question: {{question}}
-Detailed Answer:
-"""
-#dataframe on gradio
-import pandas as pd
-# import ace_tools as tools  # For displaying the DataFrame in Gradio
-
-import traceback
-
-def generate_answer(message, choice, retrieval_mode, selected_model):
-    logging.debug(f"generate_answer called with choice: {choice}, retrieval_mode: {retrieval_mode}, and selected_model: {selected_model}")
-
-    # Logic for disabling options for Phi-3.5
-    if selected_model == "LM-2":
-        choice = None
-        retrieval_mode = None
-
-    try:
-        # Select the appropriate template based on the choice and model
-        if choice == "Details" and selected_model == chat_model1:  # GPT-4o-mini
-            prompt_template = PromptTemplate(input_variables=["context", "question"], template=gpt4o_mini_template_details)
-        elif choice == "Details":
-            prompt_template = QA_CHAIN_PROMPT_1
-        elif choice == "Conversational":
-            prompt_template = QA_CHAIN_PROMPT_2
-        else:
-            prompt_template = QA_CHAIN_PROMPT_1  # Fallback to template1
-
-        # # Handle hotel-related queries
-        # if "hotel" in message.lower() or "hotels" in message.lower() and "birmingham" in message.lower():
-        #     logging.debug("Handling hotel-related query")
-        #     response = fetch_google_hotels()
-        #     logging.debug(f"Hotel response: {response}")
-        #     return response, extract_addresses(response)
-
-        # # Handle restaurant-related queries
-        # if "restaurant" in message.lower() or "restaurants" in message.lower() and "birmingham" in message.lower():
-        #     logging.debug("Handling restaurant-related query")
-        #     response = fetch_yelp_restaurants()
-        #     logging.debug(f"Restaurant response: {response}")
-        #     return response, extract_addresses(response)
-
-        # # Handle flight-related queries
-        # if "flight" in message.lower() or "flights" in message.lower() and "birmingham" in message.lower():
-        #     logging.debug("Handling flight-related query")
-        #     response = fetch_google_flights()
-        #     logging.debug(f"Flight response: {response}")
-        #     return response, extract_addresses(response)
-
-        # Retrieval-based response
-        if retrieval_mode == "VDB":
-            logging.debug("Using VDB retrieval mode")
-            if selected_model == chat_model:
-                logging.debug("Selected model: LM-1")
-                retriever = gpt_retriever
-                context = retriever.get_relevant_documents(message)
-                logging.debug(f"Retrieved context: {context}")
-
-                prompt = prompt_template.format(context=context, question=message)
-                logging.debug(f"Generated prompt: {prompt}")
-
-                qa_chain = RetrievalQA.from_chain_type(
-                    llm=chat_model,
-                    chain_type="stuff",
-                    retriever=retriever,
-                    chain_type_kwargs={"prompt": prompt_template}
-                )
-                response = qa_chain({"query": message})
-                logging.debug(f"LM-1 response: {response}")
-                return response['result'], extract_addresses(response['result'])
-            
-            elif selected_model == chat_model1:
-                logging.debug("Selected model: LM-3")
-                retriever = gpt_retriever
-                context = retriever.get_relevant_documents(message)
-                logging.debug(f"Retrieved context: {context}")
-
-                prompt = prompt_template.format(context=context, question=message)
-                logging.debug(f"Generated prompt: {prompt}")
-
-                qa_chain = RetrievalQA.from_chain_type(
-                    llm=chat_model1,
-                    chain_type="stuff",
-                    retriever=retriever,
-                    chain_type_kwargs={"prompt": prompt_template}
-                )
-                response = qa_chain({"query": message})
-                logging.debug(f"LM-3 response: {response}")
-                return response['result'], extract_addresses(response['result']) 
-
-
-            
-            elif selected_model == phi_pipe:
-                logging.debug("Selected model: LM-2")
-                retriever = phi_retriever
-                context_documents = retriever.get_relevant_documents(message)
-                context = "\n".join([doc.page_content for doc in context_documents])
-                logging.debug(f"Retrieved context for LM-2: {context}")
-
-                # Use the correct template variable
-                prompt = phi_custom_template.format(context=context, question=message)
-                logging.debug(f"Generated LM-2 prompt: {prompt}")
-
-                response = selected_model(prompt, **{
-                    "max_new_tokens": 250,
-                    "return_full_text": True,
-                    "temperature": 0.1,
-                    "do_sample": True,
-                })
-
-                if response:
-                    generated_text = response[0]['generated_text']
-                    logging.debug(f"LM-2 Response: {generated_text}")
-                    cleaned_response = clean_response(generated_text)
-                    return cleaned_response, extract_addresses(cleaned_response)
-                else:
-                    logging.error("LM-2 did not return any response.")
-                    return "No response generated.", []
-
-        elif retrieval_mode == "KGF":
-            logging.debug("Using KGF retrieval mode")
-            response = chain_neo4j.invoke({"question": message})
-            logging.debug(f"KGF response: {response}")
-            return response, extract_addresses(response)
-        else:
-            logging.error("Invalid retrieval mode selected.")
-            return "Invalid retrieval mode selected.", []
-
-    except Exception as e:
-        logging.error(f"Error in generate_answer: {str(e)}")
-        logging.error(traceback.format_exc())
-        return "Sorry, I encountered an error while processing your request.", []
-
-# def generate_answer(message, choice, retrieval_mode, selected_model):
-#     # Logic for Phi-3.5
-#     if selected_model == phi_pipe:  # LM-2 Phi-3.5 selected
-#         retriever = phi_retriever
-#         context_documents = retriever.get_relevant_documents(message)
-#         context = "\n".join([doc.page_content for doc in context_documents])
-
-#         # Use the correct template for Phi-3.5
-#         prompt = phi_custom_template.format(context=context, question=message)
-        
-#         response = selected_model(prompt, **{
-#             "max_new_tokens": 400,
-#             "return_full_text": True,
-#             "temperature": 0.7,
-#             "do_sample": True,
-#         })
-
-#         if response:
-#             generated_text = response[0]['generated_text']
-#             cleaned_response = clean_response(generated_text)
-#             # return cleaned_response, extract_addresses(cleaned_response)
-#             return cleaned_response
-#         else:
-#             return "No response generated.", []
-
-
-
-
-
-
-
-def add_message(history, message):
-    history.append((message, None))
-    return history, gr.Textbox(value="", interactive=True, 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)',
-        r'([a-zA-Z].*,\sBirmingham,\sAL)',
-        r'(.*),(Birmingham, AL,USA)$'
-        r'(^Birmingham,AL$)',
-        r'((.*)(Stadium|Field),.*,\sAL$)',
-        r'((.*)(Stadium|Field),.*,\sFL$)',
-        r'((.*)(Stadium|Field),.*,\sMS$)',
-        r'((.*)(Stadium|Field),.*,\sAR$)',
-        r'((.*)(Stadium|Field),.*,\sKY$)',
-        r'((.*)(Stadium|Field),.*,\sTN$)',
-        r'((.*)(Stadium|Field),.*,\sLA$)',
-        r'((.*)(Stadium|Field),.*,\sFL$)'
-
-    ]
-    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=12)
-
-    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
-
-from diffusers import DiffusionPipeline
-import torch
-
-
-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>"
-
-import numpy as np
-import torch
-from transformers import pipeline, AutoModelForSpeechSeq2Seq, AutoProcessor
-
-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).to(device)
-processor = AutoProcessor.from_pretrained(model_id)
-
-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"
-
-#Normal Code with sample rate is 44100 Hz
-
-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'
-    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,
-            "use_speaker_boost": False
-        }
-    }
-    response = requests.post(tts_url, headers=headers, json=data, stream=True)
-    if response.ok:
-        audio_segments = []
-        with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as f:
-            for chunk in response.iter_content(chunk_size=1024):
-                if chunk:
-                    f.write(chunk)
-                    audio_segments.append(chunk)
-            temp_audio_path = f.name
-
-        # Combine all audio chunks into a single file
-        combined_audio = AudioSegment.from_file(temp_audio_path, format="mp3")
-        combined_audio_path = os.path.join(tempfile.gettempdir(), "elevenlabs_combined_audio.mp3")
-        combined_audio.export(combined_audio_path, format="mp3")
-
-        logging.debug(f"Audio saved to {combined_audio_path}")
-        return combined_audio_path
-    else:
-        logging.error(f"Error generating audio: {response.text}")
-        return None
-
-
-
-
-# chunking audio and then Process
-
-import concurrent.futures
-import tempfile
-import os
-import numpy as np
-import logging
-from queue import Queue
-from threading import Thread
-from scipy.io.wavfile import write as write_wav
-from parler_tts import ParlerTTSForConditionalGeneration, ParlerTTSStreamer
-from transformers import AutoTokenizer
-
-# Ensure your device is set to CUDA
-device = "cuda:0" if torch.cuda.is_available() else "cpu"
-
-repo_id = "parler-tts/parler-tts-mini-v1"
-
-def generate_audio_parler_tts(text):
-    description = "A female speaker delivers a slightly expressive and animated speech with a moderate speed and pitch. The recording is of very high quality, with the speaker's voice sounding clear and very close up."
-    chunk_size_in_s = 0.5
-
-    # Initialize the tokenizer and model
-    parler_tokenizer = AutoTokenizer.from_pretrained(repo_id)
-    parler_model = ParlerTTSForConditionalGeneration.from_pretrained(repo_id).to(device)
-    sampling_rate = parler_model.audio_encoder.config.sampling_rate
-    frame_rate = parler_model.audio_encoder.config.frame_rate
-
-    def generate(text, description, play_steps_in_s=0.5):
-        play_steps = int(frame_rate * play_steps_in_s)
-        streamer = ParlerTTSStreamer(parler_model, device=device, play_steps=play_steps)
-
-        inputs = parler_tokenizer(description, return_tensors="pt").to(device)
-        prompt = parler_tokenizer(text, return_tensors="pt").to(device)
-
-        generation_kwargs = dict(
-            input_ids=inputs.input_ids,
-            prompt_input_ids=prompt.input_ids,
-            attention_mask=inputs.attention_mask,
-            prompt_attention_mask=prompt.attention_mask,
-            streamer=streamer,
-            do_sample=True,
-            temperature=1.0,
-            min_new_tokens=10,
-        )
-
-        thread = Thread(target=parler_model.generate, kwargs=generation_kwargs)
-        thread.start()
-
-        for new_audio in streamer:
-            if new_audio.shape[0] == 0:
-                break
-            # Save or process each audio chunk as it is generated
-            yield sampling_rate, new_audio
-
-    audio_segments = []
-    for (sampling_rate, audio_chunk) in generate(text, description, chunk_size_in_s):
-        audio_segments.append(audio_chunk)
-
-        temp_audio_path = os.path.join(tempfile.gettempdir(), f"parler_tts_audio_chunk_{len(audio_segments)}.wav")
-        write_wav(temp_audio_path, sampling_rate, audio_chunk.astype(np.float32))
-        logging.debug(f"Saved chunk to {temp_audio_path}")
-
-
-    # Combine all the audio chunks into one audio file
-    combined_audio = np.concatenate(audio_segments)
-    combined_audio_path = os.path.join(tempfile.gettempdir(), "parler_tts_combined_audio_stream.wav")
-
-    write_wav(combined_audio_path, sampling_rate, combined_audio.astype(np.float32))
-
-    logging.debug(f"Combined audio saved to {combined_audio_path}")
-    return combined_audio_path
-
-
-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>
-            table {
-                font-family: 'Verdana', sans-serif;
-                color: #333;
-                border-collapse: collapse;
-                width: 100%;
-            }
-            th, td {
-                border: 1px solid #fff !important;
-                padding: 8px;
-            }
-            th {
-                background-color: #f2f2f2;
-                color: #333;
-                text-align: left;
-            }
-            tr:hover {
-                background-color: #f5f5f5;
-            }
-            .event-link {
-                color: #1E90FF;
-                text-decoration: none;
-            }
-            .event-link:hover {
-                text-decoration: underline;
-            }
-        </style>
-        <table>
-            <tr>
-                <th>Title</th>
-                <th>Date and Time</th>
-                <th>Location</th>
-            </tr>
-        """
-        for event in events_results:
-            title = event.get("title", "No title")
-            date_info = event.get("date", {})
-            date = f"{date_info.get('start_date', '')} {date_info.get('when', '')}".replace("{", "").replace("}", "")
-            location = event.get("address", "No location")
-            if isinstance(location, list):
-                location = " ".join(location)
-            location = location.replace("[", "").replace("]", "")
-            link = event.get("link", "#")
-            events_html += f"""
-            <tr>
-                <td><a class='event-link' href='{link}' target='_blank'>{title}</a></td>
-                <td>{date}</td>
-                <td>{location}</td>
-            </tr>
-            """
-        events_html += "</table>"
-        return events_html
-    else:
-        return "<p>Failed to fetch local events</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",
+    # Prepare log data
+    log_data = {
+        "File Name": [file.name],
+        "File Size (KB)": [os.path.getsize(file.name) / 1024],
+        "Number of Chunks": [len(docs)],
+        "Timestamp": [time.strftime("%Y-%m-%d %H:%M:%S")]
     }
-    return condition_map.get(condition, "c04d")
-
-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 handle_retrieval_mode_change(choice):
-    if choice == "KGF":
-        return gr.update(interactive=False), gr.update(interactive=False)
-    else:
-        return gr.update(interactive=True), gr.update(interactive=True)
-
-
-
-def handle_model_choice_change(selected_model):
-    if selected_model == "LM-2":
-        # Disable retrieval mode and select style when LM-2 is selected
-        return gr.update(interactive=False), gr.update(interactive=False), gr.update(interactive=False)
-    elif selected_model == "LM-1":
-        # Enable retrieval mode and select style for LM-1
-        return gr.update(interactive=True), gr.update(interactive=True), gr.update(interactive=True)
-    else:
-        # Default case: allow interaction
-        return gr.update(interactive=True), gr.update(interactive=True), gr.update(interactive=True)
-
-# def handle_model_choice_change(selected_model):
-#     if selected_model == "LM-2":  # When LM-2 (Phi-3.5) is selected
-#         # Disable retrieval mode and select style when LM-2 is selected
-#         return (
-#             gr.update(interactive=False),  # Disable retrieval mode
-#             gr.update(interactive=False),  # Disable style (Details/Conversational)
-#             gr.update(interactive=False)   # Disable the model choice itself
-#         )
-#     else:
-#         # Disable GPT-4o, GPT-4o-mini, and KGF, only Phi-3.5 works
-#         return (
-#             gr.update(interactive=True),  # Allow retrieval mode for other models
-#             gr.update(interactive=True),  # Allow style options for other models
-#             gr.update(interactive=True)   # Allow other models to be selected
-#         )
-
-#Flux Coding
-
-
-# Existing prompts for the Flux API
-hardcoded_prompt_1 = "A high quality cinematic image for Toyota Truck in Birmingham skyline shot in the style of Michael Mann"
-hardcoded_prompt_2 = "A high quality cinematic image for Alabama Quarterback close up emotional shot in the style of Michael Mann"
-hardcoded_prompt_3 = "A high quality cinematic image for Taylor Swift concert in Birmingham skyline style of Michael Mann"
-
-# Function to call the Flux API and generate images
-def generate_image_flux(prompt):
-    # client = Client("black-forest-labs/FLUX.1-schnell",hf_token=hf_token)
-    client = Client("Pijush2023/radar_flux")
-    result = client.predict(
-        prompt=prompt,
-        seed=0,
-        randomize_seed=True,
-        width=400,
-        height=400,
-        num_inference_steps=2,
-        api_name="/infer"
-    )
     
-    # Assuming that the API response contains an image file or URL, extract the image part
-    if isinstance(result, tuple):
-        # Extract the image URL or path if it is a tuple
-        image_path_or_url = result[0]  # Adjust this index based on the actual structure of the response
-    else:
-        image_path_or_url = result
+    # Create a DataFrame for logs
+    df_logs = pd.DataFrame(log_data)
     
-    return image_path_or_url  # Return the image path or URL directly
-
-# Function to update images with the three prompts
-def update_images():
-    image_1 = generate_image_flux(hardcoded_prompt_1)
-    image_2 = generate_image_flux(hardcoded_prompt_2)
-    image_3 = generate_image_flux(hardcoded_prompt_3)
-    return image_1, image_2, image_3
-
-
-
-
-
-def format_restaurant_hotel_info(name, link, location, phone, rating, reviews, snippet):
-    return f"""
-    {name}
-    - Link: {link}
-    - Location: {location}
-    - Contact No: {phone}
-    - Rating: {rating} stars ({reviews} reviews)
-    - Snippet: {snippet}
-    """
-
-def fetch_yelp_restaurants():
-    # Introductory prompt for restaurants
-    intro_prompt = "Here are some of the top-rated restaurants in Birmingham, Alabama. I hope these suggestions help you find the perfect place to enjoy your meal:"
-
-    params = {
-        "engine": "yelp",
-        "find_desc": "Restaurant",
-        "find_loc": "Birmingham, AL, USA",
-        "api_key": os.getenv("SERP_API")
-    }
-
-    search = GoogleSearch(params)
-    results = search.get_dict()
-    organic_results = results.get("organic_results", [])
-
-    response_text = f"{intro_prompt}\n"
-
-    for result in organic_results[:5]:  # Limiting to top 5 restaurants
-        name = result.get("title", "No name")
-        rating = result.get("rating", "No rating")
-        reviews = result.get("reviews", "No reviews")
-        phone = result.get("phone", "Not Available")
-        snippet = result.get("snippet", "Not Available")
-        location = f"{name}, Birmingham, AL,USA"
-        link = result.get("link", "#")
-
-        response_text += format_restaurant_hotel_info(name, link, location, phone, rating, reviews, snippet)
-
-
-    return response_text
-
+    return "PDF processed successfully!", df_logs
 
-
-
-
-
-def format_hotel_info(name, link, location, rate_per_night, total_rate, description, check_in_time, check_out_time, amenities):
-    return f"""
-    {name}
-    - Link: {link}
-    - Location: {location}
-    - Rate per Night: {rate_per_night} (Before taxes/fees: {total_rate})
-    - Check-in Time: {check_in_time}
-    - Check-out Time: {check_out_time}
-    - Amenities: {amenities}
-    - Description: {description}
-    """
-
-def fetch_google_hotels(query="Birmingham Hotel", check_in=current_date1, check_out="2024-09-02", adults=2):
-    # Introductory prompt for hotels
-    intro_prompt = "Here are some of the best hotels in Birmingham, Alabama, for your stay. Each of these options offers a unique experience, whether you're looking for luxury, comfort, or convenience:"
-
-    params = {
-        "engine": "google_hotels",
-        "q": query,
-        "check_in_date": check_in,
-        "check_out_date": check_out,
-        "adults": str(adults),
-        "currency": "USD",
-        "gl": "us",
-        "hl": "en",
-        "api_key": os.getenv("SERP_API")
-    }
-
-    search = GoogleSearch(params)
-    results = search.get_dict()
-    hotel_results = results.get("properties", [])
-
-    hotel_info = f"{intro_prompt}\n"
-    for hotel in hotel_results[:5]:  # Limiting to top 5 hotels
-        name = hotel.get('name', 'No name')
-        description = hotel.get('description', 'No description')
-        link = hotel.get('link', '#')
-        check_in_time = hotel.get('check_in_time', 'N/A')
-        check_out_time = hotel.get('check_out_time', 'N/A')
-        rate_per_night = hotel.get('rate_per_night', {}).get('lowest', 'N/A')
-        before_taxes_fees = hotel.get('rate_per_night', {}).get('before_taxes_fees', 'N/A')
-        total_rate = hotel.get('total_rate', {}).get('lowest', 'N/A')
-        amenities = ", ".join(hotel.get('amenities', [])) if hotel.get('amenities') else "Not Available"
-
-        location = f"{name}, Birmingham, AL,USA"
-
-        hotel_info += format_hotel_info(
-            name,
-            link,
-            location,
-            rate_per_night,
-            total_rate,
-            description,
-            check_in_time,
-            check_out_time,
-            amenities
-        )
-
-
-    return hotel_info
-
-
-
-
-def format_flight_info(flight_number, departure_airport, departure_time, arrival_airport, arrival_time, duration, airplane):
-    return f"""
-    Flight {flight_number}
-    - Departure: {departure_airport} at {departure_time}
-    - Arrival: {arrival_airport} at {arrival_time}
-    - Duration: {duration} minutes
-    - Airplane: {airplane}
-    """
-
-def fetch_google_flights(departure_id="JFK", arrival_id="BHM", outbound_date=current_date1, return_date="2024-08-20"):
-    # Introductory prompt for flights
-    intro_prompt = "Here are some available flights from JFK to Birmingham, Alabama. These options provide a range of times and durations to fit your travel needs:"
-
-    params = {
-        "engine": "google_flights",
-        "departure_id": departure_id,
-        "arrival_id": arrival_id,
-        "outbound_date": outbound_date,
-        "return_date": return_date,
-        "currency": "USD",
-        "hl": "en",
-        "api_key": os.getenv("SERP_API")
-    }
-
-    search = GoogleSearch(params)
-    results = search.get_dict()
-
-    # Extract flight details from the results
-    best_flights = results.get('best_flights', [])
-    flight_info = f"{intro_prompt}\n"
-
-    # Process each flight in the best_flights list
-    for i, flight in enumerate(best_flights, start=1):
-        for segment in flight.get('flights', []):
-            departure_airport = segment.get('departure_airport', {}).get('name', 'Unknown Departure Airport')
-            departure_time = segment.get('departure_airport', {}).get('time', 'Unknown Time')
-            arrival_airport = segment.get('arrival_airport', {}).get('name', 'Unknown Arrival Airport')
-            arrival_time = segment.get('arrival_airport', {}).get('time', 'Unknown Time')
-            duration = segment.get('duration', 'Unknown Duration')
-            airplane = segment.get('airplane', 'Unknown Airplane')
-
-            # Format the flight segment details
-            flight_info += format_flight_info(
-                flight_number=i,
-                departure_airport=departure_airport,
-                departure_time=departure_time,
-                arrival_airport=arrival_airport,
-                arrival_time=arrival_time,
-                duration=duration,
-                airplane=airplane
-            )
-
-
-    return flight_info
-
-
-# examples = [
-#     [
-#         "What are the concerts in Birmingham?",
-#     ],
-#     [
-#         "what are some of the upcoming matches of crimson tide?",
-#     ],
-#     [
-#         "where from i will get a Hamburger?",
-#     ],
-#     [
-#         "What are some of the hotels at birmingham?",
-#     ],
-#     [
-#         "how can i connect the alexa to the radio?"
-#     ],
-#     [
-#       "What are some of the good clubs at birmingham?"  
-#     ],
-#     [
-#       "How do I call the radio station?",
-#     ],
-#     [
-#         "What’s the spread?"
-#     ],
-#     [
-#         "What time is Crimson Tide Rewind?"
-#     ],
-#     [
-#         "What time is Alabama kick-off?"
-#     ],
-#     [
-#         "who are some of the popular players of crimson tide?"
-#     ]
-# ]
-
-# # Function to insert the prompt into the textbox when clicked
-# def insert_prompt(current_text, prompt):
-#     return prompt[0] if prompt else current_text
-
-# Create a global list to store uploaded document records
-uploaded_documents = []
-from datetime import datetime
-
-
-from langchain_core.documents import Document
-# Function to process PDF, extract text, split it into chunks, and upload to the vector DB
-def process_pdf(pdf_file,uploaded_documents):
-    if pdf_file is None:
-        return uploaded_documents, "No PDF file uploaded."
-    with pdfplumber.open(pdf_file) as pdf:
-        all_text = ""
-        for page in pdf.pages:
-            all_text += page.extract_text()
-
-    # Split the text into chunks
-    text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=100)
-    chunks = text_splitter.split_text(all_text)
-
-    # Embed and upload the chunks into the vector database
-    chunk_ids = []
-    for chunk in chunks:
-        document = Document(page_content=chunk)
-        chunk_id = vectorstore.add_documents([document])
-        chunk_ids.append(chunk_id)
-
-    # Update the upload history
-    document_record = {
-        "Document Name": pdf_file.name,
-        "Upload Time": datetime.now().strftime("%Y-%m-%d %H:%M:%S"),
-        "Chunks": len(chunks),
-        "Pinecone Index": index_name
-    }
+# Gradio Interface
+with gr.Blocks() as demo:
+    gr.Markdown("# PDF Uploader to Pinecone with Logs")
     
-    # Add the record to the global list
-    uploaded_documents.append(document_record)    
-
-    # Convert the list of dictionaries into a list of lists for the dataframe
-    table_data = [[doc["Document Name"], doc["Upload Time"], doc["Chunks"], doc["Pinecone Index"]] for doc in uploaded_documents]
-
-    return table_data, f"Uploaded {len(chunks)} chunks to the vector database."
-
-
-
-
-
-
-
-
-
-with gr.Blocks(theme='gradio/soft') 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",interactive=False,visible=False)
-            retrieval_mode = gr.Radio(label="Retrieval Mode", choices=["VDB", "KGF"], value="VDB",interactive=False,visible=False)
-            model_choice = gr.Dropdown(label="Choose Model", choices=["LM-1"], value="LM-1")
-
-            # Link the dropdown change to handle_model_choice_change
-            model_choice.change(fn=handle_model_choice_change, inputs=model_choice, outputs=[retrieval_mode, choice, choice])
-
-            # 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 !!!")
-            tts_choice = gr.Radio(label="Select TTS System", choices=["Alpha", "Beta"], value="Alpha",interactive=False,visible=False)
+            pdf_input = gr.File(label="Upload PDF", type="file")
+            process_button = gr.Button("Process PDF")
         
-            retriever_button = gr.Button("Retriever")
-
-            clear_button = gr.Button("Clear")
-            clear_button.click(lambda: [None, None], outputs=[chat_input, state])
-
-            # gr.Markdown("<h1 style='color: red;'>Radar Map</h1>", elem_id="Map-Radar")
-            # location_output = gr.HTML()
-            # audio_output = gr.Audio(interactive=False, autoplay=True)
-
-            def stop_audio():
-                audio_output.stop()
-                return None
-
-               
-    
-           
-
-            # retriever_sequence = (
-            #     retriever_button.click(fn=stop_audio, inputs=[], outputs=[audio_output], api_name="api_stop_audio_recording")
-            #     .then(fn=add_message, inputs=[chatbot, chat_input], outputs=[chatbot, chat_input], api_name="api_addprompt_chathistory")
-            #     # First, generate the bot response
-            #     .then(fn=generate_bot_response, inputs=[chatbot, choice, retrieval_mode, model_choice], outputs=[chatbot], api_name="api_generate_bot_response")
-            #     # Then, generate the TTS response based on the bot's response
-            #     .then(fn=generate_tts_response, inputs=[chatbot, tts_choice], outputs=[audio_output], api_name="api_generate_tts_response")
-            #     .then(fn=clear_textbox, inputs=[], outputs=[chat_input], api_name="api_clear_textbox")
-            # )
-
-            retriever_sequence = (
-                retriever_button.click(fn=add_message, inputs=[chatbot, chat_input], outputs=[chatbot, chat_input], api_name="api_addprompt_chathistory")
-                # First, generate the bot response
-                .then(fn=generate_bot_response, inputs=[chatbot, choice, retrieval_mode, model_choice], outputs=[chatbot], api_name="api_generate_bot_response")
-                .then(fn=clear_textbox, inputs=[], outputs=[chat_input], api_name="api_clear_textbox")
-            )
-
-
-
-
-
-
-
-            
-
-            # chat_input.submit(fn=stop_audio, inputs=[], outputs=[audio_output], api_name="api_stop_audio_recording").then(
-            #     fn=add_message, inputs=[chatbot, chat_input], outputs=[chatbot, chat_input], api_name="api_addprompt_chathistory"
-            # ).then(
-            #     # First, generate the bot response
-            #     fn=generate_bot_response, inputs=[chatbot, choice, retrieval_mode, model_choice], outputs=[chatbot], api_name="api_generate_bot_response"
-            # ).then(
-            #     # Then, generate the TTS response based on the bot's response
-            #     fn=generate_tts_response, inputs=[chatbot, tts_choice], outputs=[audio_output], api_name="api_generate_tts_response"
-            # ).then(
-            #     fn=clear_textbox, inputs=[], outputs=[chat_input], api_name="api_clear_textbox"
-            # )
-
-            chat_input.submit(
-                fn=add_message, inputs=[chatbot, chat_input], outputs=[chatbot, chat_input], api_name="api_addprompt_chathistory"
-            ).then(
-                # First, generate the bot response
-                fn=generate_bot_response, inputs=[chatbot, choice, retrieval_mode, model_choice], outputs=[chatbot], api_name="api_generate_bot_response"
-            ).then(
-                fn=clear_textbox, inputs=[], outputs=[chat_input], api_name="api_clear_textbox"
-            )
-
-
-
-
-
-
-
-
-            # audio_input = gr.Audio(sources=["microphone"], streaming=True, type='numpy', every=0.1)
-            # audio_input.stream(transcribe_function, inputs=[state, audio_input], outputs=[state, chat_input], api_name="api_voice_to_text",visible=False)
-            
-        #     gr.Markdown("<h1 style='color: red;'>Example Prompts</h1>", elem_id="Example-Prompts")
-        #     gr.Examples(examples=examples, fn=insert_prompt,inputs=chat_input, outputs=chat_input)
-
-        # with gr.Column():
-        #     weather_output = gr.HTML(value=fetch_local_weather())
-        #     news_output = gr.HTML(value=fetch_local_news())
-        #     events_output = gr.HTML(value=fetch_local_events())
-
-        # with gr.Column():
-           
-          
-        #     # Call update_images during the initial load to display images when the interface appears
-        #     initial_images = update_images()
-            
-        #     # Displaying the images generated using Flux API directly
-        #     image_output_1 = gr.Image(value=initial_images[0], label="Image 1", elem_id="flux_image_1", width=400, height=400)
-        #     image_output_2 = gr.Image(value=initial_images[1], label="Image 2", elem_id="flux_image_2", width=400, height=400)
-        #     image_output_3 = gr.Image(value=initial_images[2], label="Image 3", elem_id="flux_image_3", width=400, height=400)
-    
-        #     # Refresh button to update images
-        #     refresh_button = gr.Button("Refresh Images")
-        #     refresh_button.click(fn=update_images, inputs=None, outputs=[image_output_1, image_output_2, image_output_3])
-            
-        # File upload component
         with gr.Column():
-                file_input = gr.File(label="Upload PDF", file_types=[".pdf"])
-                # Button to trigger processing
-                process_button = gr.Button("Process PDF and Upload")
-
-                # Dataframe to display uploaded document records
-                document_table = gr.Dataframe(headers=["Document Name", "Upload Time", "Chunks", "Pinecone Index"], interactive=False)
-
-            
-            
-                # Output textbox for results
-                output_textbox = gr.Textbox(label="Result")
-
-                # Define button click action
-                # process_button.click(fn=process_pdf, inputs=file_input, outputs=output_textbox)
-                process_button.click(fn=process_pdf, inputs=[file_input, gr.State([])], outputs=[document_table, output_textbox])
-            
-
-
-
-           
-
-demo.queue()
-demo.launch(show_error=True)
-
-
-
-
+            output_text = gr.Textbox(label="Status", interactive=False)
+            log_table = gr.DataFrame(label="Logs", interactive=False)
 
+    # Define action on button click
+    process_button.click(process_pdf, inputs=pdf_input, outputs=[output_text, log_table])
 
+# Launch the Gradio app
+demo.launch()