Update app.py

app.py

@@ -1,4 +1,4 @@
-from langchain.document_loaders import TextLoader
+
 import streamlit as st
 from streamlit_chat import message as st_message
 import pandas as pd
@@ -12,6 +12,9 @@ import json
 import torch
 from tqdm.auto import tqdm
 from langchain.text_splitter import RecursiveCharacterTextSplitter
+from dotenv import load_dotenv
+from PyPDF2 import PdfReader
+from langchain.text_splitter import CharacterTextSplitter
 
 
 from transformers import AutoModel
@@ -34,76 +37,33 @@ from langchain.chains.conversational_retrieval.prompts import CONDENSE_QUESTION_
 
 
 
-prompt_template = """
-
-You are the chatbot and the advanced legal assitant that can give answers to all the legal questions a common citizen would have . Your job is to give answers when questions about General legal information, Family law, Employment law, Consumer rights, Housing and tenancy, Personal injury, Wills and estates, Criminal law are asked.
-Your job is to answer questions only and only related to Legal aspect. Anything unrelated should be responded with the fact that your main job is solely to provide assistance regarding Legality.
-MUST only use the following pieces of context to answer the question at the end. If the answers are not in the context or you are not sure of the answer, just say that you don't know, don't try to make up an answer.
-{context}
-Question: {question}
-When encountering abusive, offensive, or harmful language, such as fuck, bitch,etc,  just politely ask the users to maintain appropriate behaviours.
-Always make sure to elaborate your response and use vibrant, positive tone to represent good branding of the school.
-Never answer with any unfinished response
-Answer:
-"""
-
-PROMPT = PromptTemplate(
-    template=prompt_template, input_variables=["context", "question"]
-)
-chain_type_kwargs = {"prompt": PROMPT}
-
-
-#This version includes the memory and custom prompt, representing the final version
-
-import streamlit as st
-from streamlit_chat import message as st_message
-import pandas as pd
-import numpy as np
-import datetime
-import gspread
-import pickle
-import os
-import csv
-import json
-import torch
-from tqdm.auto import tqdm
-from langchain.text_splitter import RecursiveCharacterTextSplitter
-
 
 # from langchain.vectorstores import Chroma
 from langchain.vectorstores import FAISS
-from langchain.embeddings import HuggingFaceInstructEmbeddings
+from langchain.embeddings import HuggingFaceInstructEmbeddings,OpenAIEmbeddings
 
 
-from langchain import HuggingFacePipeline
 from langchain.chains import RetrievalQA
-from langchain.prompts import PromptTemplate
-from langchain.memory import ConversationBufferWindowMemory
 
 
-from langchain.chains import LLMChain
-from langchain.chains import ConversationalRetrievalChain
-from langchain.chains.question_answering import load_qa_chain
-from langchain.chains.conversational_retrieval.prompts import CONDENSE_QUESTION_PROMPT
+
 
 
 
 prompt_template = """
 
-You are the chatbot and the face of Asian Institute of Technology (AIT). Your job is to give answers to prospective and current students about the school.
-Your job is to answer questions only and only related to the AIT. Anything unrelated should be responded with the fact that your main job is solely to provide assistance regarding AIT.
+You are the chatbot and the advanced legal assitant that can give answers to all the legal questions a common citizen would have . Your job is to give answers when questions about General legal information, Family law, Employment law, Consumer rights, Housing and tenancy, Personal injury, Wills and estates, Criminal law are asked.
+Your job is to answer questions only and only related to Legal aspect. Anything unrelated should be responded with the fact that your main job is solely to provide assistance regarding Legality.
 MUST only use the following pieces of context to answer the question at the end. If the answers are not in the context or you are not sure of the answer, just say that you don't know, don't try to make up an answer.
-
-
 {context}
 Question: {question}
-
 When encountering abusive, offensive, or harmful language, such as fuck, bitch,etc,  just politely ask the users to maintain appropriate behaviours.
 Always make sure to elaborate your response and use vibrant, positive tone to represent good branding of the school.
 Never answer with any unfinished response
-
 Answer:
 """
+
+
 PROMPT = PromptTemplate(
     template=prompt_template, input_variables=["context", "question"]
 )
@@ -115,33 +75,22 @@ st.set_page_config(
     page_icon = '🕵')
 
 
-
-
-banksandbanking_web_documents = None
-
-@st.cache
-def load_scraped_web_info():
-  global banksandbanking_web_documents 
-  text_splitter = RecursiveCharacterTextSplitter(...)
-  chunked_text = text_splitter.create_documents(banksandbanking_web_documents)
-  return chunked_text
-
-# Load pickle only once at startup
-with open("Banksandbanking.htm", "rb") as fp:
-  banksandbanking_web_documents = pickle.load(fp)
-
+@st.cache_data
+def get_pdf_text(pdf_docs):
+    text = ""
+    for pdf in pdf_docs:
+        pdf_reader = PdfReader(pdf)
+        for page in pdf_reader.pages:
+            text += page.extract_text()
+    return text
 
 @st.cache_resource
-def load_embedding_model():
-    embedding_model = HuggingFaceInstructEmbeddings(model_name='hkunlp/instructor-base',
-                                                model_kwargs = {'device': torch.device('cuda' if torch.cuda.is_available() else 'cpu')})
-    return embedding_model
-
-@st.cache_data
-def load_faiss_index():
-    vector_database = FAISS.load_local("faiss_index_web_and_curri_new", embedding_model) #CHANGE THIS FAISS EMBEDDED KNOWLEDGE
+def get_vectorstore(text_chunks):
+    embeddings = HuggingFaceInstructEmbeddings(model_name="hkunlp/instructor-xl")
+    vector_database = FAISS.from_texts(texts=text_chunks, embedding=embeddings)
     return vector_database
 
+
 @st.cache_resource
 def load_llm_model():
     llm = HuggingFacePipeline.from_model_id(model_id= 'PyaeSoneK/pythia_70m_legalQA',