helper.py

import tempfile 
import streamlit as st 
from streamlit_chat import message

import torch
import torch.nn

import transformers
from transformers import (
    AutoModelForCausalLM,
    AutoTokenizer,
    BitsAndBytesConfig,
    HfArgumentParser,
    TrainingArguments,
    pipeline,
    logging,
)


import pandas as pd
import numpy as np
import os
import io

from langchain.document_loaders import TextLoader
from langchain import PromptTemplate
from langchain.text_splitter import CharacterTextSplitter
from langchain.document_loaders import PyPDFLoader
from langchain.embeddings import HuggingFaceEmbeddings
from langchain.vectorstores import FAISS
from langchain.chains.question_answering import load_qa_chain
from langchain.chains import RetrievalQA
from langchain import HuggingFacePipeline


def pdf_loader(file_path): 
  '''This is a function for loading the PDFs
  Params: 
  file_path: The path of the PDF file
  '''
  output_file = "Loaded_PDF.txt"
  loader = PyPDFLoader(file_path)
  pdf_file_as_loaded_docs = loader.load()
  return pdf_file_as_loaded_docs

def splitDoc(loaded_docs):
    '''This is a function that creates the chunks of our loaded Document
    Params: 
    loaded_docs:The loaded document from the pdf_loader function'''
    splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=10)
    chunked_docs = splitter.split_documents(loaded_docs)
    return chunked_docs

def makeEmbeddings(chunked_docs):
    '''This is a functuon for making the embeddings of the chunked document
    Params: 
    chunked_docs:The chunked docs'''
    embedder = HuggingFaceEmbeddings()
    vector_store = FAISS.from_documents(chunked_docs, embedder)#making a FAISS based vector data
    return vector_store


def create_flan_t5_base(load_in_8bit=False):
    ''''Loading the Flan T5 base in the form of pipeline'''
    # Wrap it in HF pipeline for use with LangChain
    model="google/flan-t5-base"
    tokenizer = AutoTokenizer.from_pretrained(model)
    return pipeline(
        task="text2text-generation",
        model=model,
        tokenizer = tokenizer,
        max_new_tokens=100,
        model_kwargs={ "load_in_8bit": load_in_8bit, "max_length": 512, "temperature": 0.}
    )