llm_ans.py

import os
import glob
import textwrap
import time
import langchain
from langchain.document_loaders import PyPDFLoader, DirectoryLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain import PromptTemplate, LLMChain
from langchain.vectorstores import FAISS
from langchain.llms import HuggingFacePipeline
from langchain.embeddings import HuggingFaceInstructEmbeddings
from langchain.chains import RetrievalQA
import torch
import transformers
from model import qa_chain

def wrap_text_preserve_newlines(text, width=700):
    # Split the input text into lines based on newline characters
    lines = text.split('\n')

    # Wrap each line individually
    wrapped_lines = [textwrap.fill(line, width=width) for line in lines]

    # Join the wrapped lines back together using newline characters
    wrapped_text = '\n'.join(wrapped_lines)

    return wrapped_text


def process_llm_response(llm_response):
    ans = wrap_text_preserve_newlines(llm_response['result'])
    
    sources_used = ' \n'.join(
        [
            source.metadata['source'].split('/')[-1][:-4]
            + ' - page: '
            + str(source.metadata['page'])
            for source in llm_response['source_documents']
        ]
    )
    
    ans = ans + '\n\nSources: \n' + sources_used
    return ans

def llm_ans(query):
    start = time.time()
    
    llm_response = qa_chain.invoke(query)
    ans = process_llm_response(llm_response)
    
    end = time.time()

    time_elapsed = int(round(end - start, 0))
    time_elapsed_str = f'\n\nTime elapsed: {time_elapsed} s'
    ans_loc=ans.find("Answer:")
    ans_loc+=len("Answer: ")
    return ans[ans_loc:]


# query = "what are computer networks?"
# result=llm_ans(query)
# print(result)
# print(type(result))