app.py

from langchain.document_loaders import PyPDFLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.embeddings import HuggingFaceEmbeddings
from langchain.vectorstores import Pinecone 
from sentence_transformers import SentenceTransformer
from langchain.chains.question_answering import load_qa_chain
import pinecone
import os
from langchain.llms import LlamaCpp
from langchain.callbacks.manager import CallbackManager
from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
from huggingface_hub import hf_hub_download
import streamlit as st
from streamlit_chat import message

# loader = OnlinePDFLoader("food.pdf")
loader = PyPDFLoader("Ramesh_kumar_Resume.pdf")
data = loader.load()
# data
text_splitter=RecursiveCharacterTextSplitter(chunk_size=200, chunk_overlap=0)
docs=text_splitter.split_documents(data)
len(docs)

os.environ["HUGGINGFACEHUB_API_TOKEN"] = st.secrets["HF_TOKEN"]
PINECONE_API_KEY = os.environ.get('PINECONE_API_KEY', st.secrets["PINECONE"])
PINECONE_API_ENV = os.environ.get('PINECONE_API_ENV', 'gcp-starter')
os.environ['PINECONE_API_KEY'] = PINECONE_API_KEY

embeddings=HuggingFaceEmbeddings(model_name='sentence-transformers/all-MiniLM-L6-v2')

# initialize pinecone
pinecone.Pinecone(
    api_key=PINECONE_API_KEY,  # find at app.pinecone.io
    environment=PINECONE_API_ENV  # next to api key in console
)
index_name = "testindex" # put in the name of your pinecone index here


docsearch= Pinecone.from_texts([t.page_content for t in docs],embeddings, index_name=index_name)

query="what languages ramesh know?"

docs=docsearch.similarity_search(query,k=1)

# docs

# !CMAKE_ARGS="-DLLAMA_CUBLAS=on" FORCE_CMAKE=1 pip install llama-cpp-python --force-reinstall --upgrade --no-cache-dir --verbose



# Callbacks support token-wise streaming
callback_manager = CallbackManager([StreamingStdOutCallbackHandler()])
# Verbose is required to pass to the callback manager

model_name_or_path = "anakin87/gemma-2b-orpo-GGUF"
filename = "gemma-2b-orpo.Q5_K_M.gguf" # the model is in bin format
model_path = hf_hub_download(repo_id=model_name_or_path, filename=filename)

n_gpu_layers = 40  # Change this value based on your model and your GPU VRAM pool.
n_batch = 256  # Should be between 1 and n_ctx, consider the amount of VRAM in your GPU.

# Loading model,


query=st.text_input("Ask questions:")





if query:
    search=docsearch.similarity_search(query)
 
    llm = LlamaCpp(
    model_path=model_path,
    max_tokens=256,
    # n_gpu_layers=n_gpu_layers,
    # n_batch=n_batch,
    callback_manager=callback_manager,
    n_ctx=1024,
    verbose=True,
    )

    chain=load_qa_chain(llm, chain_type="stuff")
    
    response = chain.run(input_documents=search, question=query)

    st.write(response)

    # st.session_state['messages'].append({"role": "user", "content": query})
    # st.session_state['messages'].append({"role": "assistant", "content": response})

    # response_container = st.container()
    # # container for text box
    # container = st.container()

    # with container:
    #     if query:
    #         output = response
    #         st.session_state['past'].append(query)
    #         st.session_state['generated'].append(output)
        
    # if st.session_state['generated']:
    #     with response_container:
    #         for i in range(len(st.session_state['generated'])):
    #             message(st.session_state["past"][i], is_user=True, key=str(i) + '_user')
    #             message(st.session_state["generated"][i], key=str(i))