Update pages/ChatPDF_Reader.py

pages/ChatPDF_Reader.py

@@ -11,12 +11,14 @@ from langchain.text_splitter import RecursiveCharacterTextSplitter
 
 import streamlit as st
 import sys,yaml,Utilities as ut
-
+import os
+print('HF_TOKEN',os.getenv('HF_TOKEN'))
 def get_data(query):
     chat_history = []
     initdict={}
     initdict = ut.get_tokens()
-    hf_token = initdict["hf_token"]
+    hf_token = os.getenv('HF_TOKEN') 
+    #hf_token = initdict["hf_token"]
     embedding_model_id = initdict["embedding_model"]
     chromadbpath = initdict["chatPDF_chroma_db"]
     llm_repo_id = initdict["llm_repoid"]
@@ -30,10 +32,10 @@ def get_data(query):
     db = Chroma(persist_directory=chromadbpath, embedding_function=embeddings)
     retriever = db.as_retriever(search_type="mmr", search_kwargs={'k': 2})
 
-    """llm = HuggingFaceHub(huggingfacehub_api_token=hf_token, 
-                        repo_id=llm_repo_id, model_kwargs={"temperature":0.2, "max_new_tokens":50})"""
+    llm = HuggingFaceHub(huggingfacehub_api_token=hf_token, 
+                        repo_id=llm_repo_id, model_kwargs={"temperature":0.2, "max_new_tokens":50})
     
-    llm = HuggingFaceHub(repo_id=llm_repo_id, model_kwargs={"temperature":0.2, "max_new_tokens":50})
+    #llm = HuggingFaceHub(repo_id=llm_repo_id, model_kwargs={"temperature":0.2, "max_new_tokens":50})
 
     # Create the Conversational Retrieval Chain
     qa_chain = ConversationalRetrievalChain.from_llm(llm, retriever,return_source_documents=True)