Update app.py

app.py

@@ -3,19 +3,20 @@ from langchain.embeddings.openai import OpenAIEmbeddings
 from langchain.vectorstores import Qdrant
 from langchain.text_splitter import CharacterTextSplitter
 from langchain.chains import ConversationalRetrievalChain
-from langchain.llms import OpenAI
 from langchain.memory import ConversationBufferMemory
 from qdrant_client import QdrantClient
 from qdrant_client.http import models as qdrant_models
 import os
 from langchain.document_loaders import PyPDFLoader
 import gradio as gr
+from langchain_qdrant import QdrantVectorStore, FastEmbedSparse, RetrievalMode
 
+# Set OpenAI API Key
 openai_api_key = os.getenv('OPENAI_API_KEY')
-openai_api_key = openai_api_key
+os.environ["OPENAI_API_KEY"] = openai_api_key
 
+# Load PDF documents
 pdf_folder_path = "files"
-
 documents = []
 for filename in os.listdir(pdf_folder_path):
     if filename.endswith(".pdf"):
@@ -23,24 +24,25 @@ for filename in os.listdir(pdf_folder_path):
         loader = PyPDFLoader(file_path)
         documents.extend(loader.load())
 
+# Split the documents into manageable chunks
 text_splitter = CharacterTextSplitter(chunk_size=512, chunk_overlap=25)
 docs = text_splitter.split_documents(documents)
 
+# Initialize embeddings and Qdrant client
 embeddings = OpenAIEmbeddings()
-
 qdrant_client = QdrantClient(":memory:")
 
+# Recreate Qdrant collection
 qdrant_client.recreate_collection(
     collection_name="langchain_collection",
     vectors_config=qdrant_models.VectorParams(size=1536, distance=qdrant_models.Distance.COSINE)
 )
 
-from langchain_qdrant import QdrantVectorStore,FastEmbedSparse,RetrievalMode
-
+# Set up the sparse embeddings for hybrid retrieval
 sparse_embeddings = FastEmbedSparse(model_name="Qdrant/bm25")
 
-
-vector_store =  QdrantVectorStore.from_documents(
+# Initialize the vector store with hybrid retrieval mode
+vector_store = QdrantVectorStore.from_documents(
     docs,
     embedding=embeddings,
     sparse_embedding=sparse_embeddings,
@@ -48,32 +50,33 @@ vector_store =  QdrantVectorStore.from_documents(
     collection_name="langchain_collection",
     retrieval_mode=RetrievalMode.HYBRID,
 )
+
+# Set up conversational memory
 memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True)
 
-retriever = vector_store.as_retriever()
-# llm = OpenAI(temperature=0.4)
+# Set up the retriever
+retriever = vector_store.as_retriever(search_type="hybrid", search_kwargs={"k": 3})
+
+# Set up the language model
 llm = ChatOpenAI(model="gpt-3.5-turbo", temperature=0.3)
+
+# Set up the conversational retrieval chain with memory
 conversational_chain = ConversationalRetrievalChain.from_llm(
     llm=llm,
     retriever=retriever,
-    memory=memory
+    memory=memory,
+    verbose=True 
 )
 
-query = "What is COMVIVA CDR"
-response = conversational_chain.invoke({"question": query})
-print(response['answer'])
-
-
-
 
 def chat_with_ai(user_input, chat_history):
-    response = conversational_chain.invoke({"question":user_input})
-
-    chat_history.append((user_input, str(response['answer'])))
+    response = conversational_chain({"question": user_input})
+    
+    chat_history.append((user_input, response['answer']))
 
     return chat_history, ""
 
-
+# Gradio interface
 def gradio_chatbot():
     with gr.Blocks() as demo:
         gr.Markdown("# Chat Interface for Langchain")
@@ -87,10 +90,11 @@ def gradio_chatbot():
 
         chat_history = gr.State([])
 
+        # Bind button and textbox to chat function
         submit_button.click(chat_with_ai, inputs=[user_input, chat_history], outputs=[chatbot, user_input])
-
         user_input.submit(chat_with_ai, inputs=[user_input, chat_history], outputs=[chatbot, user_input])
 
     return demo
 
+# Launch Gradio interface
 gradio_chatbot().launch(debug=True)