Update app.py

app.py

@@ -10,35 +10,43 @@ from langchain_community.vectorstores import Qdrant
 from langchain.chains import RetrievalQA
 from langchain.llms import HuggingFaceHub
 hf_token=os.environ["HUGGINGFACEHUB_API_TOKEN"]
-with open("brookline_data.txt", "r") as f:
-    data = f.read()
-text_splitter = RecursiveCharacterTextSplitter(chunk_size=100, chunk_overlap=20)
-splitted_data = text_splitter.split_text(data)
-embeddings = HuggingFaceEmbeddings(model_name='sentence-transformers/all-mpnet-base-v2')
-#retriever = Qdrant.as_retriever()
-client = QdrantClient(":memory:") 
-qdrant_vectorstore = Qdrant(
-    client,
-    embeddings.embed_query,
-    collection_name="my_documents" 
-)
-retriever = qdrant_vectorstore.as_retriever() 
-llm = HuggingFaceHub(repo_id="ahmadmac/Trained-T5-large", model_kwargs={"temperature": 0.5, "max_length": 512},huggingfacehub_api_token=hf_token)
-qna = RetrievalQA.from_chain_type(llm=llm, chain_type="stuff", retriever=retriever)
 prompt_template = """ you are a highly knowledgeable AI assistant. Engage in a conversation with the user. Your main goal is to provide clear and informative answers to the user's questions.
 User: {question}
 Assistant:"""
 prompt = PromptTemplate(template=prompt_template, input_variables=["question"])
 chain = LLMChain(llm=llm, prompt=prompt)
 
-def chatbot(question, chat_history):
-    result = qna(question)
-    if result['result']:
-        return result['result']
+# Load and split the document for the retrieval-based QA system
+with open("brookline_data.txt", "r") as f:
+    data = f.read()
+text_splitter = RecursiveCharacterTextSplitter(chunk_size=100, chunk_overlap=20)
+splitted_data = text_splitter.split_text(data)
 
-    response = chain.run(question)
-    return response
+# Generate embeddings for the split text
+embeddings = HuggingFaceEmbeddings(model_name='sentence-transformers/all-mpnet-base-v2')
+metadatas = [{"source": f"source_{i}"} for i in range(len(splitted_data))]
+documents = [Document(page_content=text, metadata=metadata) for text, metadata in zip(splitted_data, metadatas)]
 
+# Initialize Qdrant vector store
+qdrant = Qdrant.from_documents(
+    documents,
+    embeddings,
+    location=":memory:",
+    collection_name="my_documents",
+)
+retriever = qdrant.as_retriever()
+qna = RetrievalQA.from_chain_type(
+    llm=HuggingFaceHub(repo_id="ahmadmac/Trained-T5-large", model_kwargs={"temperature": 0.5, "max_length": 512}, 
+                       huggingfacehub_api_token=os.environ["HUGGINGFACEHUB_API_TOKEN"]),
+    chain_type="stuff",
+    retriever=retriever
+)
+def chatbot(question, chat_history):
+    response = chain.run(question)
+    retrieval_result = qna(question)
+    retrieval_answer = retrieval_result['result']
+    combined_response = f"{response}\n\nBased on the information available:\n{retrieval_answer}"
+    return combined_response
 demo = gr.ChatInterface(
     fn=chatbot,
     title="Chatbot",