Update app.py

app.py

@@ -12,6 +12,7 @@ from langchain_core.output_parsers import StrOutputParser
 from langchain_core.prompts import ChatPromptTemplate
 
 from langchain_community.embeddings import HuggingFaceEmbeddings
+from langchain import hub
 
 def method_get_website_text(urls):
     # Convert string of URLs to list
@@ -40,32 +41,52 @@ def method_get_vectorstore(document_chunks):
     # create a vectorstore from the chunks
     vector_store = Chroma.from_documents(document_chunks, embeddings)
     return vector_store
-    
+
 def get_context_retriever_chain(vector_store,question):
     # Initialize the retriever
     retriever = vector_store.as_retriever()
     
-    # Define the RAG template
-    after_rag_template = """Answer the question based only on the following context:
-    {context}
-    Question: {question}
-    """
-    
-    # Create the RAG prompt template
-    after_rag_prompt = ChatPromptTemplate.from_template(after_rag_template)
+    # Define the RAG template and RAG prompt template
+    prompt = hub.pull("rlm/rag-prompt")
     
     # Initialize the Hugging Face language model (LLM)
     llm = HuggingFaceHub(repo_id="mistralai/Mistral-7B-Instruct-v0.2", model_kwargs={"temperature":0.6, "max_length":512})
     
     # Construct the RAG pipeline
-    after_rag_chain = (
+    rag_chain = (
         {"context": retriever, "question": RunnablePassthrough()}
-        | after_rag_prompt
+        | prompt
         | llm
         | StrOutputParser()
     )
+    return rag_chain.invoke(str(question))
+
+    
+# def get_context_retriever_chain(vector_store,question):
+#     # Initialize the retriever
+#     retriever = vector_store.as_retriever()
+    
+#     # Define the RAG template
+#     after_rag_template = """Answer the question based only on the following context:
+#     {context}
+#     Question: {question}
+#     """
+    
+#     # Create the RAG prompt template
+#     after_rag_prompt = ChatPromptTemplate.from_template(after_rag_template)
+    
+#     # Initialize the Hugging Face language model (LLM)
+#     llm = HuggingFaceHub(repo_id="mistralai/Mistral-7B-Instruct-v0.2", model_kwargs={"temperature":0.6, "max_length":512})
+    
+#     # Construct the RAG pipeline
+#     after_rag_chain = (
+#         {"context": retriever, "question": RunnablePassthrough()}
+#         | after_rag_prompt
+#         | llm
+#         | StrOutputParser()
+#     )
     
-    return after_rag_chain.invoke(question)
+#     return after_rag_chain.invoke(question)
 
 def main():
     st.set_page_config(page_title="Chat with websites", page_icon="🤖")