update

app.py

@@ -2,17 +2,17 @@ import chainlit as cl
 from chainlit.playground.providers import ChatOpenAI
 from dotenv import load_dotenv
 # from langchain_community.document_loaders import PyMuPDFLoader
-from langchain.text_splitter import RecursiveCharacterTextSplitter
-import tiktoken
-from langchain.prompts import ChatPromptTemplate
-from operator import  itemgetter
-from langchain_core.runnables import RunnablePassthrough
-from langchain import ChatOpenAI, OpenAIEmbeddings
-from langchain_community.vectorstores import FAISS
-from langchain.retrievers import MultiQueryRetriever
-
-from langchain.prompts import ChatPromptTemplate
-from langchain.retrievers import MultiQueryRetriever
+# from langchain.text_splitter import RecursiveCharacterTextSplitter
+# import tiktoken
+# from langchain.prompts import ChatPromptTemplate
+# from operator import  itemgetter
+# from langchain_core.runnables import RunnablePassthrough
+# from langchain import ChatOpenAI, OpenAIEmbeddings
+# from langchain_community.vectorstores import FAISS
+# from langchain.retrievers import MultiQueryRetriever
+
+# from langchain.prompts import ChatPromptTemplate
+# from langchain.retrievers import MultiQueryRetriever
 
 template = """
 you can only answer questions related to what's in the context. If it's not in the context, then you would reply with 
@@ -34,43 +34,43 @@ init_settings = {
     "presence_penalty": 0,
 }
 
-embeddings = OpenAIEmbeddings(model="text-embedding-3-small")
+# embeddings = OpenAIEmbeddings(model="text-embedding-3-small")
 
 
 load_dotenv()
 
-def tiktoken_len(text):
-    tokens = tiktoken.encoding_for_model("gpt-3.5-turbo").encode(
-        text,
-    )
-    return len(tokens)
+# def tiktoken_len(text):
+#     tokens = tiktoken.encoding_for_model("gpt-3.5-turbo").encode(
+#         text,
+#     )
+#     return len(tokens)
 
 
 @cl.on_chat_start
 async def main():
     model = ChatOpenAI(streaming=True)
     
-    prompt = ChatPromptTemplate.from_template(template)
+    # prompt = ChatPromptTemplate.from_template(template)
 
-    nvida_doc = PyMuPDFLoader('../docs/nvidia-document.pdf')
-    data = nvida_doc.load()
+    # nvida_doc = PyMuPDFLoader('../docs/nvidia-document.pdf')
+    # data = nvida_doc.load()
 
-    text_splitter = RecursiveCharacterTextSplitter(
-    chunk_size = 1700,
-    chunk_overlap = 0,
-    length_function = tiktoken_len)
+    # text_splitter = RecursiveCharacterTextSplitter(
+    # chunk_size = 1700,
+    # chunk_overlap = 0,
+    # length_function = tiktoken_len)
     
-    nvidia_doc_chunks = text_splitter.split_documents(data)
+    # nvidia_doc_chunks = text_splitter.split_documents(data)
 
-    vector_store = FAISS.from_documents(nvidia_doc_chunks, embedding=embeddings)
+    # vector_store = FAISS.from_documents(nvidia_doc_chunks, embedding=embeddings)
     
-    retriever = vector_store.as_retriever()
-    advanced_retriever = MultiQueryRetriever.from_llm(retriever=retriever, llm=model)
+    # retriever = vector_store.as_retriever()
+    # advanced_retriever = MultiQueryRetriever.from_llm(retriever=retriever, llm=model)
 
-    runnable = (
-    {"context": itemgetter("question") | retriever, "question": itemgetter("question")}
-    | RunnablePassthrough.assign(context=itemgetter("context"))
-    | {"response": prompt | model, "context": itemgetter("context")})
+    # runnable = (
+    # {"context": itemgetter("question") | retriever, "question": itemgetter("question")}
+    # | RunnablePassthrough.assign(context=itemgetter("context"))
+    # | {"response": prompt | model, "context": itemgetter("context")})
 
     # retrieval_qa_prompt = hub.pull("langchain-ai/retrieval-qa-chat")
     # document_chain = create_stuff_documents_chain(model, retrieval_qa_prompt)
@@ -79,7 +79,7 @@ async def main():
     # cl.user_session.set("settings", init_settings)
     # cl.user_session.set("nvidia_doc", data)
 
-    cl.user_session.set("runnable", runnable)
+    # cl.user_session.set("runnable", runnable)
 
 
 
@@ -87,9 +87,9 @@ async def main():
 async def on_message(message: cl.Message):
     # settings = cl.user_session.get("settings")
     # nvida_doc = cl.user_session.get("nvidia_doc")
-    runnable = cl.user_session.get("runnable")
+    # runnable = cl.user_session.get("runnable")
 
-    msg = cl.Message(content="")
+    # msg = cl.Message(content="")
 
     
     # async for chunk in runnable.astream(
@@ -100,10 +100,10 @@ async def on_message(message: cl.Message):
 
     # await msg.send()
 
-    inputs = {"question": message.content}
-    result = await runnable.ainvoke(inputs)
-    msg = cl.Message(content=result["response"].content)
-    await msg.send()
+    # inputs = {"question": message.content}
+    # result = await runnable.ainvoke(inputs)
+    # msg = cl.Message(content=result["response"].content)
+    # await msg.send()