AIE1 midterm submission

README.md

@@ -1,11 +1,12 @@
----
-title: Aie1 Midterm
-emoji: 📊
-colorFrom: blue
-colorTo: gray
-sdk: docker
-pinned: false
-license: mit
----
+## Welcome to AIE-1 Midterm demo!  🏗️,  🚢,  & 🚀
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+Hello! 👋 I'm Bikram, and I'm excited to share my midterm project for the AI Engineering bootcamp at AI Makerspace!
+
+You can ask anything related to **NVIDIA 10-k Filings**  
+
+#### Sample Questionaire you can try 👨‍💻👩‍💻
+- Who is the E-VP, Operations - and how old are they?
+- What is the gross carrying amount of Total Amortizable Intangible Assets for Jan 29, 2023?
+- What is the Total Number of Shares Purchased between December 25, 2023 - January 28, 2024?
+- Summarize S&P 500 vs NVIDIA Corporation growth
+ 

app.py

@@ -0,0 +1,152 @@
+
+import os
+import chainlit as cl
+from langchain_openai import OpenAIEmbeddings
+
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+
+from langchain.retrievers import MultiQueryRetriever
+from langchain.chains.combine_documents import create_stuff_documents_chain
+
+from langchain_openai import ChatOpenAI
+from langchain.prompts import ChatPromptTemplate
+from langchain_community.vectorstores import FAISS
+from langchain_community.document_loaders import PyMuPDFLoader
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+
+from langchain_openai import ChatOpenAI
+from langchain.prompts.chat import ChatPromptTemplate
+from langchain.chains import create_retrieval_chain
+
+
+from dotenv import load_dotenv
+
+
+load_dotenv()
+
+
+# START CODE
+openai_chat_model = ChatOpenAI(model="gpt-3.5-turbo", temperature=0)
+embeddings = OpenAIEmbeddings(model="text-embedding-3-small")
+
+
+def process_file(file: str):
+
+    pypdf_loader = PyMuPDFLoader(file)
+    texts = pypdf_loader.load()
+
+    text_splitter = RecursiveCharacterTextSplitter(
+        chunk_size=700,
+        chunk_overlap=50
+    )
+
+    documents = text_splitter.split_documents(texts)
+
+    return documents
+
+
+async def RAG_pipeline(question: str, documents: str):
+
+    template = """
+    
+    Answer the question based only on the following context. 
+    
+    If you cannot answer the question with the context, please respond with 'I don't know, can you provide more context?'.
+
+    If the user question is not related to NVIDIA or NVIDIA 10-k Filings then please respond with "I can only help you with NVIDIA 10-k Filings report.\n\n".
+
+    Always answer in full sentence.
+    
+    If a user says things like "Ok" or "thank you" or "thank you" or anything that is related to  "phatic expressions" or "phatic communication" then respond with "No problem! Always happy to help."
+    
+    Always end your sentence with "\n\n\n Thank you for asking. What else can I help you regarding NVIDIA 10-k Filings report?".
+    
+    Context:
+    {context}
+
+    Question:
+    {input}
+    """
+
+    # Create prompt template
+    prompt = ChatPromptTemplate.from_template(template)
+
+    # Initialize FAISS vector store
+    vector_store = FAISS.from_documents(documents, embeddings)
+
+    # Initialize a retriever to retrieve similar context
+    retriever = vector_store.as_retriever()
+
+    # Initialize retriever using a multi-query approach with a language model.
+    retriever = MultiQueryRetriever.from_llm(
+        retriever=retriever, llm=openai_chat_model)
+
+    # Create a document chain using OpenAI chat model and a prompt
+    document_chain = create_stuff_documents_chain(openai_chat_model, prompt)
+
+    # Create a retrieval chain using a retriever and a document chain
+    retrieval_chain = create_retrieval_chain(retriever, document_chain)
+
+    # Send a request to OpenAI with the question 
+    response = retrieval_chain.invoke({"input": question})
+
+    # Making sure we have 'answer' params so that we can give proper response
+    if 'answer' in response:
+        llm_answer = response['answer']
+    else: 
+        llm_answer = '**EMPTY RESPONSE**'
+
+    print("llm_answer: ", llm_answer)
+
+    return llm_answer
+
+
+@cl.on_chat_start  # marks a function that will be executed at the start of a user session
+async def start_chat():
+    settings = {
+        "model": "gpt-3.5-turbo",
+        "temperature": 0,
+        "max_tokens": 500,
+        "top_p": 1,
+        "frequency_penalty": 0,
+        "presence_penalty": 0,
+    }
+
+    print("A new chat session has started!")
+
+    # Process our document for tokenization
+    documents = process_file(
+        "https://d18rn0p25nwr6d.cloudfront.net/CIK-0001045810/1cbe8fe7-e08a-46e3-8dcc-b429fc06c1a4.pdf")
+
+
+    # Save session
+    cl.user_session.set("documents", documents)
+    cl.user_session.set("settings", settings)
+
+
+@cl.on_message
+async def main(message: cl.Message):
+
+    print("Human asked: ", message.content)
+
+    msg = cl.Message(content="")
+    await msg.send()
+
+    # do some work
+    await cl.sleep(2)
+
+    # Retrieve session
+    document_chunks = cl.user_session.get("documents")
+
+    # Wait for OpenAI to return a response and the good ol' RAG stuff
+    response = await RAG_pipeline(message.content, document_chunks)
+
+    
+    # If there is a response then let the user know else fallback to else statement!
+    if response:
+        await cl.Message(content=response).send()
+    else:
+        cl.Message(
+            content="Something went wrong! please kindly refresh and try again 🤝").send()
+
+ 

chainlit.md

@@ -0,0 +1,14 @@
+ 
+
+## Welcome to AIE-1 Midterm demo!  🏗️,  🚢,  & 🚀
+
+Hello! 👋 I'm Bikram, and I'm excited to share my midterm project for the AI Engineering bootcamp at AI Makerspace!
+
+You can ask anything related to **NVIDIA 10-k Filings**  
+
+#### Sample Questionaire you can try 👨‍💻👩‍💻
+- Who is the E-VP, Operations - and how old are they?
+- What is the gross carrying amount of Total Amortizable Intangible Assets for Jan 29, 2023?
+- What is the Total Number of Shares Purchased between December 25, 2023 - January 28, 2024?
+- Summarize S&P 500 vs NVIDIA Corporation growth
+ 

requirements.txt

@@ -0,0 +1,9 @@
+openai
+langchain 
+langchain-openai 
+langchain_core
+langchain-community 
+langchainhub 
+faiss_cpu 
+pymupdf 
+pandas