app.py

# Import the necessary Libraries
from warnings import filterwarnings
filterwarnings('ignore')
import os
import uuid
import json
import gradio as gr
import pandas as pd
from huggingface_hub import CommitScheduler
from pathlib import Path
from langchain.embeddings import SentenceTransformerEmbeddings
from langchain.vectorstores import Chroma
# from langchain.llms import OpenAI

from langchain_openai import ChatOpenAI
from langchain.schema import HumanMessage, AIMessage, SystemMessage

# Create Client
import os
os.environ['OPENAI_API_KEY']  = "gl-U2FsdGVkX1+0bNWD6YsVLZUYsn0m1WfLxUzrP0xUFbtWFAfk9Z1Cz+mD8u1yqKtV"; # e.g. gl-U2FsdGVkX19oG1mRO+LGAiNeC7nAeU8M65G4I6bfcdI7+9GUEjFFbplKq48J83by
os.environ["OPENAI_BASE_URL"] =  "https://aibe.mygreatlearning.com/openai/v1" # e.g. "https://aibe.mygreatlearning.com/openai/v1";

model_name = 'gpt-4o-mini' # e.g. 'gpt-3.5-turbo'

llm = ChatOpenAI(model_name=model_name, temperature=0)  # Set temperature to 0 for deterministic output

# Define the embedding model and the vectorstore
embedding_model = SentenceTransformerEmbeddings(model_name='thenlper/gte-large')

vectorstore_persisted = Chroma(
    collection_name='10k_reports',
    persist_directory='10k_reports_db',
    embedding_function=embedding_model
)

#
##
#

# Prepare the logging functionality
log_file = Path("logs/") / f"data_{uuid.uuid4()}.json"
log_folder = log_file.parent

scheduler = CommitScheduler(
    repo_id="eric-green-rag-financial-analyst",
    repo_type="dataset",
    folder_path=log_folder,
    path_in_repo="data",
    every=2
)

# Define the Q&A system message
# Create a system message for the LLM
qna_system_message = """
You are an assistant to a tech industry financial analyst. Your task is to provide relevant information about a set of companies AWS, Google, IBM, Meta, Microsoft.

User input will include the necessary context for you to answer their questions. This context will begin with the token: ###Context.
The context contains references to specific portions of documents relevant to the user's query, along with source links.
The source for a context will begin with the token ###Source.

When crafting your response:
1. Select only context relevant to answer the question.
2. Include the source links in your response.
3. User questions will begin with the token: ###Question.
4. If the question is irrelevant to financial report information for the 5 companies, respond with "I am unable to locate relevent information.  I answer questions related to the financial performance of AWS, Google, IBM, Meta and Microsoft."

Please adhere to the following guidelines:
- Your response should only be about the question asked and nothing else.
- Answer only using the context provided.
- Do not mention anything about the context in your final answer.
- If the answer is not found in the context, it is very very important for you to respond with "I am unable to locate a relevent answer."
- Always quote the source when you use the context. Cite the relevant source at the end of your response under the section - Source:
- Do not make up sources. Use the links provided in the sources section of the context and nothing else. You are prohibited from providing other links/sources.

Here is an example of how to structure your response:

Answer:
[Answer]

Source:
[Source]
"""

# Define the user message template
# Create a message template
qna_user_message_template = """
###Context
{context}

###Question
{question}
"""

# Define the llm_query function that runs when 'Submit' is clicked or when a API request is made
def llm_query(user_input,company):

    filter = "dataset/"+company+"-10-k-2023.pdf"
    
    relevant_document_chunks = vectorstore_persisted.similarity_search(user_input, k=5, filter={"source":filter})

    # 1 - Create context_for_query
    context_list = [d.page_content + "\n ###Source: " + str(d.metadata['page']) + "\n\n " for d in relevant_document_chunks]

    context_for_query = ". ".join(context_list)

    # 2 - Create messages
    prompt = [
        SystemMessage(content=qna_system_message),
        HumanMessage(content=qna_user_message_template.format(
            context=context_for_query,
            question=user_input
            )
        )
    ]

    # Get response from the LLM
    try:
        # Call the chat model with the message
        response = llm(prompt)
        
        llm_response = response.content.strip()

    except Exception as e:
        
        llm_response = f'Sorry, I encountered the following error: \n {e}'

    print(llm_response)

    # While the prediction is made, log both the inputs and outputs to a local log file
    # While writing to the log file, ensure that the commit scheduler is locked to avoid parallel access
    with scheduler.lock:
        with log_file.open("a") as f:
            f.write(json.dumps(
                {
                    'user_input': user_input,
                    'retrieved_context': context_for_query,
                    'model_response': llm_response
                }
            ))
            f.write("\n")

    return llm_response

# Set-up the Gradio UI
textbox = gr.Textbox(label='Question:') # Create a textbox for user input
company = gr.Radio(label='Company:', choices=["aws", "google", "IBM", "Meta", "msft"], value="aws") # Create a radio button for company selection

# Create Gradio interface
# For the inputs parameter of Interface provide [textbox,company] with outputs parameter of Interface provide prediction
demo = gr.Interface(fn=llm_query, inputs=[textbox, company], outputs="text", title="FY23 Financial Analyst Assistant", description="Ask questions about the financial performance of AWS, Google, IBM, Meta, and Microsoft based on their 10-K reports.\n\nPlease enter a question below.", theme=gr.themes.Soft())

demo.queue()
demo.launch()