app.py

# Import the necessary libraries
import subprocess
import sys

# Function to install a package using pip
def install(package):
    subprocess.check_call([sys.executable, "-m", "pip", "install", package])

# Install required packages
try:
    install("gradio")
    install("openai==1.23.2")
    install("tiktoken==0.6.0")
    install("pypdf==4.0.1")
    install("langchain==0.1.1")
    install("langchain-community==0.0.13")
    install("chromadb==0.4.22")
    install("sentence-transformers==2.3.1")
except subprocess.CalledProcessError as e:
    print(f"An error occurred: {e}")

import gradio as gr
import os
import uuid
import json
import pandas as pd
import subprocess
from openai import OpenAI
from huggingface_hub import HfApi
from huggingface_hub import CommitScheduler
from langchain_community.embeddings.sentence_transformer import (
    SentenceTransformerEmbeddings
)
from langchain_community.vectorstores import Chroma
#from google.colab import userdata, drive
from pathlib import Path
from langchain.document_loaders import PyPDFDirectoryLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
import json
import tiktoken
import pandas as pd
import tiktoken

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

# If dataset directory exixts, remove it and all of the contents within

#if os.path.exists('dataset'):
#  !rm -rf dataset

# If collection_db exists, remove it and all of the contents within

#if os.path.exists('collection_db'):
#    !rm -rf dataset

#Mount the Google Drive
#drive.mount('/content/drive')

#Upload Dataset-10k.zip and unzip it dataset folder using -d option
#!unzip Dataset-10k.zip -d dataset

import subprocess

# Command to unzip the file
command = "unzip kgauvin603/rag-10k-analysis/Dataset-10k.zip -d dataset"

# Execute the command
try:
    subprocess.run(command, check=True, shell=True)
except subprocess.CalledProcessError as e:
    print(f"An error occurred: {e}")

# Provide pdf_folder_location
pdf_folder_location = "kgauvin603/rag-10k-analysis/dataset"

# Load the directory to pdf_loader
pdf_loader = PyPDFDirectoryLoader(pdf_folder_location)

# Create text_splitter using recursive splitter
text_splitter = RecursiveCharacterTextSplitter.from_tiktoken_encoder(
    encoding_name='cl100k_base',
    chunk_size=512,
    chunk_overlap=16
)

# Create chunks
report_chunks = pdf_loader.load_and_split(text_splitter)

#Create a Colelction Name
collection_name = 'collection'

# Create the vector Database
vectorstore = Chroma.from_documents(
    report_chunks,
    embedding_model,
    collection_name=collection_name,
    persist_directory='./collection_db'
)

# Persist the DB
vectorstore.persist()

vectorstore_persisted = Chroma(
    collection_name=collection_name,
    persist_directory='./collection_db',
    embedding_function=embedding_model
)

retriever = vectorstore_persisted.as_retriever(
    search_type='similarity',
    search_kwargs={'k': 5}
)

#Mount the Google Drive
#drive.mount('/content/drive')

#Copy the persisted database to your drive
#command = "!cp -r collection_db /content/drive/MyDrive/"
# Execute the command
#try:
#    subprocess.run(command, check=True, shell=True)
#except subprocess.CalledProcessError as e:
#    print(f"An error occurred: {e}")

# Get anyscale api key
anyscale_api_key = userdata.get('dev-work')

# Initialise the client
client = OpenAI(
    base_url="https://api.endpoints.anyscale.com/v1",
    api_key=anyscale_api_key
)
#Provide the model name
model_name = 'mlabonne/NeuralHermes-2.5-Mistral-7B'

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

# Load the persisted DB
persisted_vectordb_location = './collection_db'

#Create a Colelction Name
collection_name = 'collection'

# Load the persisted DB
vectorstore_persisted = Chroma(
    collection_name=collection_name,
    persist_directory=persisted_vectordb_location,
    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="kgauvin603/rag-10k-analysis",
    repo_type="dataset",
    folder_path=log_folder,
    path_in_repo="data",
    every=2,
    token=hf_token
)

# Define the Q&A system message
qna_system_message = """You are an assistant to a financial services firm who answers user queries on annual reports.
User input will have the context required by you to answer user questions.
This context will begin with the token: ###Context.
The context contains references to specific portions of a document relevant to the user query.

User questions will begin with the token: ###Question.

Please answer only using the context provided in the input. Do not mention anything about the context in your final answer.

If the answer is not found in the context, respond "I don't know".
"""

# Create a message template
qna_user_message_template = """
###Context
Here are some documents that are relevant to the question mentioned below.
{context}

###Question
{question}
"""

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

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

    # Create context_for_query
    context_list = [d.page_content for d in relevant_document_chunks]
    context_for_query = ". ".join(context_list)

    # Create messages
    prompt = [
        {'role': 'system', 'content': qna_system_message},
        {'role': 'user', 'content': qna_user_message_template.format(
            context=context_for_query,
            question=user_input
        )}
    ]

    try:
        response = client.chat.completions.create(
            model=model_name,
            messages=prompt,
            temperature=0
        )

        prediction = response.choices[0].message.content.strip()
    except Exception as e:
        prediction = f'Sorry, I encountered the following error: \n{e}'

    # Log both the inputs and outputs to a local 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': prediction
                }
            ))
            f.write("\n")

    return prediction

# Set up the Gradio UI
# Add text box and radio button to the interface
# The radio button is used to select the company 10k report in which the context needs to be retrieved.

textbox = gr.Textbox(label="User Input")
#company = gr.List(label="Select Company", choices=["IBM", "Meta", "aws", "google","msft"])
company = gr.Dropdown(label="Select Company", choices=["IBM", "Meta", "aws", "google","msft"])

# Create the interface
# For the inputs parameter of Interface provide [textbox, company]
demo = gr.Interface(fn=predict, inputs=[textbox, company], outputs="text")

demo.queue()
demo.launch(share=True)