Customer Feedback Analyzer with LLM Integration

Overview

This project implements a data processing pipeline that analyzes customer feedback using a Language Model (LLM). It evaluates the results and presents the findings on a simple web page.

Table of Contents

• Objective
• Requirements
• Installation (for running locally)
• Usage (locally)
• Usage (containerized)

Objective

Build a data processing pipeline that analyzes customer feedback using a Language Model (LLM), evaluates the results, and presents the findings on a simple web page. The project will involve developing a Python script for data processing, crafting effective prompts for LLM integration, implementing a sentiment analysis evaluation, and creating a user-friendly HTML presentation of the results. By leveraging Temporal, we can easily handle retries, state management, and orchestration of tasks, which enhances the overall efficiency and resilience of the application.

Requirements

• Python 3.x
• Docker (optional)
• Docker compose (optional)

Installation (for running locally)

1. Clone the repository:

   git clone https://github.com/Danieloni1/langgraph-feedback-ingestor.git
   cd langgraph-feedback-ingestor/src

2. Install the required libraries using pip:

   pip install -r requirements.txt

Usage (locally)

1. Start the Temporal server:

   temporal server start-dev

2. Run the worker:

   python worker.py

3. Run the main application:

   python app.py

4. Prepare your CSV file with the following columns:

   • feedback_id (integer)
   • customer_name (string)
   • feedback_text (string)
   • submission_date (date)

5. After processing, the results will be disaplyed, a graph image will be saved as graph.png and evaluation will be saved to evaluation/evaluation.txt.

Usage (containerized)

You can also run the application containerized:

1. Build and run the application, simply use docker compose:

   docker compose up --build

   Then visit localhost:5001 for the app and localhost:8080 for the Temporal dashboard.