IssueSnap

This project provides a user-friendly interface to generate bug reports using a large language model (LLM). Users can describe the steps to reproduce an issue and the expected results, and the LLM will craft a detailed bug report incorporating these details.

Requirements

This project requires the following Python3 libraries:

• streamlit
• pinecone-client
• langchain-google-genai
• python-dotenv
• pandas

Project Structure

• IssueSnap.py: Main script to start and run the web tool
• generate_bug_report.py: Functional script to define the core logic for processing user input, querying Pinecone for similar issues, and using the LLM to create the bug report.
• closest_sample_finder.py: Functional script to vectorize user input for searching similar issues within Pinecone.
• upsert_csv_to_pinecone.py: Standalone script that vectorizes source csv data for upserting to Pinecone database.

Functionality Overview

1. User Input: Users describe the steps to reproduce a bug and the expected results through a Streamlit interface.
2. Vectorization and Search: The user input is vectorized and used to query Pinecone for similar bug reports (if available).
3. LLM Prompt Generation: Based on user input and potentially retrieved similar issues, a prompt is crafted to guide the LLM in generating the bug report.
4. Bug Report Generation: The LLM utilizes the provided prompt to generate a comprehensive bug report describing the issue and potential solutions.

Running the Project

Set Up Environment

1. Navigate to the project directory:

   cd your-work-directory/issuesnap
   

2. Create and activate a virtual environment:

   python3 -m venv env
   source env/bin/activate
   

3. Install required packages:

   pip install -r requirements.txt
   

4. Set up your Google and Pinecone accounts and obtain the necessary API keys.

5. Create a .env file in the project directory and store your API keys as the example in .env.example.

6. Ensure you have a Pinecone index created and populated with relevant bug report data (title, description, steps to reproduce, expected results) with appropriate metadata fields.

Start the Web Tool on Local

1. Launch without hiding debug options:

   streamlit run src/IssueSnap.py
   

2. Launch with production mode options:

   streamlit run src/IssueSnap.py --client.toolbarMode=minimal
   

Deployment

For deployment details, please see the README in folder deployment_templates.