Skip to content

Latest commit

 

History

History
71 lines (60 loc) · 2.1 KB

README.md

File metadata and controls

71 lines (60 loc) · 2.1 KB
Raw

Unstructured Data Conversion

Introduction

This is a small app that uses the EvaDB ChatGPT utility to convert unstructured data to a provided stuctured format. For the timebeing we focus on user hardware and software complaints. The data/user_complaints.csv file contains some example queries.

We run them through ChatGPT via a crafted prompt which classifies the Issue category and Issue Component along with displaying the raw issue string. The default format is the following

DEFAULT_STRUCTURE_FORMAT = [
    [
        "Issue Category",
        "What category the issue belongs to",
        "One of (hardware, software)",
    ],
    [
        "Raw Issue String",
        "Raw String containing the exact input given by the user",
        "string",
    ],
    ["Issue Component", "Component that is causing the issue", "string"],
]

Results

Query : The keyboard on my laptop is typing the wrong letters and it's driving me crazy!
✅ Answer:
{
  "Issue Category": "hardware",
  "Raw Issue String": "The keyboard on my laptop is typing the wrong letters and it's driving me crazy!",
  "Issue Component": "keyboard"
}

✅ Answer:
Query : My headphones won't connect to my phone anymore, even though they used to work just fine.
{
  "Issue Category": "hardware",
  "Raw Issue String": "My headphones won't connect to my phone anymore, even though they used to work just fine.",
  "Issue Component": "headphones"
}

Setup and running the application

Clone the project

git clone https://github.com/hershd23/eva-structure-gpt.git
cd eva-structure-gpt

Create a new python environment in conda or venv, and activate that environement (Optional but recommended)

python3 -m venv /env
source env/bin/activate

Install the dependencies

pip install -r requirements.txt

Simply run the main.py file and input the query, openai key and any additional inputs you want to add to the prompt

python3 main.py

Future Work

  • Currently the prompt and and the structure are kind of static. Allow user to add their own structure and paradigm of data.
  • More experimentation required with the prompt