---
title: Datahub Qa Bot
emoji: 👁
colorFrom: gray
colorTo: purple
sdk: streamlit
sdk_version: 1.17.0
app_file: app.py
pinned: false
license: mit
---

# DataHub documentation bot
Using [OpenAI](https://platform.openai.com/docs/introduction), [Langchain](https://python.langchain.com/en/latest/index.html) and [streamlit](https://docs.streamlit.io/) to train DataHub documentation to provide a DataHub QA BOT on [huggingface space](https://huggingface.co/spaces/abdvl/datahub_qa_bot?logs=build)

# How to run locally
1. Clone the repo
2. Run:
```
source .venv/bin/activate
pip install -r requirements.txt
streamlit run app.py
```

## How to train your own model
1. Delete the db folder
2. Copy the docs folder from [DataHub docs folder](https://github.com/datahub-project/datahub/tree/master/docs) to `./docs`
3. Update the `os.environ["OPENAI_API_KEY"] ` in the `train.py`
4. Run `python3 train.py`

The training will take 15 seconds, and cost around $0.20
```
chromadb.db.duckdb: loaded in 236 embeddings
chromadb.db.duckdb: loaded in 1 collections
```