# KankenOnline
This project intends to provide a website that can generate practice questions for Kanji Kentei level 1.
I draw from both definition data and numerous texts from Aozora Bunko to create each style of question from 1 to 9.

## Running
- `wget https://huggingface.co/datasets/globis-university/aozorabunko-clean/resolve/main/aozorabunko-dedupe-clean.jsonl.gz` to get the Aozora data
- `gunzip aozorabunko-dedupe-clean.jsonl` to extract the data to a single file

## Sources
- [Aozora Bunko cleaned corpus on GitHub](https://github.com/globis-org/aozorabunko-extractor?tab=readme-ov-file)
    - [Hugging Face download](https://huggingface.co/datasets/globis-university/aozorabunko-clean)

# This early build based on Flask tutorial