Tahrirgoh is a web platform developed by Tahrirchi for dataset collection for the Grammatical Error Correction (GEC) task.
Tahrirgoh is pronounced as T-AH-REER-GOKH
, and is made up word in Uzbek, meaning a "place to proofread".
For running the app locally, follow instructions in backend/README.md and frontend/README.md files
Use the docker-compose file to run it in a production server. Be sure to tweak the frontend/nginx.conf file to your needs.
docker compose up -d
After you have successfully launched the platform:
- Create worker users (API)
- Add sentences to the platform (API)
- Workers annotate sentences
- Repeat steps 2 and 3 until you have enough sentences annotated
- Download annotated sentences (API)
There is no admin panel implemented in the frontend, but there is a Swagger Docs initialized and available in the backend with user-friendly interface. Read more about it in API Documentation section on how to access it and use.
If you use this platform during your research, please cite us.
@software{Mamasaidov_Tahrirgoh_is_a_2023,
author = {Mamasaidov, Mukhammadsaid and Yusupov, Jasur},
month = oct,
title = {{Tahrirgoh is a web platform for dataset collection for the Grammatical Error Correction (GEC) task.}},
url = {https://github.com/tahrirchi/tahrirgoh},
version = {1.0.0},
year = {2023}
}