https://github.com/veldhub/veld_code__analyse_conllu
Code velds encapsulating creation of statistical summary on conllu data.
https://github.com/veldhub/veld_code__analyse_conllu
analysis conllu nlp
Last synced: 2 months ago
JSON representation
Code velds encapsulating creation of statistical summary on conllu data.
- Host: GitHub
- URL: https://github.com/veldhub/veld_code__analyse_conllu
- Owner: veldhub
- License: mit
- Created: 2024-12-07T17:36:28.000Z (6 months ago)
- Default Branch: main
- Last Pushed: 2025-01-20T19:26:30.000Z (4 months ago)
- Last Synced: 2025-01-30T10:18:25.165Z (4 months ago)
- Topics: analysis, conllu, nlp
- Language: Jupyter Notebook
- Homepage:
- Size: 19.5 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
#  veld_code__analyse_conllu
**\*work in progress\***
This repo contains [code velds](https://zenodo.org/records/13322913) encapsulating creation of
statistical summary on conllu data. For now, the observed features are:
- count of total tokens
- count of total lemma
- count of lemma normalized by token (to put the lemma in relation with token)
- count of occurrence of each (Universal Dependencies) part of speech tag
The code can be adapted to other use cases but is primarily used in this chain veld:
https://github.com/veldhub/veld_chain__eltec_udpipe_inference
## requirements
- git
- docker compose (note: older docker compose versions require running `docker-compose` instead of
`docker compose`)
## how to use
A code veld may be integrated into a chain veld, or used directly by adapting the configuration
within its yaml file and using the template folders provided in this repo. Open the respective veld
yaml file for more information.
**[./veld.yaml](./veld.yaml)** : Launches a jupyter notebook providing statistical analysis
```
docker compose -f veld.yaml up
```