allenai / allenai/science-parse
Extract tables in structured format
Aperta
- Lingua principale
- Java
- Stelle
- 706
- Fork
- 88
- Metriche di merge delle PR
- Nessuna PR unita negli ultimi 30g
Descrizione
Hello!
As far as I understand, science-parse extracts metadata and raw text from articles. However I would like to get also structured information from tables, for example json-like format
```
[['dataset', 'model', 'metrics'], ['MNIST', 'CNN', '0.98'], ...]
```
I understand that sometimes there are difficult-to-extract tables, e. g. when some rows/column are shared. However, I believe that in simplest case, this can be done quite easily.
Any ideas?
Guida per i contributori
Nessuna guida per i contributori indicizzata per questo repository
Valutazione
Questa issue non è ancora stata valutata.