BASA: Modern NLP toolkit for Indonesian and regional languages.
BASA helps developers clean and normalize informal Indonesian text with a simple, developer-friendly API.
from basa import quick
quick("GW GKKKK NGERTIII BNGTTTT!!!!!")
# saya tidak mengerti banget!
Features
- Indonesian slang normalization
- Repeated-character reduction (
gkkkk → gk) - Conservative typo correction (opt-in)
- Batch processing support
- Pure Python implementation
- Developer-friendly API
Installation
pip install basa
Quick Example
from basa import normalize
normalize(
"gw gk ngerti bngtttt sihhhh!!!"
)
# saya tidak mengerti banget sih!
Design Philosophy
BASA follows several core principles:
- Conservative by default — destructive operations are opt-in.
- One-line API — most users should only need
quick(). - Indonesian-first — built specifically for Indonesian NLP workflows.
- Future regional support — designed to grow toward Javanese, Sundanese, and other local languages.
Roadmap
v0.1
normalize()quick()- Slang normalization
- Typo correction
- Batch processing
Future Releases
- Javanese ↔ Indonesian translation
- Sundanese ↔ Indonesian translation
- Indonesian factual consistency evaluation
- Dataset utilities for IndoBERT and IndoBART
- Synthetic data generation for low-resource languages
License
Released under the MIT License.