BurmeseNLP¶

BurmeseNLP (burmesenlp) is an open-source Python library for rule-based Myanmar (Burmese) natural language processing.

Version 1.0 focuses on preprocessing: normalization, Zawgyi ↔ Unicode, syllable / word / sentence segmentation, multi-word expressions (BMWE), lexicon management, rule-based POS tagging, and phrase chunking.

pip install burmesenlp

from burmesenlp import process

doc = process("ကျွန်တော်ကျောင်းသို့သွားသည်။")
print(doc.words)
# ['ကျွန်တော်', 'ကျောင်း', 'သို့', 'သွား', 'သည်', '။']

Pipeline¶

flowchart TD
  raw[Raw text] --> norm[Normalize]
  norm --> syl[Syllables]
  syl --> words[Word tokenize]
  words --> mwe[BMWE]
  mwe --> pos[POS tag]
  pos --> chunk[Phrase chunk]
  chunk --> sent[Sentence segment]
  sent --> doc[Document]

What you get¶

Capability	Module
Full pipeline	`process` / `BurmeseNLP`
Syllable / word / sentence	`tokenize`
Multi-word idioms	`mwe`
POS tags	`tag`
Phrase chunks	`chunking`
Lexicon	`lexicon`
Zawgyi ↔ Unicode	`zawgyi`

Next steps¶

License¶

Apache-2.0 — see the repository LICENSE file.