Expand description
§crftag
CRF-based POS tagger and shallow parser (chunker) built on crfrs.
Fully independent — drop it into any Rust project.
§Features
POSTagger— structured-perceptron taggerChunker— IOB chunker (shallow parser)RuleBasedChunker— grammar-based chunker, no model file requiredtree2brackets— formats a parse result as a bracketed stringfeatures— raw feature extraction (for custom models)crfrs— re-exported; use for training/loading models directly
§Quick start
use crftag::POSTagger;
let mut tagger = POSTagger::new();
tagger.load_model("pos_tagger.model").unwrap();
let tagged = tagger.tag(&["من", "به", "مدرسه", "رفتم", "."]).unwrap();§Training
use crftag::{POSTagger, crfrs::TrainConfig};
let corpus: Vec<Vec<(String, String)>> = vec![]; // (word, tag) pairs
POSTagger::train_and_save(&corpus, "pos_tagger.model", TrainConfig::default()).unwrap();Re-exports§
pub use chunker::iob_to_tree;pub use chunker::tree2brackets;pub use chunker::Chunker;pub use chunker::ChunkedSentence;pub use chunker::ChunkedToken;pub use chunker::ParseNode;pub use chunker::RuleBasedChunker;pub use error::Error;pub use error::Result;pub use tagger::POSTagger;pub use crfrs;