Crate langdetect 

A port of langdetect (itself a port of Nakatani Shuyo’s language-detection) that reproduces its results bit for bit, including the random walk: with the same seed, the probabilities equal those of langdetect 1.0.9 on the CPython version selected by Compat.

use langdetect::{DetectorFactory, Seed};

assert_eq!(langdetect::detect("Ein, zwei, drei, vier").unwrap(), "de");

let mut factory = DetectorFactory::builtin().clone();
factory.seed = Seed::Int(0);
let langs = factory.detect_langs("Ceci est une phrase en français.").unwrap();
assert_eq!(langs[0].lang, "fr");

Modules

text

Text preprocessing of Detector.append, Detector.cleaning_text and NGram.

Structs

Compat

The Python behaviour to reproduce. Results are bit-identical to langdetect running on a CPython whose version matches these settings.

DetectOptions

Options of one detection, the attributes of langdetect.detector.Detector.

Detector

Detects the language of the appended text, the equivalent of langdetect.detector.Detector.

DetectorFactory

Language profiles and n-gram probabilities, the equivalent of DetectorFactory.

LangProfile

A language profile: n-gram counts and the total number of 1-, 2- and 3-grams.

Language

A detected language and its probability.

Enums

Error

An error, mirroring the exception langdetect raises in the same situation.

ErrorCode

Error codes of langdetect.lang_detect_exception.ErrorCode.

Seed

Seed of the random walk, mirroring DetectorFactory.seed.

SumMode

How sum() adds the probabilities when they are normalised.

UnicodeVersion

Unicode version that decides which characters str.isupper() reports as uppercase.

Functions

detect

The most probable language of text with the built-in profiles and a random seed.

detect_langs

Languages of text with probability above 0.1, most probable first.