langidpy

Langidpy is a Python library for automatic language identification. It implements a character n-gram based classifier to determine the language of a given text, typically returning a language code along with a confidence score. The library ships with a pre-trained model that covers a broad set of languages and performs well on short to moderately long text snippets. Output languages are reported using ISO language codes, commonly ISO 639-1.

Langidpy provides both a Python API and a command-line interface. In Python, the typical usage involves calling a classify function that returns a tuple such as (language_code, confidence). The command-line tool can process files or streams, enabling batch language detection without writing code. The implementation emphasizes speed and a small memory footprint, making it suitable for real-time processing in NLP pipelines.

A notable strength of langidpy is its ability to be retrained with user-provided data. This allows domain adaptation or the addition of custom languages, which can improve accuracy in specialized contexts. The library is designed to be easily integrated into preprocessing stages for text analytics, web crawling, or data cleaning workflows.

Limitations include reduced accuracy on very short texts or when texts contain multiple languages or code-switching. In such cases, results should be interpreted probabilistically rather than as definitive labels. Langidpy is open-source and commonly installed via pip, reflecting its role as a lightweight, out-of-the-box tool for language detection in Python projects.