MiniSBD: Fast Sentence Boundary Detection

0.9.5 · active · verified Fri Apr 17

MiniSBD is a free and open-source Python library designed for fast and efficient sentence boundary detection (SBD). It provides a lightweight solution for splitting text into sentences, supporting various punctuation and language patterns. The current version is 0.9.5, with releases occurring periodically, often driven by improvements in tokenization or punctuation handling.

Common errors

```
ModuleNotFoundError: No module named 'minisbd'
```
cause The minisbd package is not installed in your current Python environment.

fix Run `pip install minisbd` to install the library.
```
AttributeError: 'SBD' object has no attribute 'segments'
```
cause You are attempting to call a method named `segments`, but the correct method for sentence segmentation is `segment` (singular).

fix Change your code from `sbd.segments(text)` to `sbd.segment(text)`.

Warnings

gotcha Initializing the `SBD` object is an operation that should ideally be done once. Re-initializing it inside a loop will incur unnecessary overhead and degrade performance.
Fix: Always initialize `sbd = SBD()` outside of performance-critical loops or functions that will be called repeatedly. Pass the initialized object around if necessary.
gotcha MiniSBD is optimized for speed and general English text. While robust, its accuracy might vary for highly specialized domains, informal text (e.g., social media), or languages with very different sentence boundary rules than common Western languages. It may not handle all edge cases perfectly.
Fix: For highly critical applications or non-standard text, evaluate its performance on a representative dataset. Consider pre-processing text or post-processing results if specific patterns cause issues. For deeply linguistic tasks, more complex NLP libraries might be required.

Install

pip install minisbd Install stable release

Imports

SBD
wrong:
```
import minisbd; sbd = minisbd.SBD()
```
correct:
```
from minisbd import SBD
```
While functional, direct import of SBD is idiomatic and often preferred for clarity.

Quickstart

Initialize the SBD class once, then use its `segment` method to split a string into a list of sentences.

from minisbd import SBD

sbd = SBD() # Initialize the SBD object once

text1 = "Hello world. This is a test. Is it working?"
sentences1 = sbd.segment(text1)
print(f"Text 1: {sentences1}")

text2 = "Hello world! This is another test. Is it working now?"
sentences2 = sbd.segment(text2)
print(f"Text 2: {sentences2}")

view raw JSON →