Installation

Requirements

• Python 3.9 or higher

• click 8.0 or higher

• rich 13.0 or higher

• spaCy 3.5 or higher (optional, for high-accuracy mode)

Installation Options

phrasplit can be installed with or without spaCy support, depending on your needs.

pip install phrasplit

pip install phrasplit[nlp]

pip install phrasplit
pip install spacy>=3.5.0

Installing from Source

To install from source, clone the repository and install:

git clone https://github.com/buchwandler/phrasplit.git
cd phrasplit
pip install -e .

For development, install with dev dependencies:

pip install -e ".[dev]"

For spaCy support when installing from source:

pip install -e ".[nlp]"

Installing spaCy Language Models

If you installed with spaCy support, you’ll need a language model. The default model is en_core_web_sm (English). Install it with:

python -m spacy download en_core_web_sm

For better accuracy, you can use larger models:

# Medium model (more accurate)
python -m spacy download en_core_web_md

# Large model (most accurate)
python -m spacy download en_core_web_lg

For other languages, see the spaCy models documentation.

Verifying Installation

You can verify your installation by running:

import phrasplit
print(phrasplit.__version__)

from phrasplit import split_sentences

# Works with or without spaCy
print(split_sentences("Hello world. How are you?"))
# ['Hello world.', 'How are you?']

# Explicitly use simple mode (no spaCy required)
print(split_sentences("Hello world. How are you?", use_spacy=False))
# ['Hello world.', 'How are you?']

Choosing Between Modes

Simple Mode (use_spacy=False)

• No spaCy installation required

• Faster processing

• Lower memory usage

• Good for straightforward text

• Uses regex-based splitting

spaCy Mode (use_spacy=True, default if available)

• Requires spaCy and language models

• Higher accuracy

• Better handling of complex cases

• Uses NLP-based analysis

• Recommended for production use