Tokenizer Viz

Tokenizer Viz is a Python package that generates HTML to visualize the tokenization of text. It highlights tokens with different colors and customizable styles, making it easier to understand how a text is tokenized.

Project Layout

tokenizer-viz/
│
├── tokenizer_viz/
│   ├── __init__.py
│   └── visualization.py
│
├── .gitignore
├── LICENSE
├── README.md
└── setup.py

<br>

## Installation

You can install the **`tokenizer-viz`** package using pip:

```bash
pip install tokenizer-viz

Usage

Here's a quick example of how to use the package:

Usage with a provided encoder and decoder

from tokenizer_viz import TokenVisualization
from IPython.display import HTML

# Define sample encoder and decoder functions for demonstration purposes
def sample_encoder(text):
    return list(text)

def sample_decoder(token):
    return token

# Initialize the TokenVisualization class with the encoder and decoder functions
token_viz = TokenVisualization(
    encoder=sample_encoder,
    decoder=sample_decoder
)

# Define a sample text to visualize tokenization boundaries
sample_text = "This is a sample text.\nIt has multiple lines."

# Visualize the tokenization boundaries
html = token_viz.visualize(sample_text)
HTML(html)

EXAMPLE OUTPUT

TBD EXAMPLE IMAGE

ARGUMENTS

The TokenVisualization class accepts several optional parameters to customize the appearance and layout of the tokens:

• cmap (defualt='Pastel2'),
• font_family (defualt='Courier New'),
• transparency (default=0.675),
• font_size (defualt='1.1em'),
• unk_token (defualt='???'),
• font_weight (defualt=300),
• padding (defualt='0px'),
• margin_right (defualt='0px'),
• border_radius (defualt='0px'),
• background_color (defualt='#F0F0F0'),

Please refer to the class docstrings and method docstrings for a detailed description of each parameter.

License

This project is licensed under the MIT License.