https://github.com/cyclecycle/visualise-spacy-tree

Create dependency tree plots from SpaCy Doc objects
https://github.com/cyclecycle/visualise-spacy-tree

nlp python spacy

Last synced: 2 months ago
JSON representation

Create dependency tree plots from SpaCy Doc objects

• Host: GitHub
• URL: https://github.com/cyclecycle/visualise-spacy-tree
• Owner: cyclecycle
• License: mit
• Created: 2019-05-23T11:25:09.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2022-12-08T05:09:45.000Z (about 2 years ago)
• Last Synced: 2024-10-01T14:50:04.619Z (3 months ago)
• Topics: nlp, python, spacy
• Language: Python
• Size: 115 KB
• Stars: 7
• Watchers: 1
• Forks: 3
• Open Issues: 4
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# visualise-spacy-tree

An alternative to [SpaCy's](https://spacy.io) [visualizer](https://spacy.io/usage/visualizers#dep), built on [GraphViz](https://graphviz.gitlab.io/). 

![Example plot image](https://github.com/cyclecycle/visualise-spacy-tree/blob/master/visualise_spacy_tree/example_plots/custom_plot.png)

## Prerequisites

- You need [GraphViz](https://graphviz.gitlab.io/download/) installed.

## Installation

With pip:

```bash

pip install visualise-spacy-tree

```

Or from source:

```bash

git clone https://github.com/cyclecycle/visualise-spacy-tree.git visualise_spacy_tree

cd visualise_spacy_tree

python setup.py install

```

## Usage

```python

# Parse a string to create SpaCy Doc object

import en_core_web_sm

text = 'Forging involves the shaping of metal using localized compressive forces.'

nlp = en_core_web_sm.load()

doc = nlp(text)

# Create the plot

import visualise_spacy_tree

png = visualise_spacy_tree.create_png(doc)

# Write it to a file

with open('parse_tree.png', 'wb') as f:

    f.write(png)

# If you're using Jupyter notebook, you can render it inline

from IPython.display import Image, display

display(Image(png))

# Override node attributes to customise the plot

from spacy.tokens import Token

Token.set_extension('plot', default={})  # Create a token underscore extension

for token in doc:

    node_label = '{0} [{1}])'.format(token.orth_, token.i)

    token._.plot['label'] = node_label

    if token.dep_ == 'ROOT':

        token._.plot['color'] = 'green'

'''

You can set any valid GraphViz dot attribute in 'plot'.

See GraphViz docs for reference of possible node attributes:

https://graphviz.gitlab.io/_pages/doc/info/attrs.html

'''

```

## Running the tests

Run 

```bash

pytest

```

from the root directory.

## Acknowledgements

Uses:

- [pydot](https://github.com/pydot/pydot)

## Contributions

Are welcome :)