a Python library for wordnets

Available Wordnets | Documentation | FAQ | Migrating from NLTK | Roadmap

Wn is a Python library for exploring information in wordnets. Install it from PyPI and download some data:

$ pip install wn
$ python -m wn download oewn:2021  # the Open English WordNet 2021

Then start exploring:

>>> import wn
>>> en = wn.Wordnet('oewn:2021')  # Create Wordnet object to query
>>> ss = en.synsets('win')[0]     # Get the first synset for 'win'
>>> ss.definition()               # Get the synset's definition
'be the winner in a contest or competition; be victorious'

• Multilingual by design; first-class support for wordnets in any language
• Interlingual queries via the Collaborative Interlingual Index
• Six similarity metrics
• Functions for exploring taxonomies
• Support for lemmatization (Morphy for English is built-in) and unicode normalization
• Full support of the WN-LMF 1.1 format, including word pronunciations and lexicon extensions
• SQL-based backend offers very fast startup and improved performance on many kinds of queries

Any WN-LMF-formatted wordnet can be added to Wn's database from a local file or remote URL, but Wn also maintains an index (see wn/index.toml) of available projects, similar to a package manager for software, to aid in the discovery and downloading of new wordnets. The projects in this index are listed below.

There are several English wordnets available. In general it is recommended to use the latest Open English Wordnet, but if you have stricter compatibility needs for, e.g., experiment replicability, you may try the OMW English Wordnet based on WordNet 3.0 (compatible with the Princeton WordNet 3.0 and with the NLTK), or OpenWordnet-EN (for use with the Portuguese wordnet OpenWordnet-PT).

These are standalone non-English wordnets and collections. The wordnets of each collection are listed further down.

The Open Multilingual Wordnet collection (omw:1.4) installs the following lexicons (from here) which can also be downloaded and installed independently:

The Open Wordnets for Portuguese and English collection (own:1.0.0) installs the following lexicons (from here) which can also be downloaded and installed independently:

While not a wordnet, the Collaborative Interlingual Index (CILI) represents the interlingual backbone of many wordnets. Wn, including interlingual queries, will function without CILI loaded, but adding it to the database makes available the full list of concepts, their status (active, deprecated, etc.), and their definitions.

The 2021 version of the Open English WordNet (oewn:2021) has changed its lexicon ID from ewn to oewn, so the index is updated accordingly. The previous versions are still available as ewn:2019 and ewn:2020.

The wordnet formerly called the Princeton WordNet (pwn:3.0, pwn:3.1) is now called the OMW English Wordnet based on WordNet 3.0 (omw-en) and the OMW English Wordnet based on WordNet 3.1 (omw-en31). This is more accurate, as it is a OMW-produced derivative of the original WordNet data, and it also avoids license or trademark issues.

All OMW wordnets have changed their ID scheme from ...wn to omw-.. and the version no longer includes omw (e.g., bulwn:1.3 omw is now omw-bg:1.4).