ckanext-tagmanager
Tagmanager offers tools and interfaces for cleaning tags in CKAN open data portals.
What can you do with the tagmanager plugin?
The main features are the detection of similar tags, and the possibility to merge them. This is useful for portals with many contributors, where tags are not always kept consistent.
We offer three modes for detecting similar tags:
- Strong similarity: detect tags that differ only by capitals special characters
- Medium similarity: detect tags that have a Levenshtein edit distance smaller than three
- Synonyms: show tags that are synonyms (only for English)
Why a tagmanager plugin?
This extension is intended to fill the tag management gap of CKAN. In the future, we plan to offer the creation of relationships between tags, and a tag recommendation structure.
Requirements
Before installing tagmanager, make sure you have:
- CKAN 2.5+
- Levenshtein python library: pip install python-Levenshtein
- Unidecode python library: pip install unidecode
- NLTK library: pip install nltk
- NLTK data: python -m nltk.downloader all
Installation
- Activate your CKAN virtual environment
- Install the ckanext-tagmanager Python package:
pip install ckanext-tagmanager
- Run the database migration:
paster tagmanager migrate -c /etc/ckan/default/production.ini
- Add
tagmanager to the ckan.plugins setting
- Restart CKAN
Development Installation
git clone https://github.com/alantygel/ckanext-tagmanager.git
cd ckanext-tagmanager
python setup.py develop
paster tagmanager migrate -c /etc/ckan/default/development.ini
pip install -r dev-requirements.txt
Use
Navigate to yoursite/tagmanager
Acknowledgements
This work was driven in the context of the research STODaP project, developed at the Federal University of Rio de Janeiro (Brazil) and the University of Bonn (Germany).
License
AGPL-3.0