Canonizer is a REST service that tries to match Hugo nomination entry with a canonized form. It uses Decision Tree machine learning algorithm for matching, so it needs to be trained with a set of nominations and their correct canonized forms before usage.
Following software are needed to run the canonizer. Versions listed are what the canonizer has been tested with - it might work with other versions but no guarantees.
python 2.7.11
scikit-learn 0.17.1 numpy 1.11.0 scipy 0.17.0 bottle 0.12.10 requests 2.11.1 (Needed for running tests)
From command line:
~/canonizer/python main.py
The server should start at address http://127.0.0.1:8188/
From command line:
~/canonizer/Test/python rest_tester.py
- Parameters:
category,title,author,publisher,set,editor,example
Returns a potential canonization id for a single nomination. Any parameter can be empty or omitted.
{ "category": "Novel", "title": "3 Little Pigs", "author": "Ms Piggy", "publisher": "Example books" }
{ "canon_id": 23 }
400 Bad Request (Malformed JSON in request)
Trains the canonizer with training data.
200 OK
- Parameters:
entries(required),category(required),nominations(required),category,title,author,publisher,set,editor,example, id (required for each nomination)
Adds a set of nominations to training data
{
"entries": [
{
"category": 'Fancast',
"nominations": [
[ { title: '3 pigs' }, 1],
[ { title: 'Three Little Piggies' }, 2 ]
]
}
]
}
200 OK
400 Bad Request (Malformed JSON in request)
Clears training data set and resets the canonizer.
200 OK