Public | Automated Build

Last pushed: 2 years ago
Short Description
Experimental recommendation engine API using Flask
Full Description


Content recommendation API
Using Python3

Running the app with Docker

  1. Install Docker, see for the instructions to install it on a Mac.
  2. Then, in the top directory of this repository, you should be able to run the app using docker-compose. Just type docker-compose up --build. This will download the needed packages and run the app in the foreground.
  3. The first time you run the container, it will download and crunch the simple_wiki example. This will take about 10 minutes. You will see Running on (Press CTRL+C to quit) when it is done.
  4. Browse to localhost:5000, you should see 'Content recommendation rocks!' as message.
  5. Try out the test API by using curl: curl localhost:5000/api/recommend/v1.0/<corpus_name> -X GET -d 'url=', you should see a json document returned from the engine. See the list of available corpora below, you will need to add it to the API call.

Available corpora

The available corpora are:

  1. wiki-simple: this currently returns the document ids and scores of the 5 most relevant articles from the simple wikipedia corpus.

Adding a recommendation engine

If you write a new recommendation engine, it needs to inherit from Recommender base class, see /app/recommender/ The API currently picks a recommendation engine randomly. To register your engine, add the module to the /app/recommender/ directory. Then, explicitly import the class in /app/ This way, the class will be added to the list of subclasses of the Recommender base class.

Docker Pull Command
Source Repository