efre-lod-api

Flask API to access your Linked-Open-Data contained in an elasticsearch-cluster.

Requirements

python3 and other packages specified in requirements.txt

Install

• clone repository

  git clone https://github.com/efre-lod/efre-lod-api.git
  

• create and activate a virtual environment for this project

  cd efre-lod-api
  python3 -m venv env
  source ./env/bin/activate
  

• install package and its requirements into venv

  pip3 install -e .
  

Systemd unit-file (optional)

The systemd template file can be used to start the lod-api via systemd even automatically at boot time.

• copy systemd template file to /etc/systemd/system.

  cp systemd/[email protected] /etc/systemd/system/[email protected]
  

  hereby, replace the username with the user which should finally run the LOD-API and has it locally installed to /home/username/.local/bin

via Makefile (optional)

There is the posibility to use make to install the software into your home-directory (not using a virtualenv) and setup the systemd unit file. You must run the make as root in the project's base directory.

Usage

Copy and configure apiconfig.yml.example to suit to your Elasticsearch-Infrastructure containing your JSON-LD. Possible places for storing the config are:

• in /etc as /etc/lod-apiconfig.yml
• specify the config file directly via -c, e.g.

  lod-api -d --config apiconfig.yml
  

For starting the api in debug mode, do:

lod-api -d

For controlling the daemon, use:

lod-api [--config apiconfig.yml] {start|stop|restart}

For a productive environment, we recommend to put the API behind a load-balancer (like nginx).

systemd

Enable the lod-api-systemd-service to start it at boot using the username of the local user which installed the lod-api software.

systemctl enable lod-api@username

And start/stop it via systemctl:

service lod-api@username {start|stop|restart|status}

Example

One example configuration file can be found in tests/docker/lod-api/docker_apiconfig.yml. This configuration fits to the data contained in the test set tests/data/LDTestSet.tar.bz2, which can be stored in an elasticsearch and be used as test set for the lod-api. See data mocking integration.

Tests

There is a subset of tests that can be triggered offline (i.e. without depending on elasticsearch), therefore the offline marker is used and can be given to pytest:

python3 -m pytest -m offline tests

For triggering all tests, the api must be started separately. Tests depend on the configuration file of the api. Especially the debug-host and -port are important to determine where to run the tests against. The default configuration file is assumed to be apiconfig.yml.example in the home directory of the application. However, another configuration file can be provided using the --config switch. You can run the tests via

python3 -m pytest tests [--config apiconfig.yml]

Be aware though that some of the tests depend on mock data which are provided in tests/data and have to be stored in a local elasticsearch instance. When you want to test the API with your own data you can, of course, generate your own test set of mock data

If you want just a single test to be triggered, e.g. the connection test, you can do this with

python3 -m pytest tests/test_apis_http_status

If the output should be a bit more verbose, you can turn on print() statements via

python3 -m pytest -s tests/

For more information on the tests have a look into tests.

Data mocking integration

We provide an elasticsearch Dockerfile (in tests/docker/elasticsearch) from which a container running an elasticsearch instance can be build. There are also test set with linked data (in tests/data/LDTestSet.tar.bz2 which can be extraced and loaded into the locally running elasticsearch. See tests for more information on howto prepare the elasticsearch docker image.

In the case you are using the mock data test set you will most likely use its provided configuration file as well for the api:

lod-api -d -c tests/docker/lod-api/docker-apiconfig.yml