souley/lofar_workflow_api

Continuation of https://github.com/EOSC-LOFAR/lofar_workflow_api for PROCESS

Readme for LOFAR pipeline REST API

This project has been continued as part of PROCESS.

This is an experiment to see if we can make the LOFAR data more accessible. This is a django REST api using the django rest_framework.

Get it working

• Download this repo
• Create a venv in which you want to do your things (using python3) or use conda
• Install django (e.g. pip install django)
• Install the rest_framework (e.g. pip install djangorestframework)
• Install all the pipelines (as of 27JUN18 this is: https://github.com/EOSC-LOFAR/LGPPP_LOFAR_pipeline)
• Navigate to the folder containing the manage.py
• Start your local host: python manage.py runserver
• You can now send http requests to the localhost:8000. For example in your browser: localhost:8000/sessions/

Install

Requirements:

• git
• pipenv, https://docs.pipenv.org/

Run the following commands to install

git clone https://github.com/EOSC-LOFAR/lofar_workflow_api.git
cd lofar_workflow_api
pipenv install
pipenv shell
cd lofar_workflow_api/

Run

Start web service with following command:

python manage.py runserver

Goto http://localhost:8000/sessions

Follow an example

Check out the jupyter notebook: example_for_lofar_pilot_REST_api.ipynb.

Current models in the api

• Session: a session that will run a pipeline on an observation
• pipelineschemas: this gives you a list with implemented pipelines and configuration schemas.
• A post of a session using the request package could look like this:

	data = {
			"email": "[email protected]",
			"description": "Add your description to figure out later what this is.",
			"pipeline" : "LGPPP_LOFAR_pipeline",
			"config": "{\"avg_freq_step\": 1, \"avg_time_step\": 1, \"do_demix\": 1, \"demix_freq_step\": 1, \"demix_time_step\": 1, \"demix_sources\": 1, \"select_NL\": 1,\"parset\": 1}",
			"observation": "an observation code",
			}
	response = s.post('http://localhost:8000/sessions/', data=data)

• You can do a get to pipelineschemas to get a json with pipelines and their schemas like this:

	response = s.get('http://localhost:8000/pipelineschemas/')
	response_data = response.json()
	pp.pprint(response_data)

Adding your own pipeline

Please follow the intructions at the pipeline template python package: https://github.com/EOSC-LOFAR/LOFAR_pipeline_template