PredicSis python SDK Documentation Status


A complete documentation is available there: documentation


Install the python SDK package:

pip install predicsis

or using the git repository:

git clone
python install

(optional) Setup your environment variables

You can used the SDK with an explicit configuration of url instance and token or you can provide them as environment variables: PREDICSIS_URL and PREDICSIS_TOKEN.

For example on unix-like system:

$ export PREDICSIS_URL=http://your_predicsis.ai_url/api
$ export PREDICSIS_TOKEN='eyJ0eXAiOiJKV1QiLCJhbGciOiJ...'

Warning URL have to end with /api.


The explicit configuration overrides the implicit one.


Token can be provided by the administrator using GUI (see documentation) or using the SDK (see predicsis.PredicSis._Admin.create_token():)

Test your installation

You can run tests with the following python lines:

  • Test if it is properly installed:
>>> import predicsis
>>> predicsis.__version__

The python SDK is properly connected to your

>>> from predicsis import PredicSis
>>> # if environment variables are setup
>>> prds = PredicSis()
>>> # otherwise
>>> # prds = PredicSis(predicsis_url="...", token="...")
>>> prds.admin.versions()
        'name': 'atlas',
        'version': '0.0.1'
    }, {
        'name': 'license_authority',
        'version': u'1.3.2'
    }, {
        'name': 'dmgr',
        'version': '0.0.1'
    }, {
        'name': 'modeler',
        'version': '0.0.1'
    }, {
        'name': 'exporter',
        'version': '0.0.1'
    }, {
        'name': 'prds_sdk',
        'version': '2.0.0'

It is well configured! Congratulation!

Let’s begin with PredicSis.


Please refer to Logger section if you have any unexpected behavior using the SDK.

Indices and tables