What is Datature's Python SDK?

Datature constantly strives to make MLOps accessible for all users, from individual developers to large enterprises with established codebases. With that in mind, one of the tools we are making accessible is our Python SDK, which allows you to easily perform general interactions at all the essential steps of the MLOps pipeline through Python code.

The main function categories are as follows:

Function Category	Description
Workspace Management	This gives you the essential functions to getting basic information about your workspace.
Project Management	This gives you the essential functions to getting and changing basic information about your project.
Asset Management	This deals with the upload, retrieval, download, and removal of assets on the Nexus platform.
Asset Tag Management	This deals with the retrieval and manipulation of asset tags on the Nexus platform.
Annotation Management	This deals with the upload and retrieval of annotations stored on the Nexus platform.
Workflow Management	This controls the retrieval of created workflows on the platform.
Model Training Management	This allows users to start and monitor trainings without going on the platform.
Artifact Management	This allows users to retrieve and export artifacts in their preferred framework type.
API Deployment Management	This allows users to start, monitor, and stop Inference API deployments.
Inference Batch Job Management	This allows users to manage resources related to batch job inference.
Operation Management	This allows users to get insight on processes called in the SDK that are ongoing.

How to Get Started

Install Python 3.8 or Above

As this is an SDK for Python, users will need to ensure that they have Python installed. As of now, we currently fully support all versions of Python from 3.8 or above. If you are having issues with the SDK, please ensure that your environment uses an Python version that is fully supported, or else we are not able to guarantee functionality or fixes.

Install Datature's Python SDK

To make installation as simple as possible, we have made the Python package available on PIP, Python's most popular package installation tool. After ensuring you have pip installed in your environment, which should come with a standard Python installation, you can simply enter the following command below.

pip install --upgrade datature

Authentication

The final step that is essential to all successful requests is to ensure that you log on to the platform, access the relevant project, and store the project secret key which can be found in Integrations. As mentioned in the following link, only the project owner or those with relevant permissions can have access to the project secret on the platform. For more detail on the project key and secret key, check out this link for more information.

Once you have the project secret, you will now be able to make API requests using the Python SDK by setting this:

from datature import nexus

client = nexus.Client("965d5031........c")

You will now be able to run your desired commands as outlined above. To see all possible functions as well as view the required inputs and expected outputs, check out the above documentation.