discovery-sandbox 0.3.0

No project description provided

Pureinsights Discovery Platform: Sandbox SDK

Discovery Sandbox SDK is a Python package that allows developers to programatically access Discovery features. Currently, it supports executing one or multiple QueryFlow processors.

[!IMPORTANT] The current version of the Discovery Sandbox supports connections to internet-accessible services only. Support for local deployments is planned for an upcoming release.

Requirements

• Python 3.13+
• pip

Installation

pip install discovery-sandbox

Testing

Testing is done using the pytest framework. These commands are run from the root folder.

To run the full test suite:

pytest --verbose

To run tests from a single file or directory:

pytest <path_to_test> --verbose

To generate a XML coverage report:

pytest --verbose --cov=. --cov-report xml:coverage.xml 

To generate a HTML coverage report:

pytest --cov=. --cov-report html

Implementation

The SDK currently provides classes that represent Discovery entities (Server, Credential, Processor) as well as clients to interact with distinct endpoints. The API requests are made using the httpx library, and serialization is done using the json built-in module. This client does the following:

• Provides methods to execute standalone QueryFlow processors using text_to_text/text_to_stream that return the JSON execution output as a dictionary or text stream respectively.
• Supports overloading with the multimethod library to allow the usage of UUIDs instead of full entity objects.
• Supports the execution of a sequence of processors using the QueryFlowSequenceProcessor and QueryFlowSequence classes and the execute method.

Currently, the SDK provides a QueryFlowClient class, that can be instanced with the base url of the QueryFlow instance and an API key.