madsci.common 0.4.0

The Modular Autonomous Discovery for Science (MADSci) Common Definitions and Utilities.

MADSci Common

Common types, validators, serializers, utilities, and other flotsam and jetsam used across the MADSci toolkit.

Installation

The MADSci common components are available via the Python Package Index, and can be installed via:

pip install madsci.common

This python package is also included as part of the madsci Docker image.

MADSci Types

The MADSci toolkit uses a variety of "types", implemented as Pydantic Data Models. These data models allow us to easily create, validate, serialize, and de-serialize data structures used throughout the distributed systems. They can easily be serialized to JSON when being sent between system components over REST or stored in JSON-friendly databases like MongoDB or Redis, or to YAML for human-readable and editable definition files.

You can import these types from madsci.common.types, where they are organized by subsystem.

Settings

MADSci uses Pydantic Settings to configure many of it's subsystems. This allows users to configure managers, clients, nodes, and other MADSci components using command line arguments, environment variables, settings files in various formats (.env, .toml, .yaml, .json), and secrets files.

Detailed documentation for what configuration can be set is included in the Configuration.md, and an example .env is included in the root of the MADSci repository.

In general, each subsystem supports configuration via both a generic file (.env, settings.yaml, etc), and a subsystem-specific file (event_client.env, event_client.yaml, etc). In such cases, the subsystem specific version takes precedence over the generic version.