exobrain-actions-client

This package provides a REST API client for interacting with the Exobrain Actions API.

Key features:

• Access to various API resources such as version status and health checks (ping).
• Retrieval, triggering, or updating of action data specific to a service or organization.
• Standardized HTTP calls to different API endpoints (version, health, action data, calculation triggers, scores, AI summaries).

User Guide

Installation

To install the Exobrain Actions Client, use the following command:

pip install exobrain-actions-client

Usage

To retrieve the AI-generated summary of a given action, create a client, instantiate the accessor, and call get_summary with the service name, organization ID, action ID, and the target language.

from uuid import UUID

from exobrain.actions.client.accessor import ActionsAccessor
from exobrain.actions.client.http_client import ClientType, create_client
from exobrain.actions.client.http_client.interfaces import HTTPStatusError

org_id = UUID("12345678-1234-5678-1234-567812345678")
action_id = UUID("87654321-4321-8765-4321-876543218765")

base_url = "https://actions-service.com/"
service = "stock-rebalancing"

client = create_client(
    ClientType.REQUESTS,
    base_url=base_url,
    verify=True,  # Set to False if you want to disable SSL verification
    timeout=10,
)
accessor = ActionsAccessor(client)
try:
    summary = accessor.get_summary(service, org_id, action_id, lang="en")
except HTTPStatusError as e:
    print(f"Failed to retrieve the summary: {e.response.status_code} - {e}")
else:
    print(f"Summary: {summary}")

Note — updating the "ACTUAL" KPI values is no longer done through this client. exobrain-server writes them directly into the kpi_mapping JSONB column of the running action context (EXO-1510).

Development workflow

Prerequisites

Before starting development, make sure you have the following tools installed:

• Hatch: a modern Python project, environment, and packaging manager.

• uv: a fast Python package installer and resolver.

Project Structure

• Code: Place your models in exobrain/actions/client/.
• Tests: Add your tests in tests/.

Development Commands

All tasks (testing, linting, type checking, packaging, etc.) are managed via hatch environments and scripts:

1. Installing Dependencies

Install all dependencies (including development tools):

hatch env create

2. Running Tests

Launch all tests using pytest:

hatch test

3. Linting and Formatting

Format your code:

hatch fmt -f

Check code style:

hatch fmt

4. Type Checking

Run type checks with mypy:

hatch run types:check

5. Coverage

Check test coverage with:

hatch test --cover

6. Building the Package

Build the package for distribution:

hatch build

7. Releasing

The release process uses uv and Github Actions.

• For pre-release on TestPyPI: push a tag with rc or beta in the version (0.1rc1, 0.2b1 etc.).
• For release: push a tag type v0.1 (main release) on main branch.
• Configure secrets for PyPI tokens in repository settings (PYPI_TOKEN, TEST_PYPI_TOKEN).

Continuous Integration

All pushes and pull requests on main or development branches trigger the Continuous Integration pipeline
(see .github/workflows/ directory for details).

License

This project is private and confidential.