Skip to content

athena-sdk-lite

A Pythonic, single-import SDK for building Athena workflows.

from athena_sdk_lite import Workflow
from athena_sdk_lite.nodes import postgres, ai_tagging, branch, output

with Workflow("triage") as wf:
    rows = postgres("load", operation="select", query="SELECT ... FROM events", connection={...})
    tagged = ai_tagging("classify", inputs=rows, agent_url="https://...")
    gate = branch("is-critical", inputs=tagged, condition="$input.data.get('severity') == 'critical'")
    output("alerts", inputs=gate.out("true"), format="json")
    output("queue",  inputs=gate.out("false"), format="json")

print(wf.visualize())
issues = wf.validate()

That is the entire surface a normal user sees.


Where to go next

What this is, in one paragraph

A small Python library for building data + AI workflows as DAGs. You import it, declare nodes (Postgres read, AI classification, transform, branch), wire them with inputs=, and the library produces a workflow object you can validate, visualize, and run locally, in-process. No backend. No API key. No service to deploy. The vendored engine does the actual work; this package is a thin, obvious surface on top.

When to reach for this

Use this when Reach for something heavier when
Workflow runs in one Python process You need distributed execution across machines
You want stakeholders to read the workflow code You need non-engineers to author workflows
The 11 helpers + escape hatch cover your nodes You need first-class support for many bespoke node types
Local-first development and testing You need a UI / registry / scheduler