aetherflow-scheduler 0.0.14

APScheduler integration for aetherflow-core.

aetherflow-scheduler

aetherflow-scheduler is a thin cron-style scheduler for AetherFlow, built on APScheduler.

It reads a YAML file and triggers aetherflow runs on a schedule. The design goal is simplicity:

• Core stays run-once and deterministic
• Scheduler is the run-on-cron wrapper

This distribution ships the CLI aetherflow-scheduler.

---

Install

pip install aetherflow-scheduler

This depends on aetherflow-core and installs it automatically.

Python: 3.10+.

---

CLI

aetherflow-scheduler --help

aetherflow-scheduler run path/to/scheduler.yaml

---

Scheduler YAML (schema)

Top-level keys:

• timezone (string, optional): timezone name (IANA tz database). Default: "Europe/Berlin".
• items (list, required): list of scheduled items

Each item:

• id (string, required): APScheduler job id
• cron (string, required): crontab expression (e.g. "0 * * * *")
• flow_yaml (string, required): path to a flow YAML file
• flow (string, optional): flow id inside the YAML
• flow_job (string, optional): job id within the flow
• bundle_manifest (string, optional): bundle manifest path (sync before run)
• allow_stale_bundle (bool, optional): default false
• misfire_grace_time (int seconds, optional): default 300

Example:

timezone: Europe/Berlin
items:
  - id: hourly_sales
    cron: "0 * * * *"
    flow_yaml: "flows/sales.yaml"
    flow: "sales_flow"
    flow_job: "main"
    bundle_manifest: "bundles/sales_bundle.yaml"
    allow_stale_bundle: true
    misfire_grace_time: 600

Run it:

aetherflow-scheduler run scheduler.yaml

---

Overlap prevention and reliability

AetherFlow itself can use locks/state to avoid overlapping executions and to support recovery patterns. Scheduler is intentionally minimal: it triggers runs; the “safety” lives in core.

Relevant docs:

• Locking guide
• State
• Concurrency
• Failure Recovery Playbook

---

Namespace package rule (important)

AetherFlow is a PEP 420 implicit namespace across distributions.

Use:

• import aetherflow.scheduler
• import aetherflow.core
• from aetherflow.core.api import ...

Do not rely on ambiguous imports like aetherflow.x.

---

Docs (in this repository)

Scheduler-specific doc:

• Scheduler yaml guide

Start here:

• Home Docs.