onloq 0.1.0

Single-execution lock decorator with Redis and file-based backends.

Onloq

Lightweight worker concurrency control via redis or file locks.

When multiple workers run the same task simultaneously, Onloq ensures only one proceeds — the rest skip silently. Supports TTL-based expiry, decorator and direct-call usage, and custom error handling.

---

Install

pip install onloq

---

Quick Start

from onloq import Onloq

loq = Onloq()  # uses filesystem locks by default

@loq
def my_task():
    ...

# Redis-backed
import redis
loq = Onloq(redis_connection=redis.Redis())

---

Usage

Decorator

@loq
def send_report():
    ...

@loq(name="custom.lock.name", ttl=60)
def sync_data():
    ...

Direct call

loq.execute(send_report, ttl=30)

---

Options

Parameter	Default	Description
redis_connection	None	Redis client instance. Falls back to file locks if not set.
locks_dir	tempdir/onloq_locks	Directory for file-based lock files.
ttl	25	Lock lifetime in seconds.
on_error	'raise'	Error behavior: 'raise', None (log and suppress), or a callable (name, exc, *args, **kwargs).
logger	None	Standard logger for warnings and errors.

---

Behavior

• If the lock is already held, the call returns None immediately — no waiting, no retry.
• File locks use mtime to enforce TTL, preventing stale locks from blocking indefinitely.
• Redis locks use atomic Lua scripts to ensure safe release by the original owner only.
• Lock names default to module.qualname of the wrapped function.

---

License

MIT