llm_queue.py¶

Generated: 2026-03-30T04:38:47.530192+00:00

Commit: 615f1354a103

Language: python

Scope: dazzle_back

Purpose¶

Background job queue for async LLM intent execution.

Provides: - In-process async queue with configurable workers - Per-model semaphore (concurrency cap) - Per-model token bucket (rate limiting) - AIJob lifecycle management (pending → running → completed/failed) - LLM event emission on completion/f...

Key Exports¶

TokenBucket
LLMJob
LLMJobQueue
acquire
submit
start
shutdown
pending_count

Dependencies¶

Imports¶

__future__.annotations
asyncio
logging
time
collections.abc.Awaitable
collections.abc.Callable
dataclasses.dataclass
typing.TYPE_CHECKING
typing.Any
dazzle_back.runtime.event_bus.LLMEventType
dazzle_back.runtime.llm_executor.ExecutionResult
dazzle_back.runtime.llm_executor.LLMIntentExecutor
uuid
dazzle_back.runtime.event_bus.EntityEvent
dazzle_back.runtime.event_bus.EntityEventType

Event Interactions¶

Consumes: EntityEvent

Invariants and Assumptions¶

No invariants documented.

Tests¶

tests/unit/test_llm_queue.py
tests/integration/test_runtime_e2e.py
tests/integration/test_runtime_pipeline.py
tests/unit/test_runtime_test_routes.py
tests/unit/test_runtime_services.py

Notes¶

This page was automatically generated by tools/gen_reference_docs.py.