ML Practice
…

Thread Pool

hardpy_threads

Implement a reusable ThreadPool that executes submitted callables across a fixed number of background worker threads, returning Future handles for asynchronous result retrieval.

The concurrent and multiprocessing modules are off-limits.

Interface

  • Future:
    • result(timeout: float | None = None) -> Any: Block until the computation finishes and return its return value. If the task raised an exception, re-raise that exception here. If timeout is given and expires before completion, raise TimeoutError.
    • exception(timeout: float | None = None) -> BaseException | None: Block until the computation finishes and return the exception raised, or None if it succeeded. If timeout expires, raise TimeoutError.
    • done() -> bool: Return True if the computation has finished (succeeded or raised).
  • ThreadPool(num_workers: int):
    • submit(fn: Callable[..., Any], *args: Any, **kwargs: Any) -> Future: Submit a task to be executed by a worker. Return a Future representing the pending result. If called after shutdown(), raise RuntimeError.
    • shutdown(wait: bool = True) -> None: Signal workers to shut down. No further tasks can be submitted. If wait=True, block until all pending tasks and worker threads have finished.
    • Context manager: Can be used in a with ThreadPool(...) as pool: block; calls shutdown(wait=True) on exit.
Constraints
  • no concurrent, multiprocessing

Your solution

Edit ThreadPool and run the real pytest suite in your browser — no install required. Your code is saved locally.

⌘/Ctrl + ↵

Loading the Python runtime (first run only)…

Hints

Hint 1

Create worker threads that loop pulling tasks from a work queue until a sentinel object (e.g. None) is received.

Hint 2

Future can use threading.Event to signal when the task has produced a result or raised an exception.

Hint 3

When a worker encounters an exception from a task, catch BaseException and store it in the Future so future.result() can re-raise it.

Hint 4

submit() should check if the pool is shutdown and raise RuntimeError if so.

Hint 5

shutdown(wait=True) should enqueue one sentinel per worker thread and join all worker threads.

Stuck? Try these first

easyThread-Safe Bounded Queue