chapter_20_concurrent_executors

Chapter 20 — Concurrent Executors

Theme: concurrent.futures elevates Python concurrency from raw low-level thread/process spawning into a high-level orchestration model based on Futures.

What You'll Learn

Method	Returns	Blocking Behavior	Best For
executor.submit(func, args)	A single Future object	Non-blocking.	Disparate tasks with varied arguments.
executor.map(func, iter)	A generator of results	Blocks until the next item in order is ready.	Drop-in replacement for standard map.
futures.as_completed(futures)	An iterator of completed Futures	Blocks until any future completes.	Progress bars, responsive UIs, logging.
future.result()	The return value of the task	Blocks until that specific task is done.	Final retrieval of data.

Key Files

• examples.py — Demonstrating map, submit, and as_completed.
• exercises.py — The absolute necessity of checking .result() or .exception() to avoid silent thread failures.
• mini_project.py — A live server health dashboard that yields real-time statuses as servers respond, rather than batching them.
• benchmarks.py — A Time-To-First-Result benchmark. Proves that if Task 0 takes 10s and Task 1 takes 1s, map() blocks you for 10s, while as_completed() yields Task 1 immediately.
• notes.md — What a Future actually represents and the __exit__ guarantees of Executors.
• pitfalls.md — The silent memory explosion trap of submitting 1 million tasks into an unbounded executor queue.
• interview_questions.md — L3 to L6 interview prep.
• architecture_notes.md — How Python manages internal work queues and thread pooling.

30-Second Rules

# Rule 1: Always retrieve the results of your futures to catch exceptions!
# BAD: Exception is silently swallowed.
executor.submit(flaky_task)

# GOOD: Exception will be raised and caught here.
future = executor.submit(flaky_task)
future.result() 

# Rule 2: If you want to show a progress bar (like tqdm), you MUST use as_completed.
with ThreadPoolExecutor() as ex:
    futures = [ex.submit(work, x) for x in range(100)]
    for f in tqdm(as_completed(futures)):
        result = f.result()

Reference: Fluent Python 2nd ed., Chapter 20 — pages 765–810