Repair19 of 26 · about 10 min

Repair: the API that lies about the total

Northstar's case API pages at 50 and reports a total on page 1 that is computed BEFORE the account filter is applied. A client that stops when it has collected total records stops in the wrong place — sometimes early, sometimes never.

It also has two failure modes that do not look like failures. About one call in fifty returns **200 with an empty body**, and it rate-limits at an undocumented threshold with a **429 and no Retry-After**. response.ok is true for the first and the second needs a backoff you have to choose yourself.

Walk until the server says there is no more. Treat an empty body as a retryable failure, not as the end of the data — those two are indistinguishable to a loop that only checks for an empty list.

Example

# Wrong: trusts a number the vendor computes before filtering
while len(collected) < body["total"]:
    ...

# Right: walks until the server stops saying there is more
while has_more:
    ...

Your task

The recorded responses are in pages.json, in the order the server returned them. Fix fetch_all so it collects every case exactly once, retries an empty body and a 429, and prints the ids it collected followed by the retry count.

Stuck?

Explain it

Real retries need a delay. With no Retry-After header, how would you choose one, and what stops the loop forever?

Not graded — this is for you.

Where this goes

D3 integrates with this API for real. Its rubric asks you to demonstrate the empty-body case specifically, because it is the one most adapters never handle.

import json

responses = json.loads(open("pages.json").read())["responses"]


def next_response():
    """Hand back the next recorded response, in order."""
    if not responses:
        return None
    return responses.pop(0)


def fetch_all():
    collected = []
    retries = 0
    # BUG: stops when it has as many records as page 1 claimed, and treats a
    # missing body as the end of the data.
    total = None
    while True:
        response = next_response()
        if response is None:
            break
        body = response["body"]
        if body is None:
            break
        if total is None:
            total = body["total"]
        for case in body["data"]:
            collected.append(case["case_id"])
        if len(collected) >= total:
            break
    return collected, retries


ids, retries = fetch_all()
print(",".join(ids))
print("retries:", retries)

Files you can open: pages.json

Ctrl+Enter checks

This is a teaching runtime for a subset of Python, running in your browser. Integers are exact and the errors are written in plain English, but the standard library is a small subset and anything missing is refused by name rather than approximated. Your project runs on real Python.

Press Run to see what your program does, or Check when you think it is right. Everything runs here in your browser.