Tirsvad
ec6ed213fb
Delivers MIL-004 (lecture Step 3): `DEFAULT_HEADERS` (User-Agent, Accept-Language, Accept, Accept-Encoding, Sec-Fetch-*, Priority, Upgrade-Insecure-Requests; no Host and no X-Amzn-Trace-Id) is defined once in `constants.py` and sent by `RequestsPageFetcher`. Tests assert the headers, the absence of Host/trace id, the timeout and read-only behaviour, with no network. Review: code against QC-PY-001 as RC-013 (Go, autonomous, not independent). 38 pytest tests, ruff and mypy --strict pass. Screenshots compare the browser's headers and the bot's headers at httpbin.org/headers. Task: MIL-004#1 Task: MIL-004#2 Task: MIL-004#3 Task: MIL-004#4
86 lines
2.5 KiB
Python
86 lines
2.5 KiB
Python
import pytest
|
|
import requests
|
|
|
|
from amazon_price_tracker.constants import DEFAULT_HEADERS, REQUEST_TIMEOUT_SECONDS
|
|
from amazon_price_tracker.infrastructure.http_client import RequestsPageFetcher
|
|
|
|
Call = tuple[str, dict[str, str], float]
|
|
|
|
|
|
class FakeResponse:
|
|
def __init__(self, text: str, status_code: int) -> None:
|
|
self.text = text
|
|
self.status_code = status_code
|
|
|
|
def raise_for_status(self) -> None:
|
|
if self.status_code >= 400:
|
|
raise requests.HTTPError(f"status {self.status_code}")
|
|
|
|
|
|
def _patch_get(monkeypatch: pytest.MonkeyPatch, response: FakeResponse) -> list[Call]:
|
|
calls: list[Call] = []
|
|
|
|
def fake_get(url: str, headers: dict[str, str], timeout: float) -> FakeResponse:
|
|
calls.append((url, headers, timeout))
|
|
return response
|
|
|
|
monkeypatch.setattr(requests, "get", fake_get)
|
|
return calls
|
|
|
|
|
|
def test_fetch_returns_text_using_url_and_timeout(
|
|
monkeypatch: pytest.MonkeyPatch,
|
|
) -> None:
|
|
calls = _patch_get(monkeypatch, FakeResponse("<p>hi</p>", 200))
|
|
|
|
html = RequestsPageFetcher().fetch("https://example.test/")
|
|
|
|
assert html == "<p>hi</p>"
|
|
assert calls == [
|
|
("https://example.test/", dict(DEFAULT_HEADERS), REQUEST_TIMEOUT_SECONDS)
|
|
]
|
|
|
|
|
|
def test_fetch_sends_user_agent_and_accept_language(
|
|
monkeypatch: pytest.MonkeyPatch,
|
|
) -> None:
|
|
calls = _patch_get(monkeypatch, FakeResponse("", 200))
|
|
|
|
RequestsPageFetcher().fetch("https://example.test/")
|
|
|
|
headers = calls[0][1]
|
|
assert headers["User-Agent"]
|
|
assert headers["Accept-Language"]
|
|
|
|
|
|
def test_fetch_omits_host_and_trace_id_headers(
|
|
monkeypatch: pytest.MonkeyPatch,
|
|
) -> None:
|
|
calls = _patch_get(monkeypatch, FakeResponse("", 200))
|
|
|
|
RequestsPageFetcher().fetch("https://example.test/")
|
|
|
|
assert "Host" not in calls[0][1]
|
|
assert "X-Amzn-Trace-Id" not in calls[0][1]
|
|
|
|
|
|
def test_fetch_uses_overriding_headers(monkeypatch: pytest.MonkeyPatch) -> None:
|
|
calls = _patch_get(monkeypatch, FakeResponse("", 200))
|
|
|
|
RequestsPageFetcher(headers={"User-Agent": "custom"}).fetch("https://example.test/")
|
|
|
|
assert calls[0][1] == {"User-Agent": "custom"}
|
|
assert calls[0][2] == REQUEST_TIMEOUT_SECONDS
|
|
|
|
|
|
def test_default_headers_are_read_only() -> None:
|
|
with pytest.raises(TypeError):
|
|
DEFAULT_HEADERS["User-Agent"] = "x" # type: ignore[index]
|
|
|
|
|
|
def test_fetch_raises_on_http_error(monkeypatch: pytest.MonkeyPatch) -> None:
|
|
_patch_get(monkeypatch, FakeResponse("", 404))
|
|
|
|
with pytest.raises(requests.HTTPError):
|
|
RequestsPageFetcher().fetch("https://example.test/")
|