Files
jobboard/tests/test_ats.py
T
WolfandClaude Sonnet 5 867ebe6405 Raise the ats body cap to 96 MB, retry Greenhouse without content=true
openings.go's 8 MB cap was written against arbitrary URLs; every URL here is
on a fixed allow-list of seven ATS hosts, and the real pull of 2026-09-23
refused Anduril, Anthropic, OpenAI, Databricks, Carvana, Axon and Airwallex
for it. A board that still exceeds 96 MB now retries once without
content=true (Greenhouse only), trading the description for the listing.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-09-23 11:39:39 +02:00

176 lines
7.0 KiB
Python

"""The `ats` plugin: each host's parser against a saved feed, the function
classifier, and the fetch allow-list. No network — every fixture is a file
under `tests/fixtures/ats/`."""
from __future__ import annotations
import sys
from pathlib import Path
import pytest
ROOT = Path(__file__).resolve().parents[1]
if str(ROOT) not in sys.path:
sys.path.insert(0, str(ROOT))
from plugins import ats # noqa: E402
FIXTURES = ROOT / "tests" / "fixtures" / "ats"
def _rows(host: str, slug: str = "acme") -> list[ats.Row]:
path = FIXTURES / f"{host}.{'xml' if host == 'personio' else 'json'}"
body = path.read_bytes()
return ats._PARSERS[host](body, slug)
# ------------------------------------------------------------------ hosts
def test_greenhouse_parses_titles_locations_and_html_description() -> None:
rows = _rows("greenhouse")
# The javascript: and data: rows are refused, same as the two real ones
# are kept: a feed is external content, never trusted as a URL.
assert [r["source_id"] for r in rows] == ["1001", "1002"]
first = rows[0]
assert first["title"] == "Senior Backend Engineer" # trimmed
assert first["location"] == "Berlin, Germany"
assert first["department"] == "Engineering"
assert first["description"] == "Build our core platform."
assert first["posted"] is not None and first["posted"].year == 2026
assert first["remote"] is None
def test_lever_parses_categories_and_workplace_type() -> None:
rows = _rows("lever")
assert [r["source_id"] for r in rows] == ["2001", "2002"]
first = rows[0]
assert first["department"] == "Engineering"
assert first["remote"] == "remote"
assert first["description"] == "Own our platform reliability."
# createdAt=0 means "no date", not epoch zero.
assert rows[1]["posted"] is None
def test_ashby_parses_is_remote_and_compensation() -> None:
rows = _rows("ashby")
assert [r["source_id"] for r in rows] == ["3001", "3002"]
assert rows[0]["remote"] is False
assert rows[0]["salary"] == "70,000-90,000 EUR"
assert rows[1]["remote"] is True
def test_workable_parses_location_parts_and_telecommute() -> None:
rows = _rows("workable")
assert [r["source_id"] for r in rows] == ["4001", "4002"]
assert rows[0]["location"] == "Berlin, Germany"
assert rows[0]["remote"] is True
assert "Ship our API." in rows[0]["description"]
def test_smartrecruiters_builds_the_job_url_from_id_and_slug() -> None:
rows = _rows("smartrecruiters", slug="Acme")
assert [r["source_id"] for r in rows] == ["744000151326819", "744000147485112"]
assert rows[0]["url"] == "https://jobs.smartrecruiters.com/Acme/744000151326819"
assert rows[0]["location"] == "Berlin, Germany"
assert rows[1]["remote"] is True
def test_recruitee_parses_salary_bounds_and_careers_url() -> None:
rows = _rows("recruitee")
assert [r["source_id"] for r in rows] == ["5001", "5002"]
assert rows[0]["url"] == "https://acme.example/o/player-support-agent-german-1"
assert rows[0]["salary"] == "35000-45000 EUR"
assert rows[1]["salary"] == ""
def test_personio_builds_the_job_url_and_reads_cdata_description() -> None:
rows = _rows("personio", slug="acme")
assert [r["source_id"] for r in rows] == ["6001", "6002"]
assert rows[0]["url"] == "https://acme.jobs.personio.de/job/6001"
assert rows[0]["description"] == "Grow our pipeline."
assert rows[0]["remote"] is True # "Remote - Germany"
assert rows[1]["remote"] is None # "Munich"
# --------------------------------------------------------------- classify
def test_classify_matches_department_before_title() -> None:
assert ats.classify("Engineering", "Marketing Lead") == "engineering"
assert ats.classify("", "Senior Backend Engineer") == "engineering"
assert ats.classify("", "Product Manager, Payments") == "product"
assert ats.classify("", "Data Product Manager") == "product"
assert ats.classify("", "Machine Learning Engineer") == "data"
assert ats.classify("", "Executive Assistant to the CEO") == "finance/legal/ops"
assert ats.classify("", "") == "other"
# ------------------------------------------------------------- allow-list
def test_is_known_target_accepts_the_seven_hosts_and_refuses_others() -> None:
for host in ("greenhouse", "lever", "ashby", "workable", "smartrecruiters",
"recruitee", "personio"):
assert ats.is_known_target(f"{host}:acme")
assert not ats.is_known_target("linkedin:acme")
assert not ats.is_known_target("workday:acme")
assert not ats.is_known_target("greenhouse:")
assert not ats.is_known_target("nohostnoseparator")
def test_allowed_host_refuses_a_foreign_host_per_board() -> None:
allow = ats._allowed_host("greenhouse", "acme")
assert allow("boards-api.greenhouse.io")
assert not allow("boards-api.greenhouse.io.evil.example")
assert not allow("api.lever.co")
# recruitee and personio key off the company's own subdomain, so one
# company's slug must not pass for another's.
allow_recruitee = ats._allowed_host("recruitee", "acme")
assert allow_recruitee("acme.recruitee.com")
assert not allow_recruitee("other.recruitee.com")
def test_careers_url_is_filled_per_host_and_empty_for_unknown() -> None:
assert ats.careers_url("greenhouse:acme") == "https://job-boards.greenhouse.io/acme"
assert ats.careers_url("ashby:acme") == "https://jobs.ashbyhq.com/acme"
assert ats.careers_url("recruitee:acme") == "https://acme.recruitee.com"
assert ats.careers_url("workday:acme") == ""
# --------------------------------------------------------------- body cap
def test_max_bytes_is_96_mb() -> None:
"""Raised from openings.go's 8 MB: every URL here is on a fixed allow-list
of ATS API hosts, not an arbitrary one, and the real run of 2026-09-23
refused boards up to 41 MB under it."""
assert ats.MAX_BYTES == 96 << 20
def test_fetch_rows_retries_greenhouse_without_content_on_a_too_large_body(
monkeypatch: pytest.MonkeyPatch) -> None:
calls = []
def fake_get(client, url, host, slug, *, method="GET"):
calls.append(url)
if "?content=true" in url:
raise ats._TooLarge("response too large (100000000 bytes)")
return (FIXTURES / "greenhouse.json").read_bytes()
monkeypatch.setattr(ats, "_get", fake_get)
rows = ats._fetch_rows(object(), "greenhouse:acme") # type: ignore[arg-type]
assert calls == [
"https://boards-api.greenhouse.io/v1/boards/acme/jobs?content=true",
"https://boards-api.greenhouse.io/v1/boards/acme/jobs",
]
assert [r["source_id"] for r in rows] == ["1001", "1002"]
def test_fetch_rows_does_not_retry_a_too_large_body_on_other_hosts(
monkeypatch: pytest.MonkeyPatch) -> None:
def fake_get(client, url, host, slug, *, method="GET"):
raise ats._TooLarge("response too large (100000000 bytes)")
monkeypatch.setattr(ats, "_get", fake_get)
with pytest.raises(ats._TooLarge):
ats._fetch_rows(object(), "lever:acme") # type: ignore[arg-type]