105 lines
3.9 KiB
Python
105 lines
3.9 KiB
Python
"""
|
|
P014: every connector used by pipeline.py's adaptive plan (gleif, brreg,
|
|
companies_house, prh, cvr) used to collapse EVERY non-200 response into a
|
|
bare `None`, indistinguishable from "record genuinely doesn't exist". This
|
|
tests that the block-status subset (401/403/429/451) now raises
|
|
SourceBlockedError instead, while every other non-200 status (404, 500, ...)
|
|
still returns None exactly as before -- this is an ADDITIVE signal, not a
|
|
behavior change for the non-block case.
|
|
"""
|
|
from __future__ import annotations
|
|
|
|
import httpx
|
|
import pytest
|
|
import respx
|
|
|
|
from sources.connectors import (
|
|
SourceBlockedError,
|
|
brreg_by_orgnr,
|
|
companies_house_by_number,
|
|
cvr_by_number,
|
|
gleif_by_lei,
|
|
prh_by_businessid,
|
|
)
|
|
|
|
GLEIF_URL = "https://api.gleif.org/api/v1/lei-records/5493001KJTIIGC8Y1R12"
|
|
BRREG_URL = "https://data.brreg.no/enhetsregisteret/api/enheter/923609016"
|
|
COMPANIES_HOUSE_URL = "https://api.company-information.service.gov.uk/company/00000006"
|
|
PRH_URL = "https://avoindata.prh.fi/opendata-ytj-api/v3/companies?businessId=2077474-0"
|
|
CVR_URL = "https://cvrapi.dk/api?search=10150817&country=dk"
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
@respx.mock
|
|
@pytest.mark.parametrize("status", [401, 403, 429, 451])
|
|
async def test_gleif_raises_source_blocked_on_block_statuses(status):
|
|
respx.get(GLEIF_URL).mock(return_value=httpx.Response(status, json={}))
|
|
with pytest.raises(SourceBlockedError) as exc_info:
|
|
await gleif_by_lei("5493001KJTIIGC8Y1R12")
|
|
assert exc_info.value.source_id == "gleif"
|
|
assert exc_info.value.status_code == status
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
@respx.mock
|
|
async def test_gleif_a_404_still_returns_none_not_blocked(monkeypatch):
|
|
respx.get(GLEIF_URL).mock(return_value=httpx.Response(404, json={}))
|
|
result = await gleif_by_lei("5493001KJTIIGC8Y1R12")
|
|
assert result is None # unchanged pre-P014 behavior for a genuine "not found"
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
@respx.mock
|
|
async def test_brreg_raises_source_blocked_on_429():
|
|
respx.get(BRREG_URL).mock(return_value=httpx.Response(429, json={}))
|
|
with pytest.raises(SourceBlockedError) as exc_info:
|
|
await brreg_by_orgnr("923609016")
|
|
assert exc_info.value.source_id == "brreg"
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
@respx.mock
|
|
async def test_companies_house_raises_source_blocked_on_403(monkeypatch):
|
|
monkeypatch.setenv("COMPANIES_HOUSE_API_KEY", "test-key")
|
|
respx.get(COMPANIES_HOUSE_URL).mock(return_value=httpx.Response(403, json={}))
|
|
with pytest.raises(SourceBlockedError) as exc_info:
|
|
await companies_house_by_number("00000006")
|
|
assert exc_info.value.source_id == "companies_house"
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_companies_house_missing_key_still_delivers_honest_level_0_not_a_block():
|
|
# No COMPANIES_HOUSE_API_KEY set -- this is a config gap, not a source
|
|
# block, and must keep returning the honest "unavailable" result it
|
|
# already did (never raise SourceBlockedError for a problem on OUR side).
|
|
result = await companies_house_by_number("00000006")
|
|
assert result is not None
|
|
assert result.provenance[0].level == 0
|
|
assert result.provenance[0].transform == "unavailable_no_key"
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
@respx.mock
|
|
async def test_prh_raises_source_blocked_on_401():
|
|
respx.get(PRH_URL).mock(return_value=httpx.Response(401, json={}))
|
|
with pytest.raises(SourceBlockedError) as exc_info:
|
|
await prh_by_businessid("2077474-0")
|
|
assert exc_info.value.source_id == "prh"
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
@respx.mock
|
|
async def test_cvr_raises_source_blocked_on_451():
|
|
respx.get(CVR_URL).mock(return_value=httpx.Response(451, json={}))
|
|
with pytest.raises(SourceBlockedError) as exc_info:
|
|
await cvr_by_number("10150817")
|
|
assert exc_info.value.source_id == "cvr"
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
@respx.mock
|
|
async def test_cvr_a_500_still_returns_none_not_blocked():
|
|
respx.get(CVR_URL).mock(return_value=httpx.Response(500, json={}))
|
|
result = await cvr_by_number("10150817")
|
|
assert result is None
|