bisco/hoopscout
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(v2-ingestion): harden LBA/BCL snapshot contract for public data

Browse Source
This commit is contained in:
Alfredo Di Stasio
2026-03-20 14:46:25 +01:00
parent 1aad6945c7
commit 6066d2a0bb
10 changed files with 339 additions and 31 deletions
Download Patch File Download Diff File Expand all files Collapse all files

25
tests/fixtures/bcl/bcl_players_stats_partial_public.json vendored Normal file
View File

@ -0,0 +1,25 @@
{
"data": [
{
"player": {
"id": "bcl-player-99",
"name": "Alex Novak"
},
"team": {
"id": "bcl-team-tenerife",
"name": "Lenovo Tenerife"
},
"gp": 10,
"mpg": 27.2,
"ppg": 14.8,
"rpg": 4.1,
"apg": 3.3,
"spg": 1.2,
"bpg": 0.4,
"tov": 2.0,
"fg_pct": 47.3,
"three_pct": 38.0,
"ft_pct": 79.1
}
]
}

25
tests/fixtures/lba/lba_players_stats_partial_public.json vendored Normal file
View File

@ -0,0 +1,25 @@
{
"data": [
{
"player": {
"id": "p-002",
"name": "Andrea Bianchi"
},
"team": {
"id": "team-olimpia-milano",
"name": "Olimpia Milano"
},
"gp": 18,
"mpg": 24.7,
"ppg": 12.3,
"rpg": 2.9,
"apg": 4.2,
"spg": 1.1,
"bpg": 0.1,
"tov": 1.8,
"fg_pct": 45.0,
"three_pct": 35.4,
"ft_pct": 82.7
}
]
}

31
tests/test_bcl_extractor.py
View File

@ -51,6 +51,37 @@ def test_bcl_extractor_normalizes_fixture_payload(tmp_path, settings):
assert row["three_pt_pct"] == 37.2
@pytest.mark.django_db
def test_bcl_extractor_accepts_partial_public_player_bio_fields(tmp_path, settings):
settings.EXTRACTOR_BCL_STATS_URL = "https://www.championsleague.basketball/public/stats.json"
settings.EXTRACTOR_BCL_SEASON_LABEL = "2025-2026"
settings.EXTRACTOR_BCL_COMPETITION_EXTERNAL_ID = "bcl"
settings.EXTRACTOR_BCL_COMPETITION_NAME = "Basketball Champions League"
fixture_payload = _load_fixture("bcl/bcl_players_stats_partial_public.json")
class FakeClient:
def get_json(self, *_args, **_kwargs):
return fixture_payload
extractor = BCLSnapshotExtractor(http_client=FakeClient())
output_path = tmp_path / "bcl-partial.json"
result = extractor.run(output_path=output_path, snapshot_date=date(2026, 3, 13))
assert result.records_count == 1
payload = json.loads(output_path.read_text(encoding="utf-8"))
row = payload["records"][0]
assert row["full_name"] == "Alex Novak"
assert row["first_name"] is None
assert row["last_name"] is None
assert row["birth_date"] is None
assert row["nationality"] is None
assert row["height_cm"] is None
assert row["weight_kg"] is None
assert row["position"] is None
assert row["games_played"] == 10
@pytest.mark.django_db
def test_bcl_extractor_registry_selection(settings):
settings.EXTRACTOR_BCL_STATS_URL = "https://www.championsleague.basketball/public/stats.json"

80
tests/test_import_snapshots_command.py
View File

@ -103,6 +103,86 @@ def test_valid_snapshot_import(tmp_path, settings):
assert PlayerSeasonStats.objects.count() == 1
@pytest.mark.django_db
def test_snapshot_import_succeeds_with_optional_bio_and_physical_fields_missing(tmp_path, settings):
incoming = tmp_path / "incoming"
archive = tmp_path / "archive"
failed = tmp_path / "failed"
incoming.mkdir()
archive.mkdir()
failed.mkdir()
payload = _valid_payload()
for optional_field in ("first_name", "last_name", "birth_date", "nationality", "height_cm", "weight_kg", "position", "role"):
payload["records"][0].pop(optional_field, None)
file_path = incoming / "optional-missing.json"
_write_json(file_path, payload)
settings.STATIC_DATASET_INCOMING_DIR = str(incoming)
settings.STATIC_DATASET_ARCHIVE_DIR = str(archive)
settings.STATIC_DATASET_FAILED_DIR = str(failed)
call_command("import_snapshots")
run = ImportRun.objects.get()
assert run.status == ImportRun.RunStatus.SUCCESS
player = Player.objects.get(source_uid="player-23")
assert player.first_name == "LeBron"
assert player.last_name == "James"
assert player.birth_date is None
assert player.nationality is None
assert player.nominal_position is None
assert player.height_cm is None
assert player.weight_kg is None
assert PlayerSeasonStats.objects.count() == 1
@pytest.mark.django_db
@pytest.mark.parametrize(
("source_name", "competition_id", "competition_name"),
[
("lba", "lba-serie-a", "Lega Basket Serie A"),
("bcl", "bcl", "Basketball Champions League"),
],
)
def test_partial_public_source_snapshot_imports_for_lba_and_bcl(
tmp_path,
settings,
source_name,
competition_id,
competition_name,
):
incoming = tmp_path / "incoming"
archive = tmp_path / "archive"
failed = tmp_path / "failed"
incoming.mkdir()
archive.mkdir()
failed.mkdir()
payload = _valid_payload()
payload["source_name"] = source_name
row = payload["records"][0]
row["competition_external_id"] = competition_id
row["competition_name"] = competition_name
for optional_field in ("first_name", "last_name", "birth_date", "nationality", "height_cm", "weight_kg", "position", "role"):
row.pop(optional_field, None)
_write_json(incoming / f"{source_name}.json", payload)
settings.STATIC_DATASET_INCOMING_DIR = str(incoming)
settings.STATIC_DATASET_ARCHIVE_DIR = str(archive)
settings.STATIC_DATASET_FAILED_DIR = str(failed)
call_command("import_snapshots")
run = ImportRun.objects.get()
assert run.status == ImportRun.RunStatus.SUCCESS
assert Competition.objects.filter(source_uid=competition_id, name=competition_name).exists()
assert Player.objects.filter(source_uid="player-23").exists()
assert PlayerSeasonStats.objects.count() == 1
@pytest.mark.django_db
def test_invalid_snapshot_rejected_and_moved_to_failed(tmp_path, settings):
incoming = tmp_path / "incoming"

51
tests/test_lba_extractor.py
View File

@ -8,6 +8,7 @@ import pytest
from django.core.management import call_command
from apps.ingestion.extractors.lba import LBASnapshotExtractor
from apps.ingestion.extractors.base import ExtractorNormalizationError
from apps.ingestion.extractors.registry import create_extractor
@ -51,6 +52,56 @@ def test_lba_extractor_normalizes_fixture_payload(tmp_path, settings):
assert row["three_pt_pct"] == 36.5
@pytest.mark.django_db
def test_lba_extractor_accepts_partial_public_player_bio_fields(tmp_path, settings):
settings.EXTRACTOR_LBA_STATS_URL = "https://www.legabasket.it/public/stats.json"
settings.EXTRACTOR_LBA_SEASON_LABEL = "2025-2026"
settings.EXTRACTOR_LBA_COMPETITION_EXTERNAL_ID = "lba-serie-a"
settings.EXTRACTOR_LBA_COMPETITION_NAME = "Lega Basket Serie A"
fixture_payload = _load_fixture("lba/lba_players_stats_partial_public.json")
class FakeClient:
def get_json(self, *_args, **_kwargs):
return fixture_payload
extractor = LBASnapshotExtractor(http_client=FakeClient())
output_path = tmp_path / "lba-partial.json"
result = extractor.run(output_path=output_path, snapshot_date=date(2026, 3, 13))
assert result.records_count == 1
payload = json.loads(output_path.read_text(encoding="utf-8"))
row = payload["records"][0]
assert row["full_name"] == "Andrea Bianchi"
assert row["first_name"] is None
assert row["last_name"] is None
assert row["birth_date"] is None
assert row["nationality"] is None
assert row["height_cm"] is None
assert row["weight_kg"] is None
assert row["position"] is None
assert row["games_played"] == 18
@pytest.mark.django_db
def test_lba_extractor_still_fails_when_required_stats_are_missing(settings):
settings.EXTRACTOR_LBA_STATS_URL = "https://www.legabasket.it/public/stats.json"
settings.EXTRACTOR_LBA_SEASON_LABEL = "2025-2026"
settings.EXTRACTOR_LBA_COMPETITION_EXTERNAL_ID = "lba-serie-a"
settings.EXTRACTOR_LBA_COMPETITION_NAME = "Lega Basket Serie A"
fixture_payload = _load_fixture("lba/lba_players_stats_partial_public.json")
fixture_payload["data"][0].pop("ppg")
class FakeClient:
def get_json(self, *_args, **_kwargs):
return fixture_payload
extractor = LBASnapshotExtractor(http_client=FakeClient())
with pytest.raises(ExtractorNormalizationError):
extractor.run(write_output=False, snapshot_date=date(2026, 3, 13))
@pytest.mark.django_db
def test_lba_extractor_registry_selection(settings):
settings.EXTRACTOR_LBA_STATS_URL = "https://www.legabasket.it/public/stats.json"