Add QueueRepositoryInterface with memory and file-backed adapters.

Provide a generic disk-backed FIFO queue with configurable path, retention, and dedup keys so consumers can buffer items across restarts without optional extras.

Co-authored-by: Cursor <[email protected]>
This commit is contained in:
Brian Bjarke Jensen
2026-07-16 20:18:00 +02:00
co-authored by Cursor
parent 1babb52d09
commit a768474892
15 changed files with 1092 additions and 13 deletions
+20 -4
View File
@@ -46,6 +46,8 @@ assert JsonRepositoryInterface is not None
assert "python_repositories.adapters.redis_adapter" not in sys.modules
assert "python_repositories.adapters.minio_adapter" not in sys.modules
assert "python_repositories.adapters.postgres_adapter" not in sys.modules
assert "python_repositories.adapters.memory_queue_adapter" not in sys.modules
assert "python_repositories.adapters.file_backed_queue_adapter" not in sys.modules
"""
result = subprocess.run(
[sys.executable, "-c", script],
@@ -59,11 +61,19 @@ assert "python_repositories.adapters.postgres_adapter" not in sys.modules
def test_lazy_adapter_load_succeeds_when_extra_present() -> None:
"""Adapters load when their optional dependencies are installed."""
from python_repositories import MinioAdapter, PostgresAdapter, RedisAdapter
from python_repositories import (
FileBackedQueueAdapter,
MemoryQueueAdapter,
MinioAdapter,
PostgresAdapter,
RedisAdapter,
)
assert RedisAdapter.__name__ == "RedisAdapter"
assert MinioAdapter.__name__ == "MinioAdapter"
assert PostgresAdapter.__name__ == "PostgresAdapter"
assert MemoryQueueAdapter.__name__ == "MemoryQueueAdapter"
assert FileBackedQueueAdapter.__name__ == "FileBackedQueueAdapter"
def test_redis_adapter_import_error_without_extra() -> None:
@@ -149,9 +159,13 @@ def test_adapters_dir_exposes_lazy_exports() -> None:
"""dir(adapters) includes lazy adapter names for tab completion."""
import python_repositories.adapters as adapters
assert {"RedisAdapter", "MinioAdapter", "PostgresAdapter"}.issubset(
set(dir(adapters))
)
assert {
"RedisAdapter",
"MinioAdapter",
"PostgresAdapter",
"MemoryQueueAdapter",
"FileBackedQueueAdapter",
}.issubset(set(dir(adapters)))
def test_adapters_getattr_raises_for_unknown() -> None:
@@ -167,3 +181,5 @@ def test_top_level_dir_exposes_lazy_exports() -> None:
assert "RedisAdapter" in dir(python_repositories)
assert "MinioAdapter" in dir(python_repositories)
assert "PostgresAdapter" in dir(python_repositories)
assert "MemoryQueueAdapter" in dir(python_repositories)
assert "FileBackedQueueAdapter" in dir(python_repositories)
@@ -0,0 +1,241 @@
"""Unit tests for FileBackedQueueAdapter."""
from __future__ import annotations
from datetime import UTC, datetime, timedelta
from pathlib import Path
import pytest
from python_repositories.adapters.file_backed_queue_adapter import (
FileBackedQueueAdapter,
)
from python_repositories.adapters.memory_queue_adapter import MemoryQueueAdapter
from python_repositories.config.file_queue_config import FileQueueConfig
from python_repositories.interfaces.queue_repository_interface import (
QueueRepositoryInterface,
)
def _config(path: Path, **kwargs: object) -> FileQueueConfig:
return FileQueueConfig(path=path, **kwargs) # type: ignore[arg-type]
def test_implements_interface() -> None:
assert issubclass(FileBackedQueueAdapter, QueueRepositoryInterface)
def test_ops_require_connect(tmp_path: Path) -> None:
queue = FileBackedQueueAdapter(config=_config(tmp_path / "q.jsonl"))
with pytest.raises(RuntimeError, match="not connected"):
queue.enqueue([{"id": 1}])
with pytest.raises(RuntimeError, match="not connected"):
queue.dequeue_batch()
with pytest.raises(RuntimeError, match="not connected"):
queue.size()
with pytest.raises(RuntimeError, match="not connected"):
queue.evict_older_than(datetime.now(UTC))
def test_enqueue_dequeue_persists_and_compacts(tmp_path: Path) -> None:
path = tmp_path / "items.jsonl"
with FileBackedQueueAdapter(config=_config(path, dedup_keys=("id",))) as queue:
queue.enqueue([{"id": 1}, {"id": 2}, {"id": 1}])
assert queue.size() == 2
assert path.is_file()
lines = path.read_text(encoding="utf-8").strip().splitlines()
assert len(lines) == 2
batch = queue.dequeue_batch(max_items=1)
assert batch == [{"id": 1}]
remaining = path.read_text(encoding="utf-8").strip().splitlines()
assert len(remaining) == 1
assert queue.size() == 1
def test_replay_on_connect(tmp_path: Path) -> None:
path = tmp_path / "items.jsonl"
config = _config(path, dedup_keys=("id",), age_key="created_at", max_age_hours=24)
now = datetime.now(UTC)
with FileBackedQueueAdapter(config=config) as queue:
queue.enqueue(
[
{"id": 1, "created_at": now - timedelta(hours=1)},
{"id": 2, "created_at": now - timedelta(minutes=10)},
]
)
with FileBackedQueueAdapter(config=config) as queue:
assert queue.size() == 2
assert [item["id"] for item in queue.dequeue_batch(max_items=10)] == [1, 2]
def test_replay_filters_by_max_age(tmp_path: Path) -> None:
path = tmp_path / "items.jsonl"
now = datetime.now(UTC)
path.write_text(
"\n".join(
[
f'{{"id": "old", "created_at": "{(now - timedelta(hours=30)).isoformat()}"}}',
f'{{"id": "new", "created_at": "{(now - timedelta(hours=1)).isoformat()}"}}',
]
)
+ "\n",
encoding="utf-8",
)
config = _config(path, dedup_keys=("id",), age_key="created_at", max_age_hours=24)
with FileBackedQueueAdapter(config=config) as queue:
assert queue.size() == 1
assert queue.dequeue_batch(max_items=10)[0]["id"] == "new"
def test_connect_skips_corrupt_and_invalid_lines(tmp_path: Path) -> None:
path = tmp_path / "items.jsonl"
path.write_text(
"\n".join(
[
"",
"not-json",
"[1, 2]",
'{"id": 1}',
'{"name": "missing-id"}',
]
)
+ "\n",
encoding="utf-8",
)
config = _config(path, dedup_keys=("id",))
with FileBackedQueueAdapter(config=config) as queue:
assert queue.size() == 1
assert queue.dequeue_batch(max_items=10) == [{"id": 1}]
def test_connect_skips_missing_or_invalid_age_on_replay(tmp_path: Path) -> None:
path = tmp_path / "items.jsonl"
now = datetime.now(UTC)
path.write_text(
"\n".join(
[
'{"id": 1}',
'{"id": 2, "created_at": "not-a-date"}',
f'{{"id": 3, "created_at": "{now.isoformat()}"}}',
]
)
+ "\n",
encoding="utf-8",
)
config = _config(path, dedup_keys=("id",), age_key="created_at")
with FileBackedQueueAdapter(config=config) as queue:
assert queue.size() == 1
assert queue.dequeue_batch(max_items=10)[0]["id"] == 3
def test_connect_is_idempotent(tmp_path: Path) -> None:
path = tmp_path / "items.jsonl"
queue = FileBackedQueueAdapter(config=_config(path))
queue.connect()
queue.enqueue([{"id": 1}])
queue.connect()
assert queue.size() == 1
queue.disconnect()
assert not queue.is_connected()
def test_evict_compacts_file(tmp_path: Path) -> None:
path = tmp_path / "items.jsonl"
config = _config(path, dedup_keys=("id",), age_key="created_at", max_age_hours=24)
now = datetime.now(UTC)
with FileBackedQueueAdapter(config=config) as queue:
queue.enqueue(
[
{"id": 1, "created_at": now - timedelta(hours=2)},
{"id": 2, "created_at": now - timedelta(minutes=5)},
]
)
removed = queue.evict_older_than(now - timedelta(hours=1))
assert removed == 1
assert queue.size() == 1
lines = [line for line in path.read_text(encoding="utf-8").splitlines() if line]
assert len(lines) == 1
def test_auto_evict_on_enqueue(tmp_path: Path) -> None:
path = tmp_path / "items.jsonl"
config = _config(path, dedup_keys=("id",), age_key="created_at", max_age_hours=1)
now = datetime.now(UTC)
with FileBackedQueueAdapter(config=config) as queue:
queue.enqueue([{"id": 1, "created_at": now - timedelta(hours=2)}])
# Item is enqueued then immediately age-evicted.
assert queue.size() == 0
def test_from_env_when_config_omitted(
tmp_path: Path, monkeypatch: pytest.MonkeyPatch
) -> None:
path = tmp_path / "from-env.jsonl"
monkeypatch.setenv("FILE_QUEUE_PATH", str(path))
with FileBackedQueueAdapter() as queue:
queue.enqueue([{"id": 1}])
assert queue.size() == 1
def test_injected_memory(tmp_path: Path) -> None:
path = tmp_path / "items.jsonl"
memory = MemoryQueueAdapter(dedup_keys=("id",))
config = _config(path, dedup_keys=("id",))
with FileBackedQueueAdapter(config=config, memory=memory) as queue:
queue.enqueue([{"id": 1}])
assert memory.size() == 1
def test_empty_enqueue_and_dequeue_skip_compact(tmp_path: Path) -> None:
path = tmp_path / "items.jsonl"
with FileBackedQueueAdapter(config=_config(path)) as queue:
queue.enqueue([])
assert queue.dequeue_batch(max_items=10) == []
assert queue.evict_older_than(datetime.now(UTC)) == 0
assert not path.exists() or path.read_text(encoding="utf-8") == ""
def test_reconnect_replays_after_disconnect(tmp_path: Path) -> None:
path = tmp_path / "items.jsonl"
queue = FileBackedQueueAdapter(config=_config(path, dedup_keys=("id",)))
queue.connect()
queue.enqueue([{"id": 1}])
queue.disconnect()
queue.connect()
assert queue.size() == 1
queue.disconnect()
def test_json_default_rejects_unsupported() -> None:
from python_repositories.adapters.file_backed_queue_adapter import _json_default
with pytest.raises(TypeError):
_json_default(object())
def test_disconnect_ignores_fsync_oserror(
tmp_path: Path, monkeypatch: pytest.MonkeyPatch
) -> None:
path = tmp_path / "items.jsonl"
queue = FileBackedQueueAdapter(config=_config(path))
queue.connect()
def boom(_fd: int) -> None:
raise OSError("fsync failed")
monkeypatch.setattr(
"python_repositories.adapters.file_backed_queue_adapter.os.fsync",
boom,
)
queue.disconnect()
assert not queue.is_connected()
def test_append_items_requires_open_handle(tmp_path: Path) -> None:
path = tmp_path / "items.jsonl"
queue = FileBackedQueueAdapter(config=_config(path))
queue.connect()
queue._append_handle = None # noqa: SLF001
with pytest.raises(RuntimeError, match="append handle is not open"):
queue._append_items([{"id": 1}]) # noqa: SLF001
queue._connected = False # noqa: SLF001
+56
View File
@@ -0,0 +1,56 @@
"""Unit tests for FileQueueConfig."""
from __future__ import annotations
from pathlib import Path
import pytest
from python_repositories.config import FileQueueConfig
def test_defaults() -> None:
config = FileQueueConfig(path=Path("/tmp/queue.jsonl"))
assert config.path == Path("/tmp/queue.jsonl")
assert config.max_age_hours == 24
assert config.dedup_keys == ()
assert config.age_key is None
def test_from_env_loads_path(monkeypatch: pytest.MonkeyPatch) -> None:
monkeypatch.setenv("FILE_QUEUE_PATH", "/data/items.jsonl")
monkeypatch.delenv("FILE_QUEUE_MAX_AGE_HOURS", raising=False)
config = FileQueueConfig.from_env(use_dotenv=False)
assert config.path == Path("/data/items.jsonl")
assert config.max_age_hours == 24
def test_from_env_loads_max_age_hours(monkeypatch: pytest.MonkeyPatch) -> None:
monkeypatch.setenv("FILE_QUEUE_PATH", "/data/items.jsonl")
monkeypatch.setenv("FILE_QUEUE_MAX_AGE_HOURS", "48")
config = FileQueueConfig.from_env(use_dotenv=False)
assert config.max_age_hours == 48
def test_from_env_raises_when_path_missing(monkeypatch: pytest.MonkeyPatch) -> None:
monkeypatch.delenv("FILE_QUEUE_PATH", raising=False)
with pytest.raises(Exception):
FileQueueConfig.from_env(use_dotenv=False)
def test_from_env_respects_custom_env_var_names(
monkeypatch: pytest.MonkeyPatch,
) -> None:
monkeypatch.setenv("CUSTOM_QUEUE_PATH", "/custom/q.jsonl")
monkeypatch.setenv("CUSTOM_MAX_AGE", "12")
config = FileQueueConfig.from_env(
"CUSTOM_QUEUE_PATH",
max_age_hours_env_var_name="CUSTOM_MAX_AGE",
dedup_keys=("id",),
age_key="created_at",
use_dotenv=False,
)
assert config.path == Path("/custom/q.jsonl")
assert config.max_age_hours == 12
assert config.dedup_keys == ("id",)
assert config.age_key == "created_at"
+144
View File
@@ -0,0 +1,144 @@
"""Unit tests for MemoryQueueAdapter."""
from __future__ import annotations
from datetime import UTC, datetime, timedelta
import threading
import pytest
from python_repositories.adapters.memory_queue_adapter import (
MemoryQueueAdapter,
_parse_age,
)
from python_repositories.interfaces.queue_repository_interface import (
QueueRepositoryInterface,
)
def test_implements_interface() -> None:
assert issubclass(MemoryQueueAdapter, QueueRepositoryInterface)
def test_fifo_without_dedup() -> None:
queue = MemoryQueueAdapter()
queue.enqueue([{"id": 1}, {"id": 2}, {"id": 3}])
assert queue.size() == 3
assert queue.dequeue_batch(max_items=2) == [{"id": 1}, {"id": 2}]
assert queue.dequeue_batch(max_items=10) == [{"id": 3}]
assert queue.dequeue_batch() == []
assert queue.size() == 0
def test_empty_enqueue_is_noop() -> None:
queue = MemoryQueueAdapter()
queue.enqueue([])
assert queue.size() == 0
assert queue.enqueue_and_return_added([]) == []
def test_dedup_keeps_first() -> None:
queue = MemoryQueueAdapter(dedup_keys=("id",))
queue.enqueue([{"id": 1, "v": "a"}, {"id": 1, "v": "b"}, {"id": 2, "v": "c"}])
assert queue.size() == 2
assert queue.dequeue_batch(max_items=10) == [
{"id": 1, "v": "a"},
{"id": 2, "v": "c"},
]
def test_missing_dedup_key_raises() -> None:
queue = MemoryQueueAdapter(dedup_keys=("id",))
with pytest.raises(ValueError, match="missing dedup key"):
queue.enqueue([{"name": "x"}])
def test_missing_age_key_raises() -> None:
queue = MemoryQueueAdapter(age_key="created_at")
with pytest.raises(ValueError, match="missing age key"):
queue.enqueue([{"id": 1}])
def test_evict_older_than() -> None:
queue = MemoryQueueAdapter(age_key="created_at", dedup_keys=("id",))
now = datetime.now(UTC)
queue.enqueue(
[
{"id": 1, "created_at": now - timedelta(hours=2)},
{"id": 2, "created_at": now - timedelta(minutes=30)},
{"id": 3, "created_at": (now - timedelta(hours=3)).isoformat()},
]
)
removed = queue.evict_older_than(now - timedelta(hours=1))
assert removed == 2
assert queue.size() == 1
remaining = queue.dequeue_batch(max_items=10)
assert remaining[0]["id"] == 2
def test_evict_without_age_key_is_noop() -> None:
queue = MemoryQueueAdapter()
queue.enqueue([{"id": 1}])
assert queue.evict_older_than(datetime.now(UTC)) == 0
assert queue.size() == 1
def test_parse_age_rejects_unsupported_type() -> None:
with pytest.raises(ValueError, match="age value must be"):
_parse_age(123)
def test_parse_age_aware_datetime() -> None:
from datetime import timezone
eastern = timezone(timedelta(hours=-5))
aware = datetime(2024, 1, 1, 12, 0, 0, tzinfo=eastern)
assert _parse_age(aware) == datetime(2024, 1, 1, 17, 0, 0, tzinfo=UTC)
def test_parse_age_naive_datetime() -> None:
naive = datetime(2024, 1, 1, 12, 0, 0)
assert _parse_age(naive) == datetime(2024, 1, 1, 12, 0, 0, tzinfo=UTC)
def test_parse_age_zulu_string() -> None:
assert _parse_age("2024-01-01T00:00:00Z") == datetime(2024, 1, 1, 0, 0, 0, tzinfo=UTC)
def test_parse_age_naive_string() -> None:
assert _parse_age("2024-01-01T00:00:00") == datetime(2024, 1, 1, 0, 0, 0, tzinfo=UTC)
def test_dequeue_rejects_negative_max_items() -> None:
queue = MemoryQueueAdapter()
with pytest.raises(ValueError, match="max_items"):
queue.dequeue_batch(max_items=-1)
def test_clear_and_snapshot() -> None:
queue = MemoryQueueAdapter()
queue.enqueue([{"id": 1}, {"id": 2}])
assert queue.snapshot() == [{"id": 1}, {"id": 2}]
queue.clear()
assert queue.size() == 0
assert queue.snapshot() == []
def test_thread_safety_smoke() -> None:
queue = MemoryQueueAdapter(dedup_keys=("id",))
errors: list[BaseException] = []
def worker(start: int) -> None:
try:
for i in range(start, start + 50):
queue.enqueue([{"id": i}])
except BaseException as exc: # noqa: BLE001
errors.append(exc)
threads = [threading.Thread(target=worker, args=(i * 50,)) for i in range(4)]
for thread in threads:
thread.start()
for thread in threads:
thread.join()
assert errors == []
assert queue.size() == 200
@@ -0,0 +1,100 @@
"""Unit tests for QueueRepositoryInterface."""
from datetime import UTC, datetime
from typing import Any
import pytest
from python_repositories.interfaces.queue_repository_interface import (
QueueRepositoryInterface,
)
class InMemoryQueueRepo:
"""Plain class that satisfies QueueRepositoryInterface without inheritance."""
def enqueue(self, items: list[dict[str, Any]]) -> None:
pass
def dequeue_batch(self, *, max_items: int = 1000) -> list[dict[str, Any]]:
del max_items
return []
def size(self) -> int:
return 0
def evict_older_than(self, cutoff: datetime) -> int:
del cutoff
return 0
def accepts_queue_repo(repo: QueueRepositoryInterface) -> None:
"""Type-checking hook for QueueRepositoryInterface structural subtyping."""
repo.size()
def test_instantiation_fails_when_enqueue_not_implemented() -> None:
class Incomplete(QueueRepositoryInterface):
def dequeue_batch(self, *, max_items: int = 1000) -> list[dict[str, Any]]:
return []
def size(self) -> int:
return 0
def evict_older_than(self, cutoff: datetime) -> int:
return 0
with pytest.raises(TypeError):
_ = Incomplete() # type: ignore
def test_instantiation_fails_when_dequeue_batch_not_implemented() -> None:
class Incomplete(QueueRepositoryInterface):
def enqueue(self, items: list[dict[str, Any]]) -> None:
pass
def size(self) -> int:
return 0
def evict_older_than(self, cutoff: datetime) -> int:
return 0
with pytest.raises(TypeError):
_ = Incomplete() # type: ignore
def test_instantiation_fails_when_size_not_implemented() -> None:
class Incomplete(QueueRepositoryInterface):
def enqueue(self, items: list[dict[str, Any]]) -> None:
pass
def dequeue_batch(self, *, max_items: int = 1000) -> list[dict[str, Any]]:
return []
def evict_older_than(self, cutoff: datetime) -> int:
return 0
with pytest.raises(TypeError):
_ = Incomplete() # type: ignore
def test_instantiation_fails_when_evict_older_than_not_implemented() -> None:
class Incomplete(QueueRepositoryInterface):
def enqueue(self, items: list[dict[str, Any]]) -> None:
pass
def dequeue_batch(self, *, max_items: int = 1000) -> list[dict[str, Any]]:
return []
def size(self) -> int:
return 0
with pytest.raises(TypeError):
_ = Incomplete() # type: ignore
def test_structural_subtyping() -> None:
repo: QueueRepositoryInterface = InMemoryQueueRepo()
accepts_queue_repo(repo)
assert isinstance(repo, QueueRepositoryInterface)
assert repo.evict_older_than(datetime.now(UTC)) == 0