python-repositories/tests/unit/file_backed_queue_adapter_test.py

"""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