FabledCurator/tests/test_library_layout.py

"""Milestone #421 — the shared predicate behind the consolidation.

`destination_for` is pure and tested without a database; `survey_layout` gets
the integration treatment because the counts are the number the apply is
checked against.
"""

from pathlib import Path

import pytest
from sqlalchemy import select

from backend.app.models import Artist, ImageRecord
from backend.app.services.library_layout import (
    RESERVED_TOP_LEVEL,
    _misplaced_conditions,
    canonical_dir,
    destination_for,
    survey_layout,
)

ROOT = Path("/images")


# --- destination_for (pure) -------------------------------------------------


def test_destination_rewrites_only_the_artist_segment():
    assert destination_for(
        "/images/Conto/patreon/2026-01_a_Post/x.png", ROOT, "conto"
    ) == Path("/images/conto/patreon/2026-01_a_Post/x.png")


def test_destination_is_identity_for_a_row_already_in_place():
    p = "/images/conto/patreon/x.png"
    assert destination_for(p, ROOT, "conto") == Path(p)


def test_destination_pulls_a_root_level_row_under_its_artist():
    """Diverges from canonical_subdir deliberately: the row CARRIES an
    artist_id, so a file at the root is an anomaly with a known home."""
    assert destination_for("/images/loose.png", ROOT, "conto") == Path(
        "/images/conto/loose.png"
    )


def test_destination_refuses_paths_outside_the_images_root():
    assert destination_for("/srv/elsewhere/x.png", ROOT, "conto") is None


@pytest.mark.parametrize("reserved", sorted(RESERVED_TOP_LEVEL))
def test_destination_refuses_the_reserved_stores(reserved):
    """Relocating these would move the thumbnail cache, the attachment blobs
    or the credential key into an artist folder."""
    assert destination_for(f"/images/{reserved}/aa/x.png", ROOT, "conto") is None


def test_destination_is_idempotent():
    once = destination_for("/images/Conto/patreon/x.png", ROOT, "conto")
    assert destination_for(str(once), ROOT, "conto") == once


# --- the predicate ----------------------------------------------------------


def test_canonical_prefix_carries_a_separator():
    """Without the trailing slash, artist `ara` matches every path under
    `arbuzbudesh/` — one artist reads as fully placed while another's rows
    are silently skipped."""
    conds = _misplaced_conditions(ROOT, 1, "ara")
    rendered = str(conds[-1].compile(compile_kwargs={"literal_binds": True}))
    assert "/images/ara/" in rendered


# --- survey_layout (integration) --------------------------------------------
#
# Marked per-test rather than with a module-level `pytestmark`: the
# destination_for cases above are pure and belong in the fast unit lane.


def _artist(db, name, slug):
    a = Artist(name=name, slug=slug)
    db.add(a)
    db.flush()
    return a


def _image(db, path, artist=None, n=0):
    rec = ImageRecord(
        path=path, sha256=f"{n:064d}", size_bytes=1, mime="image/png",
        width=10, height=10, origin="imported_filesystem",
        integrity_status="unknown",
        artist_id=artist.id if artist else None,
    )
    db.add(rec)
    db.flush()
    return rec


@pytest.mark.integration
def test_survey_splits_canonical_from_misplaced(db_sync):
    conto = _artist(db_sync, "Conto", "conto")
    _image(db_sync, "/images/conto/patreon/a.png", conto, 1)
    _image(db_sync, "/images/Conto/patreon/b.png", conto, 2)
    _image(db_sync, "/images/Conto/patreon/c.png", conto, 3)

    report = survey_layout(db_sync, ROOT, check_disk=False)

    assert report.misplaced_rows == 2
    assert report.canonical_rows == 1
    row = next(a for a in report.artists if a.slug == "conto")
    assert row.stray_dirs == ["Conto"]


@pytest.mark.integration
def test_survey_does_not_confuse_a_prefix_sharing_artist(db_sync):
    """`ara` vs `arbuzbudesh` — the reason the predicate anchors on a
    separator. Both are real artists in the operator's library."""
    ara = _artist(db_sync, "Ara", "ara")
    arbuz = _artist(db_sync, "ArbuzBudesh", "arbuzbudesh")
    _image(db_sync, "/images/ara/x.png", ara, 4)
    _image(db_sync, "/images/arbuzbudesh/y.png", arbuz, 5)

    report = survey_layout(db_sync, ROOT, check_disk=False)

    assert report.misplaced_rows == 0
    assert report.canonical_rows == 2


@pytest.mark.integration
def test_survey_counts_two_rows_landing_on_one_destination(db_sync):
    """A collision is the case the apply must refuse, so the report has to
    surface it rather than promise a move that cannot happen."""
    sticky = _artist(db_sync, "StickySpoodge", "stickyspoodge")
    _image(db_sync, "/images/StickySpoodge/p/dup.png", sticky, 6)
    _image(db_sync, "/images/Stickyspoodge/p/dup.png", sticky, 7)

    report = survey_layout(db_sync, ROOT, check_disk=False)

    assert report.collision_count == 1
    row = next(a for a in report.artists if a.slug == "stickyspoodge")
    assert row.collisions == ["/images/stickyspoodge/p/dup.png"]
    assert row.stray_dirs == ["StickySpoodge", "Stickyspoodge"]


@pytest.mark.integration
def test_survey_reports_unattributed_rows_without_moving_them(db_sync):
    """The 660 loose root files have no artist_id, so no predicate reaches
    them. They are counted, and left for task #4247."""
    _image(db_sync, "/images/orphan.png", None, 8)

    report = survey_layout(db_sync, ROOT, check_disk=False)

    assert report.unattributed_rows == 1
    assert report.misplaced_rows == 0


@pytest.mark.integration
def test_survey_refuses_a_row_under_a_reserved_store(db_sync):
    thumbs = _artist(db_sync, "Thumbsy", "thumbsy")
    _image(db_sync, "/images/thumbs/aa/weird.png", thumbs, 9)

    report = survey_layout(db_sync, ROOT, check_disk=False)

    assert report.unmovable == 1
    row = next(a for a in report.artists if a.slug == "thumbsy")
    assert row.misplaced_rows == 1
    assert row.collisions == []


@pytest.mark.integration
def test_survey_counts_a_missing_source_file(db_sync, tmp_path):
    """check_disk is what separates "would move" from "can move"."""
    gone = _artist(db_sync, "Gone", "gone")
    _image(db_sync, str(tmp_path / "Gone" / "missing.png"), gone, 10)

    report = survey_layout(db_sync, tmp_path, check_disk=True)

    assert report.missing_files == 1


@pytest.mark.integration
def test_survey_flags_a_destination_that_already_exists(db_sync, tmp_path):
    occupied = _artist(db_sync, "Occupied", "occupied")
    src = tmp_path / "Occupied" / "x.png"
    src.parent.mkdir(parents=True)
    src.write_bytes(b"src")
    dest = canonical_dir(tmp_path, "occupied") / "x.png"
    dest.parent.mkdir(parents=True)
    dest.write_bytes(b"already here")
    _image(db_sync, str(src), occupied, 11)

    report = survey_layout(db_sync, tmp_path, check_disk=True)

    assert report.collision_count == 1
    assert dest.read_bytes() == b"already here"  # read-only: nothing moved


@pytest.mark.integration
def test_survey_is_read_only(db_sync, tmp_path):
    a = _artist(db_sync, "Reader", "reader")
    src = tmp_path / "Reader" / "x.png"
    src.parent.mkdir(parents=True)
    src.write_bytes(b"x")
    rec = _image(db_sync, str(src), a, 12)
    before = rec.path

    survey_layout(db_sync, tmp_path, check_disk=True)

    db_sync.expire_all()
    assert db_sync.get(ImageRecord, rec.id).path == before
    assert src.exists()
    assert db_sync.execute(
        select(ImageRecord.path).where(ImageRecord.id == rec.id)
    ).scalar_one() == before