ReMe/tests/unit/test_wikilink_utils.py

"""Tests for wikilink helpers in ``reme.utils.wikilink_handler``.

Two pure async helpers used by file_move / file_delete:

  * ``retarget_links(src, dst, scope?, dry_run?)`` — rewrite wikilink
    targets across the workspace, using the file_graph's reverse index to
    find inbound sources (no fs scan).
  * ``find_inbound(target, scope?)`` — report inbound count without
    rewriting.

Retarget only matches the literal full-path form ``[[topics/x.md]]``;
short ``[[x]]`` and no-ext ``[[topics/x]]`` are left alone by design.
"""

# pylint: disable=protected-access

from contextlib import chdir as temp_chdir
import tempfile
import warnings
from pathlib import Path

import pytest

from reme.components.file_store import LocalFileStore
from reme.schema import FileNode
from reme.utils.wikilink_handler import WikilinkHandler

warnings.filterwarnings("ignore", category=DeprecationWarning, module="jieba")
warnings.filterwarnings("ignore", category=DeprecationWarning, module="pkg_resources")


async def _store_with(files: dict[str, str]) -> LocalFileStore:
    """LocalFileStore seeded with (rel → content) files: written to disk
    AND registered in the file_graph with wikilinks parsed from the body.

    Without the parsed links the reverse-index lookup yields nothing and
    retarget becomes a no-op.
    """
    store = LocalFileStore(name="t", embedding_store="")
    await store.start()
    nodes: list[FileNode] = []
    root = Path.cwd()
    for rel, content in files.items():
        abs_path = root / rel
        abs_path.parent.mkdir(parents=True, exist_ok=True)
        abs_path.write_text(content, encoding="utf-8")
        nodes.append(
            FileNode(
                path=rel,
                st_mtime=abs_path.stat().st_mtime,
                links=WikilinkHandler.extract_links(content, rel),
            ),
        )
    if nodes:
        await store.file_graph.upsert_nodes(nodes)
    return store


async def _empty_store() -> LocalFileStore:
    store = LocalFileStore(name="t", embedding_store="")
    await store.start()
    return store


@pytest.mark.asyncio
async def test_retarget_exact_full_path_match():
    """[[topics/Alice.md]] → [[people/Alice.md]]; non-matching links untouched."""

    with tempfile.TemporaryDirectory() as tmp, temp_chdir(tmp):
        root = Path(tmp)
        store = await _store_with(
            {
                "note.md": "See [[topics/Alice.md]] and [[topics/Bob.md]].",
                "people/Alice.md": "# Alice",
            },
        )
        payload = await WikilinkHandler.retarget_links(store, src="topics/Alice.md", dst="people/Alice.md")
        assert "error" not in payload
        assert payload["links_changed"] == 1
        assert payload["files_touched"] == 1
        body = (root / "note.md").read_text(encoding="utf-8")
        assert "[[people/Alice.md]]" in body
        assert "[[topics/Alice.md]]" not in body
        assert "[[topics/Bob.md]]" in body
        await store.close()


@pytest.mark.asyncio
async def test_retarget_short_and_no_ext_forms_ignored():
    """Short-form [[Alice]] and no-ext [[topics/Alice]] are NOT matched."""

    with tempfile.TemporaryDirectory() as tmp, temp_chdir(tmp):
        root = Path(tmp)
        body_in = "See [[Alice]] and [[topics/Alice]] but also [[topics/Alice.md]]."
        store = await _store_with(
            {
                "note.md": body_in,
                "people/Alice.md": "# Alice",
            },
        )
        payload = await WikilinkHandler.retarget_links(store, src="topics/Alice.md", dst="people/Alice.md")
        assert payload["links_changed"] == 1  # only the full-path form
        body = (root / "note.md").read_text(encoding="utf-8")
        assert "[[people/Alice.md]]" in body
        assert "[[Alice]]" in body  # short form untouched
        assert "[[topics/Alice]]" in body  # no-ext form untouched
        await store.close()


@pytest.mark.asyncio
async def test_retarget_anchor_preserved():
    """`[[topics/Alice.md#intro]]` → `[[people/Alice.md#intro]]` (anchor kept verbatim)."""

    with tempfile.TemporaryDirectory() as tmp, temp_chdir(tmp):
        root = Path(tmp)
        store = await _store_with({"note.md": "Jump to [[topics/Alice.md#intro]] please."})
        payload = await WikilinkHandler.retarget_links(store, src="topics/Alice.md", dst="people/Alice.md")
        assert payload["links_changed"] == 1
        body = (root / "note.md").read_text(encoding="utf-8")
        assert "[[people/Alice.md#intro]]" in body
        await store.close()


@pytest.mark.asyncio
async def test_retarget_alias_preserved():
    """`[[topics/Alice.md|Display Name]]` → `[[people/Alice.md|Display Name]]`."""

    with tempfile.TemporaryDirectory() as tmp, temp_chdir(tmp):
        root = Path(tmp)
        store = await _store_with({"note.md": "Meet [[topics/Alice.md|Alice the Architect]]."})
        payload = await WikilinkHandler.retarget_links(store, src="topics/Alice.md", dst="people/Alice.md")
        assert payload["links_changed"] == 1
        body = (root / "note.md").read_text(encoding="utf-8")
        assert "[[people/Alice.md|Alice the Architect]]" in body
        await store.close()


@pytest.mark.asyncio
async def test_retarget_anchor_and_alias_together():
    """`[[A.md#h|disp]]` → `[[B.md#h|disp]]` keeps both suffixes."""

    with tempfile.TemporaryDirectory() as tmp, temp_chdir(tmp):
        root = Path(tmp)
        store = await _store_with({"note.md": "See [[topics/Alice.md#bio|her bio]] now."})
        payload = await WikilinkHandler.retarget_links(store, src="topics/Alice.md", dst="people/Alice.md")
        assert payload["links_changed"] == 1
        body = (root / "note.md").read_text(encoding="utf-8")
        assert "[[people/Alice.md#bio|her bio]]" in body
        await store.close()


@pytest.mark.asyncio
async def test_retarget_image_marker_preserved():
    """`![[topics/diagram.md]]` (embed) keeps its `!` prefix on rewrite."""

    with tempfile.TemporaryDirectory() as tmp, temp_chdir(tmp):
        root = Path(tmp)
        store = await _store_with({"note.md": "Inline embed: ![[topics/diagram.md]] here."})
        payload = await WikilinkHandler.retarget_links(store, src="topics/diagram.md", dst="diagrams/diagram.md")
        assert payload["links_changed"] == 1
        body = (root / "note.md").read_text(encoding="utf-8")
        assert "![[diagrams/diagram.md]]" in body
        await store.close()


@pytest.mark.asyncio
async def test_retarget_preserves_surrounding_relation_text():
    """Retargeting changes only the inner wikilink target."""

    with tempfile.TemporaryDirectory() as tmp, temp_chdir(tmp):
        root = Path(tmp)
        body_in = "colleague:: [[topics/Alice.md]]\n" + "She is the [负责:: [[topics/Alice.md]]] for the migration.\n"
        store = await _store_with({"note.md": body_in})
        payload = await WikilinkHandler.retarget_links(store, src="topics/Alice.md", dst="people/Alice.md")
        assert payload["links_changed"] == 2
        body = (root / "note.md").read_text(encoding="utf-8")
        assert "colleague:: [[people/Alice.md]]" in body
        assert "[负责:: [[people/Alice.md]]]" in body
        await store.close()


def test_markdown_links_are_ignored():
    """Only double-bracket wikilinks create graph edges or get rewritten."""
    markdown_links = (
        "[label](../wiki/example.md) "
        "[label](../wiki/example.md#section) "
        "[label](../wiki/example.md#L9-L10,L15-L20)"
    )
    original = f"{markdown_links} [[wiki/example.md#L9]]"

    rewritten, count = WikilinkHandler.scan_and_rewrite(
        original,
        old="wiki/example.md",
        new="archive/example.md",
    )

    assert [(link.target_path, link.target_anchor) for link in WikilinkHandler.extract_links(original, "note.md")] == [
        ("wiki/example.md", "L9"),
    ]
    assert count == 1
    assert rewritten == f"{markdown_links} [[archive/example.md#L9]]"


@pytest.mark.asyncio
async def test_retarget_multiple_files_aggregate_counts():
    """links_changed sums across files; by_file lists per-file counts."""

    with tempfile.TemporaryDirectory() as tmp, temp_chdir(tmp):
        store = await _store_with(
            {
                "a.md": "[[topics/Alice.md]] then [[topics/Alice.md]]",
                "sub/b.md": "[[topics/Alice.md]] and [[topics/Bob.md]]",
            },
        )
        payload = await WikilinkHandler.retarget_links(store, src="topics/Alice.md", dst="people/Alice.md")
        assert payload["links_changed"] == 3
        assert payload["files_touched"] == 2
        by_file = {row["path"]: row["count"] for row in payload["by_file"]}
        assert by_file == {"a.md": 2, "sub/b.md": 1}
        await store.close()


@pytest.mark.asyncio
async def test_retarget_dry_run_does_not_write():
    """dry_run=True reports counts but leaves files on disk unchanged."""

    with tempfile.TemporaryDirectory() as tmp, temp_chdir(tmp):
        root = Path(tmp)
        original = "See [[topics/Alice.md]]."
        store = await _store_with({"note.md": original})
        payload = await WikilinkHandler.retarget_links(
            store,
            src="topics/Alice.md",
            dst="people/Alice.md",
            dry_run=True,
        )
        assert payload["dry_run"] is True
        assert payload["links_changed"] == 1
        assert payload["files_touched"] == 1
        assert (root / "note.md").read_text(encoding="utf-8") == original
        await store.close()


@pytest.mark.asyncio
async def test_retarget_scope_limits_sweep():
    """Files outside `scope` are not visited even if they contain matches."""

    with tempfile.TemporaryDirectory() as tmp, temp_chdir(tmp):
        root = Path(tmp)
        store = await _store_with(
            {
                "in_scope/note.md": "See [[topics/Alice.md]].",
                "outside/note.md": "Also [[topics/Alice.md]].",
            },
        )
        payload = await WikilinkHandler.retarget_links(
            store,
            src="topics/Alice.md",
            dst="people/Alice.md",
            scope="in_scope",
        )
        assert payload["links_changed"] == 1
        assert payload["files_touched"] == 1
        in_body = (root / "in_scope/note.md").read_text(encoding="utf-8")
        out_body = (root / "outside/note.md").read_text(encoding="utf-8")
        assert "[[people/Alice.md]]" in in_body
        assert "[[topics/Alice.md]]" in out_body  # untouched
        await store.close()


@pytest.mark.asyncio
async def test_retarget_src_eq_dst_is_noop():
    """src == dst: no scan needed, returns zero counts."""

    with tempfile.TemporaryDirectory() as tmp, temp_chdir(tmp):
        store = await _store_with({"note.md": "See [[topics/Alice.md]]."})
        payload = await WikilinkHandler.retarget_links(store, src="topics/Alice.md", dst="topics/Alice.md")
        assert payload["links_changed"] == 0
        assert payload["files_touched"] == 0
        await store.close()


@pytest.mark.asyncio
async def test_retarget_empty_src_or_dst_errors():
    """Empty src or dst produces an error payload."""

    with tempfile.TemporaryDirectory() as tmp, temp_chdir(tmp):
        store = await _empty_store()
        for kwargs in ({"src": "", "dst": "people/Alice.md"}, {"src": "topics/Alice.md", "dst": ""}):
            payload = await WikilinkHandler.retarget_links(store, **kwargs)
            assert "error" in payload, f"expected error for {kwargs}"
        await store.close()


@pytest.mark.asyncio
async def test_retarget_dst_with_forbidden_chars_errors():
    """``dst`` containing ``[ ] # |`` or newline lands as an error payload."""

    with tempfile.TemporaryDirectory() as tmp, temp_chdir(tmp):
        store = await _empty_store()
        for bad in ["people/Alice#anchor", "people/Alice|alias", "p/[A].md", "p/A]]"]:
            payload = await WikilinkHandler.retarget_links(store, src="topics/Alice.md", dst=bad)
            assert "error" in payload, f"expected error for dst={bad!r}, got {payload}"
        await store.close()


@pytest.mark.asyncio
async def test_retarget_absolute_path_rejected():
    """Absolute paths in src or dst are rejected (must be relative to the workspace)."""

    with tempfile.TemporaryDirectory() as tmp, temp_chdir(tmp):
        store = await _empty_store()
        for kwargs in (
            {"src": "/abs/old.md", "dst": "people/Alice.md"},
            {"src": "topics/Alice.md", "dst": "/abs/new.md"},
        ):
            payload = await WikilinkHandler.retarget_links(store, **kwargs)
            assert "error" in payload, f"expected error for {kwargs}"
        await store.close()


@pytest.mark.asyncio
async def test_find_inbound_counts_references():
    """find_inbound reports per-file counts and excludes self-references."""

    with tempfile.TemporaryDirectory() as tmp, temp_chdir(tmp):
        store = await _store_with(
            {
                "a.md": "[[topics/Alice.md]] and [[topics/Alice.md]] again",
                "b.md": "see [[topics/Alice.md]]",
                "topics/Alice.md": "self-ref [[topics/Alice.md]] should not count",
            },
        )
        payload = await WikilinkHandler.find_inbound(store, target="topics/Alice.md")
        assert payload["files_touched"] == 2
        assert payload["links_total"] == 3
        by_file = {row["path"]: row["count"] for row in payload["by_file"]}
        assert by_file == {"a.md": 2, "b.md": 1}
        await store.close()