feat(import): adopt files that are already in a library

tests/test_import_flow.py

@@ -1294,3 +1294,37 @@ def test_bulk_move_delete_and_count(client):
                           json={"ids": ids[1:]}).json() == {"deleted": 2}
     assert client.get("/api/requests/count?status=missing").json() == {"count": 1}
     assert [r["event"] for r in client.get("/api/history").json()] == ["deleted"] * 2
+
+
+def test_adopt_files_already_in_the_library(client):
+    root = client.tmp_path / "audio"
+    lib = client.post("/api/libraries", json={
+        "name": "Audio", "media_type": "audiobook", "root_path": str(root)}).json()["id"]
+    hobbit = client.post("/api/requests", json={
+        "library_id": lib, "title": "Der Hobbit", "authors": "Tolkien"}).json()["id"]
+    known = client.post("/api/requests", json={
+        "library_id": lib, "title": "Momo", "authors": "Ende"}).json()["id"]
+    for folder in ("Tolkien/Der Hobbit", "Ende/Momo"):
+        (root / folder).mkdir(parents=True)
+        for i in (1, 2):
+            (root / folder / f"{i:02d}.mp3").write_bytes(b"")
+    (root / "Tolkien" / "Notizen.epub").write_bytes(b"")  # other media type
+
+    # Momo already points at its folder, so only the Hobbit is offered
+    adopted = client.post(f"/api/import/library/{lib}", json={"items": [
+        {"path": str(root / "Ende/Momo"), "request_id": known}]}).json()
+    assert adopted == {"results": [{"path": str(root / "Ende/Momo"), "ok": True}]}
+    items = client.get(f"/api/import/library/{lib}").json()["items"]
+    assert [(i["name"], i["suggested_request_id"]) for i in items] == [("Der Hobbit", hobbit)]
+
+    resp = client.post(f"/api/import/library/{lib}", json={"items": [
+        {"path": items[0]["path"], "request_id": hobbit},
+        {"path": str(root / "Ende/Momo"), "request_id": known},  # no longer open
+        {"path": str(client.downloads), "request_id": hobbit},
+    ]}).json()["results"]
+    assert [r["ok"] for r in resp] == [True, False, False]
+    reqs = {r["id"]: r for r in client.get("/api/requests?status=imported").json()}
+    assert reqs[hobbit]["imported_path"] == str(root / "Tolkien/Der Hobbit")
+    assert (root / "Tolkien/Der Hobbit/01.mp3").exists()  # nothing moved
+    assert client.get(f"/api/import/library/{lib}").json()["items"] == []
+    assert [h["event"] for h in client.get("/api/history").json()] == ["adopted"] * 2

wordarr/api/imports.py

@@ -6,8 +6,8 @@ from sqlalchemy import select
 from sqlalchemy.orm import Session
 
 from .. import config, connect
-from ..db import BookRequest, get_session, record
+from ..db import BookRequest, Library, get_session, record
-from ..fsnames import decode_path, readable
+from ..fsnames import decode_path, is_broken, readable
 from ..importer import matcher, mover, scanner
 
 router = APIRouter(prefix="/api/import", tags=["import"])
@@ -128,3 +128,84 @@ def do_import(data: ImportIn, background: BackgroundTasks,
         raise HTTPException(400, "nothing to import")
     connect.schedule(background, rows)
     return {"results": results}
+
+
+def _library(session: Session, library_id: int) -> Library:
+    lib = session.get(Library, library_id)
+    if not lib:
+        raise HTTPException(404, "library not found")
+    return lib
+
+
+def _open_for(session: Session, lib: Library) -> list[BookRequest]:
+    missing = session.scalars(select(BookRequest).where(
+        BookRequest.status == "missing", BookRequest.library_id == lib.id)).all()
+    return list(missing) + [r for r in _orphaned(session) if r.library_id == lib.id]
+
+
+@router.get("/library/{library_id}")
+def scan_library(library_id: int, session: Session = Depends(get_session)):
+    """Files already in the library that no request points at, matched against
+    the library's open requests, so an existing collection need not be imported
+    again."""
+    lib = _library(session, library_id)
+    imported = session.scalars(select(BookRequest.imported_path).where(
+        BookRequest.status == "imported", BookRequest.imported_path != "")).all()
+    covered = {Path(p) for p in imported}
+    # a folder holding an imported file, or a file inside an imported folder
+    above = {a for p in covered for a in p.parents}
+
+    def known(item: dict) -> bool:
+        p = Path(decode_path(item["path"]))
+        return p in covered or p in above or any(a in covered for a in p.parents)
+
+    items = [i for i in scanner.scan(Path(lib.root_path))
+             if i["media_type"] == lib.media_type and not known(i)]
+    open_reqs = _open_for(session, lib)
+    return {"root_path": lib.root_path,
+            "items": matcher.best_matches(items, open_reqs),
+            "requests": [{"id": r.id, "title": r.title, "authors": r.authors}
+                         for r in open_reqs]}
+
+
+class AdoptItem(BaseModel):
+    path: str
+    request_id: int
+
+
+class AdoptIn(BaseModel):
+    items: list[AdoptItem]
+
+
+@router.post("/library/{library_id}")
+def adopt(library_id: int, data: AdoptIn, background: BackgroundTasks,
+          session: Session = Depends(get_session)):
+    """Point open requests at files that are already in place. Nothing moves."""
+    lib = _library(session, library_id)
+    root = Path(lib.root_path).resolve()
+    open_ids = {r.id for r in _open_for(session, lib)}
+    results, rows = [], []
+    for item in data.items:
+        path = decode_path(item.path)
+        shown = readable(path)
+        req = session.get(BookRequest, item.request_id)
+        if is_broken(path):
+            error = "file name is not valid UTF-8, rename it first"
+        elif root not in Path(path).resolve().parents:
+            error = "path outside the library"
+        elif not Path(path).exists():
+            error = "path no longer exists"
+        elif not req or req.id not in open_ids:
+            error = f"request #{item.request_id} is not open in this library"
+        else:
+            error = ""
+        if error:
+            results.append({"path": shown, "ok": False, "error": error})
+            continue
+        req.status, req.imported_path = "imported", path
+        rows.append(record(session, "adopted", req, dest=path))
+        open_ids.discard(req.id)  # two files for one request: the first wins
+        results.append({"path": shown, "ok": True})
+    session.commit()
+    connect.schedule(background, rows)
+    return {"results": results}