add(metadata): request a whole ebook series from hardcover

2026-09-01 17:20:34 +02:00
parent ffcfe24dfb
commit 169e84901c
6 changed files with 232 additions and 25 deletions
@@ -34,7 +34,8 @@ the same mount, otherwise moving turns into copy + delete.
 2. **Search** – look up a title (ebooks: Hardcover, Google Books and Open
   Library, audiobooks: Audible plus MusicBrainz, manga: AniList) and request
   it, or add it by hand. The language filter narrows ebook and audiobook hits. For a series,
-   *Ganze Serie…* requests every episode at once.
+   *Ganze Serie…* requests every part at once — audiobooks from Audible's keyword
+   crawl, ebooks from Hardcover's series listing (needs the token).
 3. **Import** – scan the download folder. wordarr suggests file→request matches,
   which you confirm or correct; entries can be merged, split, or turned into
   requests from their folder names. Every row names its format and size, and
@@ -239,7 +239,9 @@ $("#search-form").addEventListener("submit", async (e) => {
      )
      .join("");
    box.querySelectorAll("[data-series]").forEach((btn) =>
-      btn.addEventListener("click", () => openSeriesDialog(results[btn.dataset.series]))
+      btn.addEventListener("click", () =>
+        openSeriesDialog(results[btn.dataset.series], type, language)
+      )
    );
    box.querySelectorAll("[data-req]").forEach((btn) =>
      btn.addEventListener("click", async () => {
@@ -274,6 +276,7 @@ $("#search-form").addEventListener("submit", async (e) => {

 // ---- request a whole series ----
 let seriesEpisodes = [];
+let seriesType = "audiobook";

 function renderSeriesList() {
  const from = parseInt($("#series-from").value);
@@ -288,7 +291,9 @@ function renderSeriesList() {
      let gap = "";
      if (e.volume != null && previous != null && e.volume > previous + 1) {
        const missing = e.volume - previous - 1;
-        gap = `<div class="gap">… ${missing} Folge${missing > 1 ? "n" : ""} nicht bei Audible gefunden (${previous + 1}–${e.volume - 1})</div>`;
+        const ebook = seriesType === "ebook";
+        const unit = ebook ? (missing > 1 ? "Bände" : "Band") : `Folge${missing > 1 ? "n" : ""}`;
+        gap = `<div class="gap">… ${missing} ${unit} nicht ${ebook ? "bei Hardcover" : "bei Audible"} gefunden (${previous + 1}–${e.volume - 1})</div>`;
      }
      if (e.volume != null) previous = e.volume;
      return `${gap}<label>
@@ -314,20 +319,23 @@ function updateSeriesCount() {
  btn.disabled = n === 0;
 }

-async function openSeriesDialog(result) {
+async function openSeriesDialog(result, type = "audiobook", language = "") {
  const dlg = $("#series-dialog");
+  seriesType = type;
+  const unit = type === "ebook" ? "Bände" : "Folgen";
  $("#series-title").textContent = result.series || "Serie";
-  $("#series-info").innerHTML = '<span class="spinner"></span> Folgen werden gesucht…';
+  $("#series-info").innerHTML = `<span class="spinner"></span> ${unit} werden gesucht…`;
  $("#series-list").innerHTML = "";
  $("#series-from").value = "";
  $("#series-to").value = "";
-  $("#series-library").innerHTML = libOptions("audiobook") || "<option value=''>— keine Library —</option>";
+  $("#series-library").innerHTML = libOptions(type) || "<option value=''>— keine Library —</option>";
  $("#series-submit").disabled = true;
  dlg.showModal();
  try {
    seriesEpisodes = await api(
      `/api/search/series?series_id=${encodeURIComponent(result.series_id)}` +
-        `&title=${encodeURIComponent(result.series || result.title)}`
+        `&title=${encodeURIComponent(result.series || result.title)}` +
+        `&media_type=${type}&language=${encodeURIComponent(language)}`
    );
  } catch (err) {
    $("#series-info").textContent = err.message;
@@ -335,9 +343,11 @@ async function openSeriesDialog(result) {
  }
  const numbered = seriesEpisodes.filter((e) => e.volume != null);
  $("#series-info").textContent =
-    `${seriesEpisodes.length} Folgen gefunden` +
+    `${seriesEpisodes.length} ${unit} gefunden` +
    (numbered.length ? ` (Nr. ${numbered[0].volume}–${numbered[numbered.length - 1].volume})` : "") +
-    ". Audible hat keine Serien-Abfrage — einzelne Folgen können fehlen.";
+    (type === "ebook"
+      ? ". Bände ohne Ausgabe in der gewählten Sprache stehen mit ihrem Originaltitel da."
+      : ". Audible hat keine Serien-Abfrage — einzelne Folgen können fehlen.");
  renderSeriesList();
 }

@@ -356,7 +366,7 @@ $("#series-cancel").addEventListener("click", () => $("#series-dialog").close())

 $("#series-submit").addEventListener("click", async () => {
  const libId = $("#series-library").value;
-  if (!libId) { toast("Erst eine Audiobook-Library anlegen (Tab Libraries)", true); return; }
+  if (!libId) { toast(`Erst eine ${seriesType === "ebook" ? "Ebook" : "Audiobook"}-Library anlegen (Tab Libraries)`, true); return; }
  const items = seriesChecked().map((e) => ({
    title: e.title, authors: e.authors, narrator: e.narrator || "",
    external_id: e.external_id, year: e.year, series: e.series,
@@ -370,7 +380,7 @@ $("#series-submit").addEventListener("click", async () => {
      method: "POST",
      body: JSON.stringify({ library_id: parseInt(libId), items }),
    });
-    setLastLib("audiobook", libId);
+    setLastLib(seriesType, libId);
    toast(
      `${res.created.length} Anfragen angelegt` +
        (res.skipped ? `, ${res.skipped} bereits vorhanden` : "")
@@ -83,6 +83,80 @@ def test_hardcover_drops_books_without_an_edition_in_that_language(monkeypatch):
    assert asyncio.run(hardcover.search("hobbit", "german")) == []


+SERIES_ROWS = {
+    "name": "Die drei Fragezeichen",
+    "book_series": [
+        {"position": 1, "book": {
+            "id": 1, "title": "Super-Papagei", "release_year": 1968,
+            "contributions": [{"author": {"name": "Robert Arthur"}}],
+            "cached_image": {"url": "https://example.invalid/1.jpg"}}},
+        {"position": 2, "book": {"id": 2, "title": "Misteri Danau"}},
+        {"position": 2, "book": {"id": 3, "title": "Der Phantomsee"}},
+        {"position": 2, "book": {"id": 2, "title": "duplicate row"}},
+        {"position": 0, "book": {"id": 4, "title": "Sonderband"}},
+    ],
+}
+
+
+def series_client(monkeypatch, editions):
+    seen = []
+
+    def handler(request: httpx.Request) -> httpx.Response:
+        import json
+        body = json.loads(request.content)
+        seen.append(body["variables"])
+        if "editions" in body["query"]:
+            return httpx.Response(200, json={"data": {"editions": editions}})
+        return httpx.Response(200, json={"data": {"series": [SERIES_ROWS]}})
+
+    monkeypatch.setattr(hardcover, "TOKEN", "test-token")
+    mock_client(monkeypatch, hardcover, handler)
+    return seen
+
+
+def test_hardcover_series_sorts_by_volume_and_keeps_the_first_row(monkeypatch):
+    series_client(monkeypatch, [])
+    books = asyncio.run(hardcover.series_books("10146"))
+    assert [(b.volume, b.title) for b in books] == [
+        (1, "Super-Papagei"), (2, "Misteri Danau"), (None, "Sonderband"),
+    ]
+    first = books[0]
+    assert (first.series, first.series_id, first.authors, first.media_type) == (
+        "Die drei Fragezeichen", "10146", "Robert Arthur", "ebook")
+    assert first.cover_url == "https://example.invalid/1.jpg"
+
+
+def test_hardcover_series_prefers_the_row_with_an_edition_in_the_language(monkeypatch):
+    calls = series_client(monkeypatch, [
+        {"book_id": 3, "title": "Der Phantomsee", "isbn_13": "9783440032251",
+         "release_date": "1969-01-01", "cached_image": None},
+    ])
+    books = asyncio.run(hardcover.series_books("10146", "german"))
+    assert calls[1] == {"ids": [1, 2, 3, 4], "lang": "de"}
+    second = next(b for b in books if b.volume == 2)
+    assert (second.title, second.language, second.external_id, second.year) == (
+        "Der Phantomsee", "german", "9783440032251", 1969)
+    # a volume without a German edition keeps its original row instead of vanishing
+    assert [b.title for b in books] == ["Super-Papagei", "Der Phantomsee", "Sonderband"]
+
+
+def test_hardcover_series_survives_a_broken_edition_lookup(monkeypatch):
+    def handler(request: httpx.Request) -> httpx.Response:
+        import json
+        if "editions" in json.loads(request.content)["query"]:
+            return httpx.Response(500)
+        return httpx.Response(200, json={"data": {"series": [SERIES_ROWS]}})
+
+    monkeypatch.setattr(hardcover, "TOKEN", "test-token")
+    mock_client(monkeypatch, hardcover, handler)
+    assert len(asyncio.run(hardcover.series_books("10146", "german"))) == 3
+
+
+def test_hardcover_series_needs_a_numeric_id(monkeypatch):
+    monkeypatch.setattr(hardcover, "TOKEN", "test-token")
+    assert asyncio.run(hardcover.series_books("")) == []
+
+
 def test_google_books_restricts_the_language(monkeypatch):
    seen = []

@@ -1,6 +1,6 @@
 from fastapi import APIRouter, HTTPException

-from ..metadata import PROVIDERS, audible
+from ..metadata import PROVIDERS, SERIES_PROVIDERS
 from ..metadata.base import MetadataResult

 router = APIRouter(prefix="/api/search", tags=["search"])
@@ -18,13 +18,16 @@ async def search(media_type: str, q: str, language: str = "", fallback: bool = T


@router.get("/series", response_model=list[MetadataResult])
-async def series(series_id: str, title: str):
-    """All episodes of an Audible series, sorted by episode number. Audible has
-    no series endpoint, so this is a paged keyword crawl and may miss a few
-    episodes - those can still be added by hand."""
+async def series(series_id: str, title: str = "", media_type: str = "audiobook",
+                 language: str = ""):
+    """All parts of a series, by number. Audible has no series endpoint, so
+    audiobooks are a keyword crawl that may miss episodes; Hardcover is exact."""
+    lookup = SERIES_PROVIDERS.get(media_type)
+    if not lookup:
+        raise HTTPException(400, f"no series lookup for media_type: {media_type}")
    if not series_id:
        raise HTTPException(400, "series_id required")
    try:
-        return await audible.series_episodes(series_id, title)
+        return await lookup(series_id, title, language)
    except Exception as exc:
        raise HTTPException(502, f"metadata provider error: {exc}")
@@ -85,8 +85,22 @@ async def audiobook_search(query: str, language: str = "",
    return hits + [r for r in extra if (r.title.strip().lower(), r.volume) not in seen]


+async def _audiobook_series(series_id: str, title: str, language: str = ""):
+    return await audible.series_episodes(series_id, title)
+
+
+async def _ebook_series(series_id: str, title: str = "", language: str = ""):
+    return await hardcover.series_books(series_id, language)
+
+
 PROVIDERS = {
    "ebook": ebook_search,
    "audiobook": audiobook_search,
    "comic": anilist.search,
 }
+
+# a media type without an entry gets no "whole series" button
+SERIES_PROVIDERS = {
+    "audiobook": _audiobook_series,
+    "ebook": _ebook_series,
+}
@@ -8,6 +8,7 @@ The GraphQL search is Typesense backed and returns its hits as untyped JSON, so
 every field is read defensively. A book row describes the work, not one edition,
 so when a language is requested the matching editions are fetched in a second
 query and their title/ISBN/cover replace the ones of the original edition.
+Series are one query - unlike Audible, Hardcover knows them.
 """

 import logging
@@ -21,6 +22,7 @@ ENDPOINT = "https://api.hardcover.app/v1/graphql"
 TOKEN = os.environ.get("WORDARR_HARDCOVER_TOKEN", "").strip()
 LANGUAGES = {"german": "de", "english": "en"}
 RESULT_LIMIT = 10
+SERIES_LIMIT = 500  # books per series

 log = logging.getLogger(__name__)

@@ -49,6 +51,25 @@ query Editions($ids: [Int!]!, $lang: String!) {
 """


+_SERIES = """
+query Series($id: Int!, $limit: Int!) {
+  series(where: {id: {_eq: $id}}) {
+    name
+    book_series(order_by: {position: asc_nulls_last}, limit: $limit) {
+      position
+      book {
+        id
+        title
+        release_year
+        cached_image
+        contributions(limit: 3) { author { name } }
+      }
+    }
+  }
+}
+"""
+
+
 async def _post(client: httpx.AsyncClient, query: str, variables: dict) -> dict:
    resp = await client.post(
        ENDPOINT,
@@ -88,8 +109,27 @@ def _isbn(*candidates) -> str:
    return ""


+def _series_of(doc: dict) -> tuple[str, str]:
+    """Name and id of the featured series; series_names[0] can be another one."""
+    featured = ((doc.get("featured_series") or {}).get("series")) or {}
+    if featured.get("name"):
+        return featured["name"], str(featured.get("id") or "")
+    names = doc.get("series_names") or []
+    ids = doc.get("series_ids") or []
+    return (names[0] if names else ""), (str(ids[0]) if ids else "")
+
+
+def _volume(position) -> int | None:
+    # position 0 means "somewhere in this series", not "episode zero"
+    try:
+        number = int(float(position))
+    except (TypeError, ValueError):
+        return None
+    return number or None
+
+
 def _to_result(doc: dict, language: str) -> MetadataResult:
-    series = doc.get("series_names") or []
+    series, series_id = _series_of(doc)
    year = doc.get("release_year")
    return MetadataResult(
        media_type="ebook",
@@ -97,16 +137,18 @@ def _to_result(doc: dict, language: str) -> MetadataResult:
        title=doc.get("title", ""),
        authors=_authors(doc),
        external_id=_isbn(doc.get("isbns")),
-        year=int(year) if str(year).isdigit() else None,
+        year=int(str(year)) if str(year).isdigit() else None,
        language=language,
-        series=series[0] if series else "",
+        series=series,
+        series_id=series_id,
+        volume=_volume(doc.get("featured_series_position")),
        cover_url=_image_url(doc.get("image")),
    )


-async def _localize(client: httpx.AsyncClient, results: dict[int, MetadataResult],
-                    language: str) -> list[MetadataResult]:
-    """Replace each hit with its edition in the wanted language, drop the rest."""
+async def _localized(client: httpx.AsyncClient, results: dict[int, MetadataResult],
+                     language: str) -> dict[int, MetadataResult]:
+    """The hits that have an edition in the wanted language, swapped in."""
    data = await _post(client, _EDITIONS,
                       {"ids": list(results), "lang": LANGUAGES[language]})
    localized: dict[int, MetadataResult] = {}
@@ -118,11 +160,19 @@ async def _localize(client: httpx.AsyncClient, results: dict[int, MetadataResult
        release = str(ed.get("release_date") or "")
        localized[book_id] = base.model_copy(update={
            "title": ed.get("title") or base.title,
-            "external_id": _isbn(ed.get("isbn_13"), ed.get("isbn_10")) or base.external_id,
+            "external_id": (_isbn(ed.get("isbn_13"), ed.get("isbn_10"))
+                            or base.external_id),
            "year": int(release[:4]) if release[:4].isdigit() else base.year,
            "language": language,
            "cover_url": _image_url(ed.get("cached_image")) or base.cover_url,
        })
+    return localized
+
+
+async def _localize(client: httpx.AsyncClient, results: dict[int, MetadataResult],
+                    language: str) -> list[MetadataResult]:
+    """Replace each hit with its edition in the wanted language, drop the rest."""
+    localized = await _localized(client, results, language)
    return [localized[book_id] for book_id in results if book_id in localized]


@@ -139,7 +189,7 @@ async def search(query: str, language: str = "",
            doc = hit.get("document") or {}
            book_id = doc.get("id")
            try:
-                book_id = int(book_id)
+                book_id = int(str(book_id))
            except (TypeError, ValueError):
                continue
            if doc.get("title"):
@@ -152,3 +202,58 @@ async def search(query: str, language: str = "",
                # without the edition lookup the hits are still worth showing
                log.warning("hardcover edition lookup failed for %r: %s", query, exc)
    return list(results.values())
+
+
+async def series_books(series_id: str, language: str = "") -> list[MetadataResult]:
+    """Every volume of a series, by position. A position often holds several book
+    rows (the same story as another work); the one with an edition in the wanted
+    language wins the slot, else the first. Unnumbered books go last."""
+    if not TOKEN:
+        return []
+    try:
+        numeric_id = int(str(series_id))
+    except (TypeError, ValueError):
+        return []
+
+    async with httpx.AsyncClient(timeout=30) as client:
+        data = await _post(client, _SERIES, {"id": numeric_id, "limit": SERIES_LIMIT})
+        row = (data.get("series") or [{}])[0]
+        name = row.get("name", "")
+
+        slots: dict[int, list[int]] = {}  # volume -> book rows filed under it
+        loose: list[int] = []
+        books: dict[int, MetadataResult] = {}
+        for entry in row.get("book_series") or []:
+            book = entry.get("book") or {}
+            try:
+                book_id = int(str(book.get("id")))
+            except (TypeError, ValueError):
+                continue
+            if not book.get("title") or book_id in books:
+                continue
+            volume = _volume(entry.get("position"))
+            books[book_id] = _to_result(
+                {**book, "image": book.get("cached_image"),
+                 "series_names": [name], "series_ids": [numeric_id],
+                 "featured_series_position": volume},
+                "",
+            )
+            if volume is None:
+                loose.append(book_id)
+            else:
+                slots.setdefault(volume, []).append(book_id)
+
+        localized: dict[int, MetadataResult] = {}
+        if language in LANGUAGES and books:
+            try:
+                localized = await _localized(client, books, language)
+            except (httpx.HTTPError, RuntimeError) as exc:
+                log.warning("hardcover series editions failed for %s: %s",
+                            series_id, exc)
+
+    def pick(candidates: list[int]) -> MetadataResult:
+        best = next((i for i in candidates if i in localized), candidates[0])
+        return localized.get(best) or books[best]
+
+    return ([pick(slots[volume]) for volume in sorted(slots)]
+            + [localized.get(i) or books[i] for i in loose])