feat(metadata): clean pasted identifiers instead of truncating them

ISBNs copied from Amazon carry hyphens, and copied ASINs often start with a zero-width space.
The edit tab cut a hyphenated ISBN at its maxlength, Book Dock sent it on for the server to
reject, and an invisible character left an ID that looked right but never matched.

Clean ISBN, ASIN and provider id input as it is typed or pasted, and apply the same shared
rule in the book metadata and Book Dock DTOs so every client, including released iOS builds,
gets it. Provider ids keep their punctuation because Google ids and slugs use it, and an
over-long value is kept whole so validation still names the field.

Closes #1624

packages/types/src/identifier-input.ts

@@ -0,0 +1,82 @@
+/**
+ * How a hand-entered identifier is cleaned before it is stored. An identifier copied from a store
+ * page carries its print formatting (hyphens, spaces) and often an invisible character such as a
+ * zero-width space or a bidi mark. None of these belong to the identifier, and a field that looks
+ * right while holding one is very hard to diagnose. Cleaning never truncates: an over-long result
+ * stays over-long so validation can name it.
+ */
+export type IdentifierInputKind = "isbn" | "asin" | "provider-id";
+
+/**
+ * The cleaning rule for each editable identifier field. Only ISBNs and ASINs have a fixed alphabet;
+ * every other provider id keeps its punctuation, because Google Books ids and Hardcover and Kobo
+ * slugs legitimately contain `-` and `_`.
+ */
+export const IDENTIFIER_INPUT_KIND_BY_FIELD = {
+  isbn10: "isbn",
+  isbn13: "isbn",
+  amazonId: "asin",
+  audibleId: "asin",
+  googleBooksId: "provider-id",
+  goodreadsId: "provider-id",
+  hardcoverId: "provider-id",
+  hardcoverEditionId: "provider-id",
+  openLibraryId: "provider-id",
+  itunesId: "provider-id",
+  librofmId: "provider-id",
+  koboId: "provider-id",
+  comicvineId: "provider-id",
+  ranobedbId: "provider-id",
+  lubimyczytacId: "provider-id",
+  aladinId: "provider-id",
+} as const satisfies Record<string, IdentifierInputKind>;
+
+export type IdentifierInputField = keyof typeof IDENTIFIER_INPUT_KIND_BY_FIELD;
+
+// Whitespace plus Unicode control (Cc) and format (Cf) characters: zero-width spaces and joiners,
+// bidi marks, the byte-order mark and the soft hyphen.
+const INVISIBLE_OR_WHITESPACE = /[\s\p{Cc}\p{Cf}]/gu;
+
+// Amazon prints "ISBN-13 : 978-...". Without this the label's digits survive the digit filter and
+// prefix the real number. The colon is required so an ISBN-10 starting with 10 or 13 is never cut.
+const NUMBERED_ISBN_LABEL = /^ISBN\p{Pd}?1[03]:/iu;
+
+/** Keeps the digits and the ISBN-10 check character `X`, so `978-3426226223` becomes `9783426226223`. */
+export function normalizeIsbnInput(value: string): string {
+  return value
+    .normalize("NFKC")
+    .replace(INVISIBLE_OR_WHITESPACE, "")
+    .replace(NUMBERED_ISBN_LABEL, "")
+    .toUpperCase()
+    .replace(/[^0-9X]/g, "");
+}
+
+/** ASINs are upper-case alphanumerics; anything else is formatting or an invisible character. */
+export function normalizeAsinInput(value: string): string {
+  return value
+    .normalize("NFKC")
+    .toUpperCase()
+    .replace(/[^0-9A-Z]/g, "");
+}
+
+/** Removes only whitespace and invisible characters, which no provider id contains. */
+export function normalizeProviderIdInput(value: string): string {
+  return value.replace(INVISIBLE_OR_WHITESPACE, "");
+}
+
+export function normalizeIdentifierInput(kind: IdentifierInputKind, value: string): string {
+  switch (kind) {
+    case "isbn":
+      return normalizeIsbnInput(value);
+    case "asin":
+      return normalizeAsinInput(value);
+    case "provider-id":
+      return normalizeProviderIdInput(value);
+  }
+}
+
+export function identifierInputKindForField(field: string): IdentifierInputKind | null {
+  return Object.prototype.hasOwnProperty.call(IDENTIFIER_INPUT_KIND_BY_FIELD, field)
+    ? IDENTIFIER_INPUT_KIND_BY_FIELD[field as IdentifierInputField]
+    : null;
+}

packages/types/src/index.ts

@@ -12,6 +12,7 @@ export * from "./metadata-preferences";
 export * from "./custom-metadata";
 export * from "./custom-icon";
 export * from "./metadata-lock";
+export * from "./identifier-input";
 export * from "./dashboard";
 export * from "./book";
 export * from "./book-duplicates";