fix: keep the identifiers an EPUB declares

DC:identifier was validated verbatim, so hyphenated and urn:isbn: forms never
reached the checksum and every non-ISBN identifier was discarded. Normalise
first, and name whatever survives.

backend/tests/unit/test_matching.py

@@ -7,6 +7,7 @@ from chitai.services.matching import (
     normalize_identifier,
     normalize_title,
 )
+from chitai.services.metadata_extractor import parse_identifier
 from chitai.services.utils import isbn10_to_isbn13
 
 
@@ -123,3 +124,45 @@ class TestIsbnConversion:
     @pytest.mark.parametrize("isbn", ["0486282113", "9780486282114", "nonsense"])
     def test_anything_that_is_not_an_isbn_10_converts_to_nothing(self, isbn: str) -> None:
         assert isbn10_to_isbn13(isbn) is None
+
+
+class TestParseIdentifier:
+    """What an EPUB writes, and what is worth storing for it."""
+
+    @pytest.mark.parametrize(
+        "written",
+        [
+            "9780486282114",
+            "978-0-486-28211-4",
+            "urn:isbn:9780486282114",
+            "urn:isbn:978-0-486-28211-4",
+            "ISBN:978-0-486-28211-4",
+        ],
+    )
+    def test_isbns_survive_however_they_are_written(self, written: str) -> None:
+        assert parse_identifier(written) == ("isbn-13", "9780486282114")
+
+    def test_the_scheme_attribute_is_read_too(self) -> None:
+        assert parse_identifier("0-486-28211-2", "ISBN") == ("isbn-10", "0486282112")
+
+    def test_non_isbn_identifiers_are_kept(self) -> None:
+        assert parse_identifier("urn:uuid:3f2b1c4e-1111-2222-3333-444455556666") == (
+            "uuid",
+            "3f2b1c4e-1111-2222-3333-444455556666",
+        )
+        assert parse_identifier("calibre:1234") == ("calibre", "1234")
+        assert parse_identifier("B000FC0PDA", "mobi-asin") == ("asin", "B000FC0PDA")
+
+    def test_an_unrecognised_prefix_is_part_of_the_value(self) -> None:
+        """"http://example.com/book" is not an identifier called "http"."""
+        assert parse_identifier("http://www.gutenberg.org/5200") == (
+            "id",
+            "http://www.gutenberg.org/5200",
+        )
+
+    def test_a_declared_isbn_that_is_not_one_is_dropped(self) -> None:
+        assert parse_identifier("urn:isbn:not-an-isbn") is None
+
+    @pytest.mark.parametrize("written", ["", "   ", None])
+    def test_nothing_yields_nothing(self, written: str | None) -> None:
+        assert parse_identifier(written) is None