feat: split the edition out of a book's title

"Fluent Python, 2nd Edition" is one book with a field for the edition. Left in
the title it also splits the library, since the second edition never looks like
the first. Only a numbered statement is moved, so "Catch 22" keeps its number
and "Global Edition" — which has nowhere to go in an integer column — stays put.

backend/tests/integration/test_book.py

@@ -37,7 +37,8 @@ from pathlib import Path
         (
             Path("tests/data_files/Calculus Made Easy - Silvanus Thompson.pdf"),
             2,
-            "The Project Gutenberg eBook #33283: Calculus Made Easy, 2nd Edition",
+            # The ", 2nd Edition" is split off into `edition`, not kept in the title.
+            "The Project Gutenberg eBook #33283: Calculus Made Easy",
             ["Silvanus Phillips Thompson"],
         ),
     ],

backend/tests/unit/test_metadata_extractor.py

@@ -1,7 +1,12 @@
 import pytest
 from pathlib import Path
 from datetime import date
-from chitai.services.metadata_extractor import EpubExtractor, Extractor, PdfExtractor
+from chitai.services.metadata_extractor import (
+    EpubExtractor,
+    Extractor,
+    PdfExtractor,
+    split_edition,
+)
 
 
 @pytest.mark.asyncio()
@@ -74,3 +79,96 @@ class TestIdentifierMerging:
         metadata = await Extractor.extract_metadata([PDF])
 
         assert "identifiers" not in metadata
+
+
+class TestSplitEdition:
+    """An edition is a field on the book, not part of what the book is called."""
+
+    @pytest.mark.parametrize(
+        ("title", "stripped", "edition"),
+        [
+            # Every form below is one that turned up in a real library.
+            ("Fluent Python, 2nd Edition", "Fluent Python", 2),
+            ("Building Microservices, 2E", "Building Microservices", 2),
+            ("Digital Image Processing, 4e", "Digital Image Processing", 4),
+            (
+                "Network Security Essentials: Applications and Standards/6e",
+                "Network Security Essentials: Applications and Standards",
+                6,
+            ),
+            (
+                "Refactoring: Improving the Design of Existing Code (2nd edition)",
+                "Refactoring: Improving the Design of Existing Code",
+                2,
+            ),
+            # Ordinal words, including a qualifier sitting inside the statement.
+            (
+                "The Art of Computer Programming: Volume 1 / Fundamental Algorithms, Third Edition",
+                "The Art of Computer Programming: Volume 1 / Fundamental Algorithms",
+                3,
+            ),
+            (
+                "Introduction to the Theory of Computation, Third International Edition",
+                "Introduction to the Theory of Computation",
+                3,
+            ),
+            # Mid-title, before a subtitle and before a trailing author.
+            (
+                "How Linux Works, 3rd Edition: What Every Superuser Should Know",
+                "How Linux Works: What Every Superuser Should Know",
+                3,
+            ),
+            (
+                "Code Complete, 2nd Edition - Steve McConnell",
+                "Code Complete - Steve McConnell",
+                2,
+            ),
+            # An underscore between the number and the "e", beside an unnumbered
+            # qualifier that has nowhere to go in an integer column and so stays put.
+            (
+                "Cryptography and Network Security, Global Edition, 8_e - Stallings",
+                "Cryptography and Network Security, Global Edition - Stallings",
+                8,
+            ),
+        ],
+    )
+    def test_editions_are_split_out(self, title: str, stripped: str, edition: int) -> None:
+        assert split_edition(title) == (stripped, edition)
+
+    @pytest.mark.parametrize(
+        "title",
+        [
+            # A number alone is never an edition — these are titles.
+            "Catch 22",
+            "Fahrenheit 451",
+            "Blade Runner 2049",
+            "1984",
+            "Apollo 13",
+            "Slaughterhouse 5",
+            "The Art of Computer Programming: Volume 1",
+            # "Edition" with no number cannot be stored, so it stays where it can
+            # still be read.
+            "Cryptography and Network Security: Principles and Practice, Global Edition",
+            "Building Microservices",
+        ],
+    )
+    def test_titles_are_left_alone(self, title: str) -> None:
+        assert split_edition(title) == (title, None)
+
+    def test_a_title_that_is_only_an_edition_is_kept(self) -> None:
+        """Stripping must never leave a book with no title at all."""
+        assert split_edition("2nd Edition") == ("2nd Edition", None)
+
+    @pytest.mark.parametrize("title", ["", None])
+    def test_nothing_yields_nothing(self, title: str | None) -> None:
+        assert split_edition(title) == (title, None)
+
+
+@pytest.mark.asyncio()
+class TestEditionFromFiles:
+    async def test_extraction_moves_the_edition_off_the_title(self) -> None:
+        """The PDF fixture calls itself a 2nd edition in its own metadata title."""
+        metadata = await Extractor.extract_metadata([PDF])
+
+        assert metadata["title"] == "The Project Gutenberg eBook #33283: Calculus Made Easy"
+        assert metadata["edition"] == 2