feat: epub2md – EPUB nach kapitelweisem Markdown

Kommandozeilenwerkzeug, das DRM-freie EPUB-2/3-Dateien offline in
kapitelweise Markdown-Dateien umwandelt. Erzeugt werden außerdem Medien,
metadata.yaml, manifest.json, conversion-report.json und Diagnose-Dateien.

- Spine-basierte Kapitelbildung mit --chapter-level, epub:type-Abschnitten,
  Fortsetzungsdateien und Erkennung von Layout-Überschriften
- Code mit exakter Einrückung, sicheren Fences und konservativer
  Spracherkennung
- MathML->TeX (konservativ), TeX aus alt-Texten, SVG-/Bildformeln
  bleiben Bilder
- Fußnoten-Modi end-of-chapter, end-of-book, inline und keep-links
- Interne Links/Anker werden umgeschrieben, Linkcheck über die Ausgabe
- Sicher: kein Entpacken, Schutz vor Path Traversal und XXE,
  DRM-Erkennung, atomares Schreiben, --force nur für frühere Ausgaben
- 82 pytest-Tests mit synthetischen EPUBs und Invariantentests für Beispiele/

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

tests/test_codeblocks.py

@@ -0,0 +1,88 @@
+import re
+
+from conftest import Book, Doc
+
+PY = """def read_book(path):
+    if path:
+        return path.read_text(encoding="utf-8")  # <utf-8>
+\treturn None"""
+
+
+def fenced_blocks(text: str) -> list[tuple[str, str]]:
+    return re.findall(r"^(`{3,})([\w+-]*)\n(.*?)\n\1$", text, re.M | re.S)
+
+
+def test_python_indentation_and_special_chars(convert):
+    body = f'<h1>Code</h1><pre><code class="language-python">{PY}</code></pre>'
+    run = convert(Book(docs=[Doc("c.xhtml", body)]))
+    blocks = fenced_blocks(run.chapter("001"))
+    assert len(blocks) == 1
+    fence, lang, code = blocks[0]
+    assert lang == "python"
+    assert code == PY.replace("&lt;", "<").replace("&gt;", ">")
+    assert "&lt;" not in code
+
+
+def test_shell_code_with_pipes_backslashes_and_dollar(convert):
+    shell = "$ cat datei.txt | grep -v '^#' \\\n    | sort &amp;&amp; echo \"$HOME\" > out.txt"
+    body = f"<h1>Shell</h1><pre>{shell}</pre>"
+    run = convert(Book(docs=[Doc("c.xhtml", body)]))
+    (_, lang, code), = fenced_blocks(run.chapter("001"))
+    assert code == "$ cat datei.txt | grep -v '^#' \\\n    | sort && echo \"$HOME\" > out.txt"
+    assert lang == ""  # no reliable language signal -> no label
+
+
+def test_backticks_in_code_get_longer_fence(convert):
+    body = "<h1>T</h1><pre>echo ```nested``` and `x`</pre>"
+    run = convert(Book(docs=[Doc("c.xhtml", body)]))
+    (fence, _, code), = fenced_blocks(run.chapter("001"))
+    assert fence == "````"
+    assert code == "echo ```nested``` and `x`"
+
+
+def test_br_lines_and_highlight_spans_in_pre(convert):
+    body = ('<h1>C</h1><div class="listing"><pre><span class="kw">#include</span> &lt;stdio.h&gt;<br/><br/>'
+            'int main(void) {<br/>   printf("Hallo\\n");<br/>} </pre></div>')
+    run = convert(Book(docs=[Doc("c.xhtml", body)]))
+    (_, lang, code), = fenced_blocks(run.chapter("001"))
+    assert lang == "c"
+    assert code == '#include <stdio.h>\n\nint main(void) {\n   printf("Hallo\\n");\n} '
+
+
+def test_listing_div_with_span_pre(convert):
+    body = ('<h1>K</h1><div class="divUnProgramlisting"><span class="pre"><strong>def</strong> f(x):\n'
+            '    <strong>return</strong> 2 * x\n\n<strong>def</strong> g(x):\n    <strong>return</strong> x</span></div>')
+    run = convert(Book(docs=[Doc("c.xhtml", body)]))
+    (_, lang, code), = fenced_blocks(run.chapter("001"))
+    assert lang == "python"
+    assert code == "def f(x):\n    return 2 * x\n\ndef g(x):\n    return x"
+
+
+def test_line_numbers_are_removed(convert):
+    body = '<h1>L</h1><pre><span class="lineno">1</span>a = 1<br/><span class="lineno">2</span>b = 2</pre>'
+    run = convert(Book(docs=[Doc("c.xhtml", body)]))
+    (_, _, code), = fenced_blocks(run.chapter("001"))
+    assert code == "a = 1\nb = 2"
+
+
+def test_inline_code_variants(convert):
+    body = ('<h1>I</h1><p>Nutze <code>a*b_c</code>, <span class="tt">x | y</span> und <code>`tick`</code> '
+            'sowie <kbd>Strg</kbd>.</p>')
+    run = convert(Book(docs=[Doc("c.xhtml", body)]))
+    text = run.chapter("001")
+    assert "`a*b_c`" in text
+    assert "`x | y`" in text
+    assert "`` `tick` ``" in text
+    assert "<kbd>Strg</kbd>" in text
+
+
+def test_prose_is_not_forced_into_code(convert):
+    body = "<h1>P</h1><p>Ein normaler Absatz mit    Leerzeichen.</p>"
+    run = convert(Book(docs=[Doc("c.xhtml", body)]))
+    assert "```" not in run.chapter("001")
+    assert "Ein normaler Absatz mit Leerzeichen." in run.chapter("001")
+
+
+def test_unlabeled_code_is_reported(convert):
+    run = convert(Book(docs=[Doc("c.xhtml", "<h1>U</h1><pre>foo bar</pre>")]))
+    assert run.report["code_blocks"]["languages"] == {"(keine)": 1}