s are layout helpers (Springer's
+ div.ClearBoth) and paragraphs holding only anchors are index targets – neither is a scene break."""
+ return local(el.tag) == "p" and self._is_empty(el) and \
+ all(local(c.tag) == "br" for c in el if is_element(c))
+
+ def _is_prose(self, el: Any) -> bool:
+ return local(el.tag) in ("p", "div") and not self._is_empty(el) and not self._is_ornament(el) \
+ and not self._is_poem(el)
+
+ @staticmethod
+ def _element_siblings(el: Any, forward: bool) -> Any:
+ node = el.getnext() if forward else el.getprevious()
+ while node is not None:
+ if is_element(node):
+ yield node
+ node = node.getnext() if forward else node.getprevious()
+
+ def _is_ornament(self, el: Any) -> bool:
+ """Paragraph that only contains a typographic separator such as *** or ⁂."""
+ return bool(ORNAMENT.fullmatch(text_content(el).replace("\xa0", " ").strip()))
+
+ def _is_scene_gap(self, el: Any) -> bool:
+ """Empty paragraph between two prose paragraphs (first of a run of empties) → scene break."""
+ if not self._is_blank_p(el):
+ return False
+ prev = next(self._element_siblings(el, forward=False), None)
+ if prev is None or not self._is_prose(prev):
+ return False
+ for nxt in self._element_siblings(el, forward=True):
+ if self._is_blank_p(nxt):
+ continue # further empty paragraphs of the same gap
+ return self._is_prose(nxt)
+ return False
+
+ def _is_poem(self, el: Any) -> bool:
+ types = set(epub_types(el))
+ return bool(types & {"z3998:poem", "z3998:verse", "z3998:song", "z3998:lyrics"}) or \
+ any(POEM_CLASS.search(c) for c in classes(el))
+
+ def poem(self, el: Any) -> list[str]:
+ """Verse: one quoted line per verse line, stanzas separated by an empty quote line."""
+ anchor = self.anchors_of(el)
+ stanzas: list[list[str]] = [[]]
+
+ def new_stanza() -> None:
+ if stanzas[-1]:
+ stanzas.append([])
+
+ def walk(container: Any) -> None:
+ for child in container:
+ if not is_element(child) or self._hidden(child):
+ continue
+ if local(child.tag) in ("div", "section") and any(
+ is_element(c) and local(c.tag) in ("p", "div", "span") for c in child):
+ new_stanza() # stanza container
+ walk(child)
+ new_stanza()
+ elif self._is_empty(child):
+ new_stanza()
+ else:
+ line = join_inline([("r", self.anchors_of(child))] + self.inline_children(child))
+ stanzas[-1].append(escape_line_starts(line).replace("\n", "\n> "))
+
+ walk(el)
+ stanzas = [s for s in stanzas if s]
+ if not stanzas:
+ return self._flow_with_anchor(el)
+ self.poems += 1
+ text = "\n>\n".join("\\\n".join(f"> {line}" for line in stanza) for stanza in stanzas)
+ return ([anchor] if anchor else []) + [text]
+
def _flow_with_anchor(self, el: Any) -> list[str]:
anchor = self.anchors_of(el)
if self.ch is not None and {"index", "index-persons", "glossary"} & set(self.ch.types):
diff --git a/epub2md/pipeline.py b/epub2md/pipeline.py
index e0dd2fa..df4770e 100644
--- a/epub2md/pipeline.py
+++ b/epub2md/pipeline.py
@@ -517,6 +517,7 @@ def _report(input_path: Path, sha: str, generated_at: str, package: Any, opts: O
"text_cleanup": {"invisible_chars_removed": renderer.invisible_removed},
"css": {"rules": len(renderer.css.rules) if renderer.css else 0, "applied": dict(renderer.css_applied),
"hidden_elements": renderer.hidden_elements},
+ "layout": {"scene_breaks": renderer.scene_breaks, "poems": renderer.poems},
"warnings": warnings,
"errors": errors,
"infos": [d.as_dict() for d in diag.items if d.level == "info"],
diff --git a/tests/test_styles.py b/tests/test_styles.py
index 36100fd..b80f217 100644
--- a/tests/test_styles.py
+++ b/tests/test_styles.py
@@ -66,3 +66,24 @@ def test_adjacent_emphasis_is_merged(convert):
text = styled(body, convert=convert).chapter("001")
assert "
2*×π*" in text
assert "*xy*" in text and "**ab**" in text and "*c* *d*" in text
+
+
+def test_p4_scene_breaks(convert):
+ body = ('
K
Erster Absatz.
'
+ '
Neue Szene.
Weiter.
***
'
+ '
Nach Ornament.
Abschnitt
Text
⁂
Ende.
')
+ text = convert(Book(docs=[Doc("c.xhtml", body)])).chapter("001")
+ body_md = text.split("# K", 1)[1]
+ assert "Erster Absatz.\n\n* * *\n\nNeue Szene." in body_md
+ assert body_md.count("* * *") == 3 # two empties collapse into one; ***, ⁂ are ornaments; none before the heading
+ assert "Weiter.\n\n* * *\n\nNach Ornament." in body_md
+ assert "\\*" not in body_md
+ assert body_md.strip().startswith("Erster Absatz.") # no break at chapter start
+
+
+def test_p4_poems_keep_lines_and_stanzas(convert):
+ body = ('
G
Davor.
»Erste Zeile
zweite Zeile
'
+ '
Zweite Strophe«
Danach.
')
+ text = convert(Book(docs=[Doc("c.xhtml", body)])).chapter("001")
+ assert "> »Erste Zeile\\\n> zweite Zeile\n>\n> Zweite Strophe«" in text
+ assert "* * *" not in text