built ingest

python/ebook_search/epub_parse.py

@@ -0,0 +1,95 @@
+"""EPUB parsing helpers."""
+
+from __future__ import annotations
+
+import re
+from dataclasses import dataclass
+from typing import TYPE_CHECKING
+
+from bs4 import BeautifulSoup
+from ebooklib import ITEM_DOCUMENT, epub
+
+if TYPE_CHECKING:
+    from pathlib import Path
+
+WHITESPACE_RE = re.compile(r"\s+")
+
+
+@dataclass(frozen=True)
+class ParsedChapter:
+    """Text extracted from one EPUB spine document."""
+
+    title: str | None
+    href: str | None
+    text: str
+    page_labels: tuple[str, ...]
+
+
+@dataclass(frozen=True)
+class ParsedEpub:
+    """Parsed EPUB metadata and text."""
+
+    title: str
+    author: str | None
+    language: str | None
+    publisher: str | None
+    identifier: str | None
+    chapters: tuple[ParsedChapter, ...]
+
+
+def parse_epub(path: Path) -> ParsedEpub:
+    """Parse EPUB metadata and spine text."""
+    book = epub.read_epub(path)
+    chapters = []
+    for item in book.get_items_of_type(ITEM_DOCUMENT):
+        soup = BeautifulSoup(item.get_content(), "html.parser")
+        title = chapter_title(soup)
+        page_labels = tuple(extract_page_labels(soup))
+        text = clean_text(soup.get_text(" "))
+        if text:
+            chapters.append(ParsedChapter(title=title, href=item.get_name(), text=text, page_labels=page_labels))
+
+    return ParsedEpub(
+        title=metadata_value(book, "title") or path.stem,
+        author=metadata_value(book, "creator"),
+        language=metadata_value(book, "language"),
+        publisher=metadata_value(book, "publisher"),
+        identifier=metadata_value(book, "identifier"),
+        chapters=tuple(chapters),
+    )
+
+
+def metadata_value(book: epub.EpubBook, name: str) -> str | None:
+    """Return the first non-empty Dublin Core metadata value for a name."""
+    values = book.get_metadata("DC", name)
+    if not values:
+        return None
+    value = values[0][0]
+    return str(value).strip() or None
+
+
+def chapter_title(soup: BeautifulSoup) -> str | None:
+    """Extract the best available title from an EPUB document soup."""
+    heading = soup.find(["h1", "h2", "h3"])
+    if heading is None:
+        title = soup.find("title")
+        if title is None:
+            return None
+        return clean_text(title.get_text(" ")) or None
+    return clean_text(heading.get_text(" ")) or None
+
+
+def extract_page_labels(soup: BeautifulSoup) -> list[str]:
+    """Extract EPUB page-break labels from a document soup."""
+    labels: list[str] = []
+    for tag in soup.find_all(attrs={"epub:type": "pagebreak"}):
+        label = tag.get("title") or tag.get("aria-label") or tag.get_text(" ")
+        clean = clean_text(str(label))
+        if clean:
+            labels.append(clean)
+    return labels
+
+
+def clean_text(text: str) -> str:
+    """Normalize whitespace in extracted EPUB text."""
+    return WHITESPACE_RE.sub(" ", text).strip()