feat: ePub builder with chapters, images, TOC, cover

Made-with: Cursor
2026-04-06 15:17:21 -04:00
parent d88a0817b7
commit 01f8366e85
2 changed files with 220 additions and 0 deletions
--- a/src/epub_builder.py
+++ b/src/epub_builder.py
@@ -0,0 +1,118 @@
+import json
+import os
+from datetime import date
+
+from ebooklib import epub
+
+from src.models import Article, Image
+
+EPUB_CSS = """
+body { font-family: serif; margin: 1em; line-height: 1.5; }
+h1 { font-size: 1.4em; margin-bottom: 0.3em; }
+.byline { font-size: 0.85em; color: #555; margin-bottom: 0.5em; }
+.categories { font-size: 0.8em; color: #777; margin-bottom: 1em; }
+img { max-width: 100%; display: block; margin: 0.5em auto; }
+figcaption { font-size: 0.8em; text-align: center; color: #555; }
+"""
+
+
+def build_epub(
+    week_start: date,
+    week_end: date,
+    article_ids: list[int],
+    cover_path: str,
+    output_dir: str,
+) -> str:
+    os.makedirs(output_dir, exist_ok=True)
+
+    articles = (
+        Article.query
+        .filter(Article.id.in_(article_ids))
+        .order_by(Article.pub_date.asc())
+        .all()
+    )
+
+    title = (
+        f"Plymouth Independent \u2014 "
+        f"Week of {week_start.strftime('%b %d')}\u2013{week_end.strftime('%b %d, %Y')}"
+    )
+
+    book = epub.EpubBook()
+    book.set_identifier(f"pi-{week_start.isoformat()}")
+    book.set_title(title)
+    book.set_language("en")
+    book.add_author("Plymouth Independent")
+
+    with open(cover_path, "rb") as f:
+        book.set_cover("cover.jpg", f.read())
+
+    style = epub.EpubItem(
+        uid="style", file_name="style/default.css",
+        media_type="text/css", content=EPUB_CSS.encode("utf-8"),
+    )
+    book.add_item(style)
+
+    chapters = []
+    image_counter = 0
+
+    for article in articles:
+        categories = json.loads(article.categories)
+        cat_str = ", ".join(categories) if categories else ""
+
+        chapter_html = f"<h1>{article.title}</h1>\n"
+        chapter_html += (
+            f'<p class="byline">{article.author} \u00b7 '
+            f'{article.pub_date.strftime("%B %d, %Y")}</p>\n'
+        )
+        if cat_str:
+            chapter_html += f'<p class="categories">{cat_str}</p>\n'
+
+        content = article.content_html
+        article_images = Image.query.filter_by(article_id=article.id).all()
+
+        for img_record in article_images:
+            if not os.path.exists(img_record.local_path):
+                continue
+
+            image_counter += 1
+            epub_img_name = f"images/img_{image_counter}.jpg"
+
+            with open(img_record.local_path, "rb") as f:
+                img_data = f.read()
+
+            epub_img = epub.EpubItem(
+                uid=f"img_{image_counter}",
+                file_name=epub_img_name,
+                media_type="image/jpeg",
+                content=img_data,
+            )
+            book.add_item(epub_img)
+            content = content.replace(img_record.local_path, epub_img_name)
+
+        chapter_html += content
+
+        chapter = epub.EpubHtml(
+            title=article.title,
+            file_name=f"chapter_{article.id}.xhtml",
+            lang="en",
+        )
+        chapter.set_content(
+            f'<html><head><link rel="stylesheet" href="style/default.css"/>'
+            f"</head><body>{chapter_html}</body></html>"
+        )
+        chapter.add_item(style)
+        chapters.append(chapter)
+        book.add_item(chapter)
+
+    book.toc = [(c, []) for c in chapters]
+    book.add_item(epub.EpubNcx())
+    book.add_item(epub.EpubNav())
+
+    book.spine = ["nav"] + chapters
+
+    iso_week = week_start.isocalendar()[1]
+    filename = f"plymouth-independent-{week_start.year}-W{iso_week:02d}.epub"
+    epub_path = os.path.join(output_dir, filename)
+    epub.write_epub(epub_path, book)
+
+    return epub_path