build_epub.py

#!/usr/bin/env python3
"""
Build an EPUB from OCR text files and optional cover image.

When the first page is the book title/cover, it is kept as the cover image only;
OCR text from page 2 onward is used for the body.

Requires: pip install ebooklib

Usage:
  python build_epub.py [--pages-dir DIR] [--output FILE] [--title TITLE] [--author AUTHOR]
  python build_epub.py --pages-dir love-and-limerence-pages -o Love_and_Limerence.epub --title "Love and Limerence" --author "Dorothy Tennov"
"""

import argparse
import re
from pathlib import Path
from typing import List

try:
    import ebooklib
    from ebooklib import epub
except ImportError:
    print("Install ebooklib: pip install ebooklib", file=__import__("sys").stderr)
    raise SystemExit(1)


def natural_sort_key(p: Path) -> tuple:
    m = re.search(r"(\d+)\s*$", p.stem)
    return (int(m.group(1)),) if m else (0,)


def find_text_files(pages_dir: Path) -> List[Path]:
    files = list(pages_dir.glob("*.txt"))
    return sorted(set(files), key=natural_sort_key)


def html_escape(s: str) -> str:
    return (
        s.replace("&", "&")
        .replace("<", "&lt;")
        .replace(">", "&gt;")
        .replace('"', "&quot;")
    )


def main():
    ap = argparse.ArgumentParser(description="Build EPUB from OCR text files.")
    ap.add_argument("--pages-dir", type=Path, default=Path("stand-on-zanzibar-pages"), help="Directory with .txt (and optional images)")
    ap.add_argument("--output", "-o", type=Path, default=Path("Stand_on_Zanzibar.epub"), help="Output EPUB path")
    ap.add_argument("--title", default="Stand on Zanzibar", help="Book title")
    ap.add_argument("--author", default="John Brunner", help="Author")
    ap.add_argument("--chapters-per-part", type=int, default=50, help="Group N pages per chapter (0 = one chapter for whole book)")
    ap.add_argument("--first-page-is-cover", action="store_true", default=True, help="Use first page as cover image only; body text starts from page 2 (default)")
    ap.add_argument("--no-first-page-is-cover", action="store_false", dest="first_page_is_cover", help="Include first page OCR in body (don't treat as cover-only)")
    args = ap.parse_args()

    pages_dir = args.pages_dir.resolve()
    if not pages_dir.is_dir():
        ap.error(f"Directory not found: {pages_dir}")

    txt_files = find_text_files(pages_dir)
    if not txt_files:
        ap.error("No .txt files found.")

    # First page = cover image only; body text starts from page 2 (skip first .txt)
    if args.first_page_is_cover and len(txt_files) > 1:
        txt_files_for_body = txt_files[1:]
    else:
        txt_files_for_body = txt_files

    book = epub.EpubBook()
    book.set_identifier("stand-on-zanzibar-ocr")
    book.set_title(args.title)
    book.set_language("en")
    book.add_author(args.author)

    # Optional cover: first image in directory (page 1 = title/cover)
    for ext in ["png", "jpg", "jpeg"]:
        cover_candidates = sorted(pages_dir.glob(f"*.{ext}"), key=natural_sort_key)
        if cover_candidates:
            cover_path = cover_candidates[0]
            cover_content = cover_path.read_bytes()
            book.set_cover(f"cover.{ext}", cover_content)
            if args.first_page_is_cover:
                print(f"Cover: {cover_path.name} (first page as image)")
            break

    # Collect body text from OCR (page 2 onward when first page is cover)
    all_paragraphs = []  # type: List[str]
    for tf in txt_files_for_body:
        raw = tf.read_text(encoding="utf-8", errors="replace").strip()
        if not raw:
            continue
        # Normalize: split into paragraphs (double newline or single long line)
        for block in re.split(r"\n\s*\n", raw):
            block = block.strip()
            if block:
                all_paragraphs.append(block)

    if not all_paragraphs:
        ap.error("No text content in .txt files (or only one page when --first-page-is-cover).")

    chapters_per_part = max(0, args.chapters_per_part)
    if chapters_per_part <= 0:
        # One chapter for the whole book
        html_blocks = ["<p>" + html_escape(p).replace("\n", "</p><p>") + "</p>" for p in all_paragraphs]
        body = "\n".join(html_blocks)
        ch = epub.EpubHtml(title="Contents", file_name="content.xhtml", lang="en")
        ch.content = f"""<?xml version='1.0' encoding='utf-8'?>
<html xmlns="http://www.w3.org/1999/xhtml">
<head><title>Contents</title></head>
<body>{body}</body>
</html>"""
        book.add_item(ch)
        book.toc = (ch,)
        book.add_item(epub.EpubNcx())
        book.add_item(epub.EpubNav())
        book.spine = ["nav", ch]
    else:
        # Multiple chapters (e.g. every N pages)
        spine = ["nav"]
        toc_list = []
        for start in range(0, len(all_paragraphs), chapters_per_part):
            part = all_paragraphs[start : start + chapters_per_part]
            part_num = start // chapters_per_part + 1
            html_blocks = ["<p>" + html_escape(p).replace("\n", "</p><p>") + "</p>" for p in part]
            body = "\n".join(html_blocks)
            ch = epub.EpubHtml(
                title=f"Part {part_num}",
                file_name=f"part_{part_num:04d}.xhtml",
                lang="en",
            )
            ch.content = f"""<?xml version='1.0' encoding='utf-8'?>
<html xmlns="http://www.w3.org/1999/xhtml">
<head><title>Part {part_num}</title></head>
<body>{body}</body>
</html>"""
            book.add_item(ch)
            toc_list.append(ch)
            spine.append(ch)
        book.toc = tuple(toc_list)
        book.add_item(epub.EpubNcx())
        book.add_item(epub.EpubNav())
        book.spine = spine

    out_path = args.output.resolve()
    epub.write_epub(str(out_path), book)
    print(f"Wrote: {out_path}")


if __name__ == "__main__":
    main()