Workflow: sparse partial checkout (code and _incoming only), book lookup before processing, safe rebase retry; revert cancel-in-progress

_pipeline/ingest.py

@@ -40,14 +40,23 @@ def author_slug(author):
     return bookio.slug(" ".join([parts[-1]] + parts[:-1])) if parts else "unknown"
 
 
+def book_folder(cfg, author, title):
+    return os.path.join(settings.resolve(cfg, "books_dir"), author_slug(author), bookio.slug(title))
+
+
+def detect(pdf, title=None, author=None):
+    """Title and author from the PDF itself (title: PDF metadata; author: a 'By <Name>' line on the first pages)."""
+    author = author or find_author([u["text"] for u in pdftext.units(pdf, first_pages=6)])
+    title = title or find_title(pdf)
+    return title, author
+
+
 def run(cfg, pdf, title=None, author=None):
     """Returns the book folder path."""
-    pdf_units = pdftext.units(pdf)
-    author = author or find_author([u["text"] for u in pdf_units[:6]])
-    title = title or find_title(pdf)
+    title, author = detect(pdf, title, author)
     if not author or not title:
         raise SystemExit("could not find title/author in the PDF; pass --title and --author")
-    book_dir = os.path.join(settings.resolve(cfg, "books_dir"), author_slug(author), bookio.slug(title))
+    book_dir = book_folder(cfg, author, title)
     os.makedirs(book_dir, exist_ok=True)
     copy = os.path.join(book_dir, os.path.basename(pdf))
     if os.path.exists(copy) and sha256(copy) != sha256(pdf):