Add pipeline code

_pipeline/license.py

@@ -0,0 +1,40 @@
+"""License status by code. Published 1930 or earlier = US public domain as of 2026. Anything else is unknown."""
+import datetime
+import os
+import re
+
+import pdftext
+
+PATTERNS = [r"first\s+published\s+in\s+(\d{4})", r"(?:©|copyright|\(c\))\s*(\d{4})"]
+
+
+def find_year(units):
+    """Earliest year the book's own front pages give as first publication or copyright: (year, where, text) or None."""
+    for pat in PATTERNS:
+        for u in units[:8]:
+            m = re.search(pat, u["text"], re.I)
+            if m:
+                return int(m.group(1)), "pdf p.%d" % u["pdf_page"], m.group(0)
+    return None
+
+
+def run(cfg, book_dir, book, pdf):
+    found = find_year(pdftext.units(pdf))
+    limit = cfg["license"]["public_domain_max_year"]
+    if found is None:
+        status, reason, year = "unknown", "no publication year found in the book's own text", None
+    else:
+        year, where, quote = found
+        if year <= limit:
+            status = "public domain"
+            reason = "published %d (book text, %s: \"%s\"), %d or earlier, so US public domain as of 2026" % (year, where, quote, limit)
+        else:
+            status = "unknown"
+            reason = "published %d (book text, %s: \"%s\"), later than %d, so the date rule does not clear it; confirm before sharing" % (year, where, quote, limit)
+    today = datetime.date.today().isoformat()
+    book["published_year"] = year
+    book["license"] = {"status": status, "reason": reason, "checked": today}
+    with open(os.path.join(book_dir, "LICENSE_STATUS.md"), "w", encoding="utf-8") as f:
+        f.write("# License status\n\nStatus: %s\n\nReason: %s\n\nDate checked: %s\n\n"
+                "Rule: published %d or earlier is public domain in the US as of 2026; anything else is unknown until confirmed. "
+                "Status never blocks audio production. Nothing is published.\n" % (status, reason, today, limit))