Update paperworks handling of revisions

Author: vinniefalco

paperworks/lib/inventory.py

@@ -237,27 +237,38 @@ def build_inventory(watch_dirs, output_dir, remote_papers=None):
         if base:
             md_by_base[f"{base}R{rev}"] = v
 
-    pdf_by_base = {}
+    # PDF filenames are revision-ambiguous; index by base only
+    pdf_by_base_only = {}
     for k, v in pdf_papers.items():
         _, base, rev = _parse_doc_number(k)
         if base:
-            pdf_by_base[f"{base}R{rev}"] = v
+            pdf_by_base_only[base] = v
 
     all_keys = set()
     all_keys.update(md_by_base.keys())
-    all_keys.update(pdf_by_base.keys())
     all_keys.update(remote_by_base.keys())
 
+    # Add PDF-only keys (orphans with no markdown or remote)
+    bases_covered = {
+        _parse_doc_number(k)[1] for k in all_keys
+        if _parse_doc_number(k)[1]
+    }
+    for k, v in pdf_papers.items():
+        _, base, rev = _parse_doc_number(k)
+        if base and base not in bases_covered:
+            all_keys.add(f"{base}R{rev}")
+
     records = {}
     for key in all_keys:
         md = md_by_base.get(key)
-        pdf = pdf_by_base.get(key)
         remote = remote_by_base.get(key)
 
         full, base, rev = _parse_doc_number(key)
         if not base:
             continue
 
+        pdf = pdf_by_base_only.get(base)
+
         # Markdown is source of truth for metadata
         title = (md or {}).get("title") or (pdf or {}).get("title") or (remote or {}).get("title", "")
         authors = (md or {}).get("authors") or (pdf or {}).get("authors") or (remote or {}).get("author", "")

paperworks/lib/pdf_reader.py

@@ -15,7 +15,7 @@
 )
 
 _DOC_FIELD_RE = re.compile(
-    r"Document\s+Number[:\s]+([DPN]\d{3,5}(?:R\d+)?|N\d{3,5})",
+    r"Document(?:\s+Number)?[:\s]+([DPN]\d{3,5}(?:R\d+)?|N\d{3,5})",
     re.IGNORECASE,
 )
 

paperworks/tests/test_inventory.py

@@ -0,0 +1,82 @@
+"""Tests for inventory.build_inventory merge logic."""
+
+import unittest
+from unittest.mock import patch
+
+
+def _md(doc, title="Test", rev=0, base="P9999"):
+    return {
+        "doc_number": doc, "base": base, "revision": rev,
+        "title": title, "authors": "A", "date": "2026-01-01",
+        "audience": "LEWG", "intent": "ask", "brutal_summary": "Test.",
+        "md_path": f"/src/{doc.lower()}.md", "md_mtime": 1000.0,
+        "folder_idx": 1,
+    }
+
+
+def _pdf(doc, rev=0, base="P9999"):
+    return {
+        "doc_number": doc, "base": base, "revision": rev,
+        "title": "", "authors": "", "date": "", "audience": "",
+        "brutal_summary": None,
+        "pdf_path": f"/out/{doc.lower()}.pdf", "pdf_mtime": 2000.0,
+    }
+
+
+class TestBuildInventory(unittest.TestCase):
+
+    def _build(self, md_papers, pdf_papers, remote_papers=None):
+        with patch("lib.inventory.scan_markdown_dirs", return_value=md_papers), \
+             patch("lib.inventory.scan_pdf_dir", return_value=pdf_papers):
+            from lib.inventory import build_inventory
+            return build_inventory([], "/out", remote_papers)
+
+    def _find(self, papers, base):
+        return next((p for p in papers if p["base"] == base), None)
+
+    def test_pdf_matches_markdown_at_r0(self):
+        """Baseline: R0 markdown + R0-keyed PDF merge correctly."""
+        md = {"D4035R0": _md("D4035R0", rev=0, base="P4035")}
+        pdf = {"D4035R0": _pdf("D4035R0", rev=0, base="P4035")}
+        result = self._build(md, pdf)
+        p = self._find(result, "P4035")
+        self.assertIsNotNone(p)
+        self.assertIsNotNone(p["md_path"])
+        self.assertIsNotNone(p["pdf_path"])
+
+    def test_pdf_matches_markdown_at_r2(self):
+        """The bug: filename-derived R0 PDF must match R2 markdown."""
+        md = {"P4003R2": _md("P4003R2", rev=2, base="P4003")}
+        # PDF keyed as R0 (filename fallback: d4003-io-awaitables.pdf -> D4003 -> R0)
+        pdf = {"D4003R0": _pdf("D4003R0", rev=0, base="P4003")}
+        result = self._build(md, pdf)
+        p = self._find(result, "P4003")
+        self.assertIsNotNone(p)
+        self.assertEqual(p["revision"], 2)
+        self.assertIsNotNone(p["md_path"])
+        self.assertIsNotNone(p["pdf_path"])
+
+    def test_orphan_pdf_still_appears(self):
+        """PDF with no markdown or remote is still in the inventory."""
+        pdf = {"D4099R0": _pdf("D4099R0", rev=0, base="P4099")}
+        result = self._build({}, pdf)
+        p = self._find(result, "P4099")
+        self.assertIsNotNone(p)
+        self.assertIsNotNone(p["pdf_path"])
+        self.assertIsNone(p["md_path"])
+
+    def test_pdf_plus_remote_at_non_r0(self):
+        """PDF matched by base when remote provides the authoritative revision."""
+        pdf = {"D4007R0": _pdf("D4007R0", rev=0, base="P4007")}
+        remote = [{"doc_number": "P4007R2", "title": "Open Issues",
+                    "author": "A", "status": "Draft", "date": "2026-04-08",
+                    "form_id": "99", "form_url": "https://isocpp.org/papers/form/99"}]
+        result = self._build({}, pdf, remote)
+        p = self._find(result, "P4007")
+        self.assertIsNotNone(p)
+        self.assertEqual(p["revision"], 2)
+        self.assertIsNotNone(p["pdf_path"])
+
+
+if __name__ == "__main__":
+    unittest.main()

paperworks/tests/test_pdf_reader.py

@@ -25,3 +25,12 @@ def test_case_insensitive_field(self):
     def test_n_paper_structured(self):
         text = "Document Number: N4950\nWorking Draft"
         assert _extract_doc_number(text) == "N4950"
+
+    def test_single_word_document_label(self):
+        """Scrivener wg21 style renders 'Document:' not 'Document Number:'."""
+        text = "Document: P4003R2\nDate: 2026-04-05"
+        assert _extract_doc_number(text) == "P4003R2"
+
+    def test_single_word_document_label_no_revision(self):
+        text = "Document: D4035\nDate: 2026-03-20"
+        assert _extract_doc_number(text) == "D4035"