wikifier 4.6.0__tar.gz → 4.6.1__tar.gz

{wikifier-4.6.0/wikifier.egg-info → wikifier-4.6.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: wikifier
-Version: 4.6.0
+Version: 4.6.1
 Summary: Zero-dependency agent-to-agent codebase wiki for LLMs and AI agents. Autonomously maintained via record-change and mark-green for token-efficient lookup of files, dependencies, health, and summaries — across tiny scripts to 50k+ monorepos. Optional MCP server with rich tools for agents.
 Author-email: Aron Amos <aron@example.com>
 Maintainer: Aron Amos

{wikifier-4.6.0 → wikifier-4.6.1}/tests/test_agent_loop.py

@@ -78,6 +78,52 @@ class TestContentHonestDirty(TempProjectTestCase):
         self.assertIn("Yellow", data3["entries"]["mod.py"]["status"])
         self.assertIn("content", data3["entries"]["mod.py"]["reason"].lower())
 
+    def test_legacy_green_without_hash_rewrite_yellows(self):
+        """Criterion 2: Green with no source_content_hash + real rewrite must Yellow.
+
+        Must not seed post-edit hash and stay Green (post-upgrade / legacy path).
+        """
+        p = self.write("legacy.py", "print('v1')\n")
+        # Legacy Green: upsert only — no source_content_hash baseline
+        health_mod.upsert_entry(self.root, "legacy.py", "🟢 Green", "legacy green no hash")
+        data = health_mod.load_health(self.root)
+        self.assertNotIn("source_content_hash", data["entries"]["legacy.py"] or {})
+        from wikifier import import_cache as ic
+        cache = {}
+        ic.update_file_data(cache, "legacy.py", mtime=1, imports=[], resolved_pairs=[])
+        ic.save_cache(self.root, cache)
+        # Real content rewrite
+        p.write_text("print('v2-rewritten')\n", encoding="utf-8")
+        cache = ic.load_cache(self.root)
+        cache["legacy.py"]["mtime"] = 1  # force dirty vs disk
+        ic.save_cache(self.root, cache)
+        r = cli.check_changes(project_root=self.root)
+        self.assertTrue(r.get("success"), r)
+        self.assertGreaterEqual(int(r.get("changes_detected") or 0), 1)
+        data2 = health_mod.load_health(self.root)
+        ent = data2["entries"]["legacy.py"]
+        self.assertIn("Yellow", ent["status"])
+        # Must not have silently stored the post-edit hash as trusted baseline
+        self.assertNotEqual(
+            ent.get("source_content_hash"),
+            compute_source_content_hash(p),
+            "must not seed post-edit content hash without mark_green",
+        )
+        # After mark_green, baseline is set and further touch-only stays green
+        cli.mark_green("legacy.py", "re-baselined", project_root=self.root)
+        data3 = health_mod.load_health(self.root)
+        self.assertIn("source_content_hash", data3["entries"]["legacy.py"])
+        baseline = data3["entries"]["legacy.py"]["source_content_hash"]
+        cache = ic.load_cache(self.root)
+        cache["legacy.py"]["mtime"] = 1
+        ic.save_cache(self.root, cache)
+        os.utime(p, None)
+        r2 = cli.check_changes(project_root=self.root)
+        self.assertTrue(r2.get("success"), r2)
+        data4 = health_mod.load_health(self.root)
+        self.assertIn("Green", data4["entries"]["legacy.py"]["status"])
+        self.assertEqual(data4["entries"]["legacy.py"]["source_content_hash"], baseline)
+
 
 class TestSessionBootstrapAndSuggest(TempProjectTestCase):
     def test_session_bootstrap_shape(self):

{wikifier-4.6.0 → wikifier-4.6.1}/wikifier/__init__.py

@@ -76,4 +76,4 @@ from .contracts import (
     compute_acs_confidence,
 )
 
-__version__ = "4.6.0"
+__version__ = "4.6.1"

{wikifier-4.6.0 → wikifier-4.6.1}/wikifier/cli.py

@@ -1208,48 +1208,34 @@ def check_changes(project_root: Optional[Union[str, Path]] = None) -> Dict[str,
                             if stored_hash and live and stored_hash == live:
                                 verdict = {"content_dirty": False, "reason": "content_unchanged", "seed_baseline": False, "hash": live}
                             elif not stored_hash and live:
-                                verdict = {"content_dirty": False, "reason": "no_baseline", "seed_baseline": True, "hash": live}
+                                # no baseline → dirty (do not seed post-edit hash)
+                                verdict = {"content_dirty": True, "reason": "no_baseline", "seed_baseline": False, "hash": live}
                             elif live and stored_hash and stored_hash != live:
                                 verdict = {"content_dirty": True, "reason": "content_changed", "seed_baseline": False, "hash": live}
                         except Exception:
                             pass
 
-                    if verdict.get("seed_baseline") and verdict.get("hash") and isinstance(ent, dict):
-                        ent["source_content_hash"] = verdict["hash"]
-                        entries[rel] = ent
-                        health_dirty = True
-                        seeded_baselines += 1
-                        # Do not Yellow solely for first baseline seed on mtime thrash
-                        continue
                     if not verdict.get("content_dirty") and verdict.get("reason") == "content_unchanged":
                         skipped_mtime_only += 1
                         continue
-                    # Content changed (or unclassifiable): Yellow with content reason
+                    # Content changed, no baseline, or unclassifiable: Yellow.
+                    # Never write source_content_hash here — only mark_green sets the
+                    # trusted baseline (avoids seeding post-edit bytes and staying Green).
                     reason = (
                         "content changed since last trusted baseline (check_changes content-honest)"
                         if verdict.get("reason") == "content_changed"
                         else "content change or no baseline (check_changes content-honest auto-detect)"
                     )
                     _upsert(root, rel, "🟡 Yellow", reason)
-                    # refresh local entries view after upsert
                     try:
                         health_data = _health_mod.load_health(root)
                         entries = health_data.setdefault("entries", {})
-                        if verdict.get("hash") and rel in entries and isinstance(entries[rel], dict):
-                            # keep prior baseline until mark-green; do not update hash on yellow
-                            pass
                         health_dirty = False  # upsert already saved
                     except Exception:
                         pass
                     _add_to_pending(root, rel, "Content change auto-detected — review and run mark-green after wiki update")
                     _ensure_journal_entry(root, "auto-detected", rel, reason)
                     changed_count += 1
-                if health_dirty and health_data is not None:
-                    try:
-                        if hasattr(_health_mod, "save_health"):
-                            _health_mod.save_health(root, health_data)
-                    except Exception:
-                        pass
                 # Keep pending queue aligned with lean monitored_paths (no flood outside scope)
                 if hasattr(_health_mod, "prune_pending_to_monitored"):
                     try:

{wikifier-4.6.0 → wikifier-4.6.1}/wikifier/health.py

@@ -272,12 +272,16 @@ def classify_content_dirty(
             "seed_baseline": False,
         }
     if not stored_hash:
+        # No trusted baseline (legacy Green / never mark-green): when this path is
+        # under check_changes dirty observation, treat as needing attention so a
+        # post-edit rewrite cannot silently re-seed the *new* bytes and stay Green.
+        # Callers that only want to migrate baselines should use an explicit seed helper.
         return {
-            "content_dirty": False,
+            "content_dirty": True,
             "missing": False,
             "hash": live,
             "reason": "no_baseline",
-            "seed_baseline": True,
+            "seed_baseline": False,
         }
     if stored_hash == live:
         return {

{wikifier-4.6.0 → wikifier-4.6.1/wikifier.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: wikifier
-Version: 4.6.0
+Version: 4.6.1
 Summary: Zero-dependency agent-to-agent codebase wiki for LLMs and AI agents. Autonomously maintained via record-change and mark-green for token-efficient lookup of files, dependencies, health, and summaries — across tiny scripts to 50k+ monorepos. Optional MCP server with rich tools for agents.
 Author-email: Aron Amos <aron@example.com>
 Maintainer: Aron Amos