fix(watch): preserve semantic edges of re-extracted sources (#1865)

graphify/extract.py

@@ -4874,9 +4874,14 @@ def _has_import_evidence(candidate_id: str) -> bool:
     # Tag AST provenance so the incremental watch rebuild can distinguish
     # AST-extracted nodes from semantic/LLM nodes. On a full re-extraction
     # the watcher drops any AST-marked node missing from the fresh output
-    # even when its source file still exists (#1116).
+    # even when its source file still exists (#1116). Edges carry the same
+    # marker so edge eviction can be tier-scoped: re-extracting a source
+    # replaces its AST edges without evicting the semantic edges the AST
+    # pass cannot regenerate (#1865).
     for n in all_nodes:
         n["_origin"] = "ast"
+    for e in all_edges:
+        e["_origin"] = "ast"
 
     return {
         "nodes": all_nodes,

graphify/watch.py

@@ -408,11 +408,11 @@ def _reconcile_existing_graph(
         }
         node_evicted_source_identities = set(deleted_source_identities)
         hyperedge_evicted_source_identities = set(deleted_source_identities)
+        # Deletion evicts edges regardless of tier; re-extraction only owns a
+        # source's AST-tier edges (checked per-edge below, #1865).
+        edge_evicted_source_identities = set(deleted_source_identities)
         if not full_rebuild:
             node_evicted_source_identities.update(rebuilt_source_identities)
-        edge_evicted_source_identities = (
-            node_evicted_source_identities | rebuilt_source_identities
-        )
 
         # Reconcile every rebuild against the current watched corpus. Hook change
         # lists can contain only a rename destination, so explicit paths alone
@@ -485,14 +485,22 @@ def _reconcile_existing_graph(
         ]
         all_ids = new_ast_ids | {node["id"] for node in preserved_nodes}
 
-        # Edges are owned by source_file. Re-extraction must replace an owner's
-        # previous edges, while edges from unchanged or semantic sources survive.
+        # Edges are owned by source_file, but ownership is tier-scoped: the AST
+        # pass replaces a re-extracted source's AST edges, while that source's
+        # semantic/LLM edges — which the AST pass cannot regenerate — survive
+        # until a semantic re-extraction supersedes them. Same provenance rule
+        # the node reconciliation above applies via _origin (#1865). Deletion
+        # eviction stays provenance-blind.
         preserved_edges = [
             edge
             for edge in existing.get("links", existing.get("edges", []))
             if edge.get("source") in all_ids
             and edge.get("target") in all_ids
             and not source_paths.is_evicted(edge, edge_evicted_source_identities)
+            and not (
+                edge.get("_origin") == "ast"
+                and source_paths.is_evicted(edge, rebuilt_source_identities)
+            )
         ]
 
         new_hyperedge_ids = {

tests/test_watch.py

@@ -327,6 +327,72 @@ def test_rebuild_code_preserves_hyperedges_for_rebuilt_surviving_source(
     }]
 
 
+@pytest.mark.parametrize(
+    "changed_paths",
+    [None, [Path("auth.md")]],
+    ids=["full-update", "incremental-doc-update"],
+)
+def test_rebuild_code_preserves_semantic_edges_from_reextracted_doc(
+    tmp_path, changed_paths
+):
+    """#1865: AST-only updates must not evict semantic edges whose source_file
+    is a re-extracted document; only that source's AST-tier edges are replaced."""
+    from graphify.watch import _rebuild_code
+
+    corpus = tmp_path / "corpus"
+    corpus.mkdir()
+    (corpus / "auth.md").write_text(
+        "# Token Validation\n\nVerifies bearer tokens.\n", encoding="utf-8"
+    )
+    (corpus / "login.md").write_text(
+        "# Session Verification\n\nVerifies login sessions.\n", encoding="utf-8"
+    )
+
+    assert _rebuild_code(corpus, no_cluster=True, acquire_lock=False) is True
+    graph_path = corpus / "graphify-out" / "graph.json"
+    data = json.loads(graph_path.read_text(encoding="utf-8"))
+    node_ids = {n["id"] for n in data["nodes"]}
+    assert {"auth_token_validation", "login_session_verification"} <= node_ids
+
+    data["links"].extend([
+        {
+            "source": "auth_token_validation",
+            "target": "login_session_verification",
+            "relation": "semantically_similar_to",
+            "confidence": "INFERRED",
+            "source_file": "auth.md",
+        },
+        # A stale AST-tier edge of the same source must still be evicted.
+        {
+            "source": "auth_token_validation",
+            "target": "login_session_verification",
+            "relation": "references",
+            "_origin": "ast",
+            "source_file": "auth.md",
+        },
+    ])
+    graph_path.write_text(json.dumps(data), encoding="utf-8")
+
+    assert _rebuild_code(
+        corpus,
+        changed_paths=changed_paths,
+        no_cluster=True,
+        acquire_lock=False,
+    ) is True
+
+    after = json.loads(graph_path.read_text(encoding="utf-8"))
+    relations = {
+        (e.get("source"), e.get("target"), e.get("relation"))
+        for e in after["links"]
+    }
+    assert (
+        "auth_token_validation", "login_session_verification", "semantically_similar_to"
+    ) in relations, "semantic edge from a re-extracted doc must survive an AST-only update"
+    assert (
+        "auth_token_validation", "login_session_verification", "references"
+    ) not in relations, "stale AST-tier edge of a re-extracted source must be evicted"
+
+
 @pytest.mark.parametrize(
     "changed_paths",
     [None, [Path("only.py")]],