From a158a696337ecbfb0da73ee079a9a117876887e7 Mon Sep 17 00:00:00 2001
From: Wolfvin <wolfvin@users.noreply.github.com>
Date: Wed, 1 Jul 2026 03:46:37 +0000
Subject: [PATCH] =?UTF-8?q?fix(graph):=20dedupe=20node=5Frows=20+=20use=20?=
 =?UTF-8?q?INSERT=20OR=20REPLACE=20to=20avoid=20UNIQUE=20constraint=20fail?=
 =?UTF-8?q?ures=20=E2=80=94=20closes=20#114?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Bug: populate_graph_tables() in scripts/graph_model.py built node_rows
from flat_nodes, which can contain duplicate node_id values (file:line:fn)
when e.g. a class and its __init__ method share the same line, or when
two declarations collide on the same line number. The graph_nodes.node_id
column has a UNIQUE constraint, so a bare INSERT raised
sqlite3.IntegrityError and rolled back the entire BEGIN EXCLUSIVE
transaction — leaving graph_nodes empty.

Fix applies two layers of defense:

1. Dedupe node_rows by node_id (last occurrence wins, matching the
   'last match wins' semantics of the flat registry) before the batch
   insert. Applied to both populate_graph_tables() and
   incremental_graph_update().

2. Use INSERT OR REPLACE instead of bare INSERT so even if a duplicate
   slips through (e.g. from a prior partial write), the batch succeeds
   instead of failing the whole transaction.

Verified: scan on clean_app fixture now consistently populates graph_nodes
(31 nodes) and graph_edges (113 edges) without IntegrityError. Test
suite: 1183 passed, 1 pre-existing failure (#118, unrelated).
---
 scripts/graph_model.py | 28 ++++++++++++++++++++++++++--
 1 file changed, 26 insertions(+), 2 deletions(-)

diff --git a/scripts/graph_model.py b/scripts/graph_model.py
index 638f268e..105d5627 100644
--- a/scripts/graph_model.py
+++ b/scripts/graph_model.py
@@ -325,6 +325,22 @@ def populate_graph_tables(workspace: str, db_path: Optional[str] = None) -> Dict
             src_file, src_line, confidence, extra_json,
         ))
 
+    # ── Issue #114: dedupe node_rows by node_id.
+    # The flat registry can produce multiple nodes with the same
+    # ``node_id`` (file:line:fn) when e.g. a class and its __init__
+    # method share the same line, or when two declarations collide on
+    # the same line number. The graph_nodes.node_id column has a UNIQUE
+    # constraint, so a bare INSERT would raise sqlite3.IntegrityError
+    # and roll back the entire batch — leaving the graph empty.
+    # Dedupe by node_id (last occurrence wins, matching the "last match
+    # wins" semantics of the flat registry) before the batch insert.
+    if node_rows:
+        seen: Dict[str, Tuple[Any, ...]] = {}
+        for row in node_rows:
+            # row[0] is node_id
+            seen[row[0]] = row
+        node_rows = list(seen.values())
+
     # ── Issue #10: RAM-first indexing — batch write in a single
     # ``BEGIN EXCLUSIVE`` transaction. The node_rows + edge_rows are
     # collected fully in memory (above) BEFORE opening the transaction;
@@ -351,8 +367,10 @@ def populate_graph_tables(workspace: str, db_path: Optional[str] = None) -> Dict
         )
         conn.execute("DELETE FROM {}".format(GRAPH_NODES_TABLE))
         if node_rows:
+            # Issue #114: use INSERT OR REPLACE as defense-in-depth
+            # against UNIQUE constraint violations even after dedup.
             conn.executemany(
-                "INSERT INTO {t} (node_id, node_type, name, file, line, extra_json) "
+                "INSERT OR REPLACE INTO {t} (node_id, node_type, name, file, line, extra_json) "
                 "VALUES (?, ?, ?, ?, ?, ?)".format(t=GRAPH_NODES_TABLE),
                 node_rows,
             )
@@ -669,9 +687,15 @@ def incremental_graph_update(
         )
 
         # ── Step 4 (insert): Re-insert nodes from changed files ──
+        # Issue #114: dedupe + INSERT OR REPLACE to avoid UNIQUE
+        # constraint violations on duplicate node_ids.
         if node_rows:
+            _seen_inc: Dict[str, Tuple[Any, ...]] = {}
+            for _row in node_rows:
+                _seen_inc[_row[0]] = _row
+            node_rows = list(_seen_inc.values())
             conn.executemany(
-                "INSERT INTO {t} (node_id, node_type, name, file, line, extra_json) "
+                "INSERT OR REPLACE INTO {t} (node_id, node_type, name, file, line, extra_json) "
                 "VALUES (?, ?, ?, ?, ?, ?)".format(t=GRAPH_NODES_TABLE),
                 node_rows,
             )