fix(enrich): explicit QUOTE/ESCAPE so Kuzu COPY honors RFC-4180

PR #150 switched the staging file delimiter to '|' to avoid JSON-
property comma collisions. That fixes comma-bearing values but
breaks when an ID itself contains a literal '|' — Istio's EDS
cluster names are exactly this shape:

    json:istio/none_cds.json:inbound|7070|tcplocal|s1tcp.none

Go's encoding/csv writer DOES wrap such fields in '"' per RFC-4180.
But Kuzu's CSV reader defaults to BACKSLASH escaping, not the
RFC-4180 doubled-quote form Go produces. With the default Kuzu
escape rule, the pipe-bearing quoted field is parsed as multiple
fields and the COPY aborts:

    Copy exception: Error in file ... expected 6 values per row,
    but got more.

Fix: pass `QUOTE='"', ESCAPE='"'` explicitly so Kuzu interprets
the RFC-4180 form Go writes. Applies to both copyNodeBatch and
copyEdgeBatch.

End-to-end: `codeiq enrich ~/projects/polyglot-bench/istio` now
exits 0 (was exit 2 pre-fix): 36k nodes, 55k edges, 20 services.

Regression test TestBulkLoadEdgesPipeInTargetID covers the exact
Istio cluster-name shape.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Author: aksOps

go/internal/graph/bulk.go

@@ -101,9 +101,17 @@ func (s *Store) copyNodeBatch(batch []*model.CodeNode) error {
 
 	// Kuzu COPY FROM with explicit column list. ToSlash for Windows path
 	// portability — Kuzu's parser accepts forward slashes on all platforms.
-	// DELIM='|' matches the pipe-separated staging file written above.
+	//
+	// DELIM='|' matches the pipe-separated staging file written above. The
+	// explicit QUOTE/ESCAPE pair overrides Kuzu's default backslash-escape
+	// behaviour with RFC-4180 (doubled-quote) escaping so that Go's
+	// encoding/csv writer (which emits "field""with""quotes" form) round-
+	// trips correctly. Fields containing the delimiter (e.g. Istio service
+	// names like "inbound|7070|tcplocal|s1tcp.none") are wrapped by the Go
+	// writer; Kuzu then dequotes them only when the matching escape rule is
+	// set.
 	q := fmt.Sprintf(
-		"COPY CodeNode(%s) FROM '%s' (header=false, DELIM='|')",
+		`COPY CodeNode(%s) FROM '%s' (header=false, DELIM='|', QUOTE='"', ESCAPE='"')`,
 		strings.Join(nodeColumns, ", "),
 		filepath.ToSlash(tmp.Name()),
 	)
@@ -263,9 +271,10 @@ func (s *Store) copyEdgeBatch(kind model.EdgeKind, batch []*model.CodeEdge) erro
 		return fmt.Errorf("graph: csv close: %w", err)
 	}
 
-	// DELIM='|' matches the pipe-separated staging file written above.
+	// DELIM/QUOTE/ESCAPE — see copyNodeBatch for the rationale (RFC-4180
+	// round-trip with Go's encoding/csv).
 	q := fmt.Sprintf(
-		"COPY %s FROM '%s' (header=false, DELIM='|')",
+		`COPY %s FROM '%s' (header=false, DELIM='|', QUOTE='"', ESCAPE='"')`,
 		relTableName(kind),
 		filepath.ToSlash(tmp.Name()),
 	)

go/internal/graph/bulk_test.go

@@ -223,6 +223,52 @@ func TestBulkLoadNodesCommaInProperties(t *testing.T) {
 	}
 }
 
+// TestBulkLoadEdgesPipeInTargetID is a regression test for Istio-style IDs
+// that contain the field delimiter '|' literally (e.g. EDS cluster names
+// "inbound|7070|tcplocal|s1tcp.none" parsed from JSON config). Go's csv.Writer
+// RFC-4180-wraps such fields in '"', but Kuzu's default ESCAPE is backslash
+// not doubled-quote — so without explicit QUOTE='"', ESCAPE='"' the COPY
+// FROM splits the wrapped field on each interior '|' and aborts with
+// "expected N values per row, but got more". Fix: explicit QUOTE/ESCAPE in
+// the COPY FROM clause.
+func TestBulkLoadEdgesPipeInTargetID(t *testing.T) {
+	s, err := graph.Open(filepath.Join(t.TempDir(), "g.kuzu"))
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer s.Close()
+	if err := s.ApplySchema(); err != nil {
+		t.Fatal(err)
+	}
+	// Istio-flavoured target ID with literal pipes.
+	target := "json:istio/none_cds.json:inbound|7070|tcplocal|s1tcp.none"
+	nodes := []*model.CodeNode{
+		{ID: "json:istio/none_cds.json", Kind: model.NodeModule, Label: "none_cds.json"},
+		{ID: target, Kind: model.NodeConfigKey, Label: "inbound|7070|tcplocal|s1tcp.none"},
+	}
+	if err := s.BulkLoadNodes(nodes); err != nil {
+		t.Fatalf("BulkLoadNodes with pipe-bearing ID: %v", err)
+	}
+	edges := []*model.CodeEdge{{
+		ID:         "json:istio/none_cds.json->" + target,
+		Kind:       model.EdgeContains,
+		SourceID:   "json:istio/none_cds.json",
+		TargetID:   target,
+		Confidence: model.ConfidenceSyntactic,
+		Source:     "JsonStructureDetector",
+	}}
+	if err := s.BulkLoadEdges(edges); err != nil {
+		t.Fatalf("BulkLoadEdges with pipe-bearing target ID: %v", err)
+	}
+	rows, err := s.Cypher("MATCH ()-[r:CONTAINS]->() RETURN r.id AS id")
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(rows) != 1 {
+		t.Fatalf("want 1 CONTAINS row, got %d: %v", len(rows), rows)
+	}
+}
+
 // TestBulkLoadEdgesEmpty — zero edges is a no-op like the node path.
 func TestBulkLoadEdgesEmpty(t *testing.T) {
 	s, err := graph.Open(filepath.Join(t.TempDir(), "g.kuzu"))