Fix LadybugDB migration issues and JSON parsing

ast_java.py

@@ -325,7 +325,7 @@ class RouteDecl:
     filename: str
     start_line: int
     end_line: int
-    # brownfield / B2a composition (graph_enrich.resolve_routes_for_method); not a Kuzu column.
+    # brownfield / B2a composition (graph_enrich.resolve_routes_for_method); not a graph column.
     route_source_layer: str = "builtin"
 
 

graph_enrich.py

@@ -334,7 +334,7 @@ def collect_annotation_meta_chain(
 ) -> dict[str, frozenset[str]]:
     """Map annotation simple name → built-in simple names reachable via meta-annotations.
 
-    Single source of truth for Layer A: both the Kuzu writer and Lance chunk
+    Single source of truth for Layer A: both the LadybugDB writer and Lance chunk
     enrichment must use this; they must not derive `meta_chain` from separate
     filesystem walks. See ``PLAN-BROWNFIELD-ROLE-OVERRIDES`` §
     *Single source of truth (REQUIRED — read before implementation)*.
@@ -350,7 +350,7 @@ def annotation_meta_decls_from_graph_tables(
     """From `build_ast_graph.GraphTables.types`, map @interface simple name -> meta anns.
 
     Used for diagnostics; Layer A in production uses `collect_annotation_meta_chain`
-    (disk) so Kuzu and Lance share one index.
+    (disk) so LadybugDB and Lance share one index.
     """
     decls: dict[str, tuple[str, ...]] = {}
     first_fqn: dict[str, str] = {}
@@ -1702,7 +1702,7 @@ def enrich_chunk(
 
 
 def symbol_id(kind: str, fqn: str, file_path: str = "", start_byte: int = 0) -> str:
-    """Deterministic SHA1-based id for Kuzu Symbol nodes."""
+    """Deterministic SHA1-based id for LadybugDB Symbol nodes."""
     key = f"{kind}|{fqn}|{file_path}|{start_byte}".encode("utf-8")
     return hashlib.sha1(key).hexdigest()
 

java_codebase_rag/cli.py

@@ -24,10 +24,10 @@
 from java_codebase_rag.pipeline import clip, run_build_ast_graph, run_cocoindex_drop, run_cocoindex_update, run_incremental_graph
 from java_ontology import VALID_UNRESOLVED_CALL_REASONS
 
-KUZU_INCREMENTAL_TRACKING_ISSUE_URL = "https://github.com/HumanBean17/java-codebase-rag/issues/73"
+LADYBUG_INCREMENTAL_TRACKING_ISSUE_URL = "https://github.com/HumanBean17/java-codebase-rag/issues/73"
 
 _INCREMENT_WARNING_LINES = (
-    "WARNING: AST graph (Kuzu) incremental rebuild is not yet implemented.",
+    "WARNING: AST graph (LadybugDB) incremental rebuild is not yet implemented.",
     "The graph reflects the index state from the last `init` or `reprocess`,",
     "which means `find`, `neighbors`, and `describe` may return stale results",
     "for files changed since then.",
@@ -37,8 +37,8 @@
     "For an up-to-date graph, run:",
     "    java-codebase-rag reprocess",
     "",
-    "Track progress on Kuzu incremental rebuild:",
-    f"    {KUZU_INCREMENTAL_TRACKING_ISSUE_URL}",
+    "Track progress on LadybugDB incremental rebuild:",
+    f"    {LADYBUG_INCREMENTAL_TRACKING_ISSUE_URL}",
 )
 
 _REFRESH_DEPRECATION = (
@@ -47,7 +47,7 @@
 )
 
 _REPROCESS_DRIFT_VECTORS_ONLY = (
-    "java-codebase-rag reprocess: rebuilt vectors only; graph (code_graph.kuzu) was NOT rebuilt "
+    "java-codebase-rag reprocess: rebuilt vectors only; graph (code_graph.lbug) was NOT rebuilt "
     "and may now reflect a stale source snapshot."
 )
 
@@ -178,7 +178,7 @@ def _emit(value: Any) -> None:
     print(json.dumps(payload, default=_jsonable, sort_keys=True, indent=None))
 
 
-def _emit_increment_kuzu_warning() -> None:
+def _emit_increment_ladybug_warning() -> None:
     for line in _INCREMENT_WARNING_LINES:
         print(line, file=sys.stderr)
 
@@ -289,7 +289,7 @@ def work() -> int:
             print(file=sys.stderr, flush=True)
         g = run_build_ast_graph(
             source_root=cfg.source_root,
-            kuzu_path=cfg.kuzu_path,
+            ladybug_path=cfg.ladybug_path,
             verbose=verbose,
             quiet=bool(args.quiet),
             env=env,
@@ -319,7 +319,7 @@ def _cmd_increment(args: argparse.Namespace) -> int:
     # Check for --vectors-only flag
     vectors_only = bool(getattr(args, "vectors_only", False))
     if vectors_only:
-        _emit_increment_kuzu_warning()
+        _emit_increment_ladybug_warning()
 
     def work() -> int:
         env = cfg.subprocess_env()
@@ -350,7 +350,7 @@ def work() -> int:
         # Run incremental graph update
         g = run_incremental_graph(
             source_root=cfg.source_root,
-            kuzu_path=cfg.kuzu_path,
+            ladybug_path=cfg.ladybug_path,
             verbose=bool(args.verbose),
             quiet=bool(args.quiet),
             env=env,
@@ -437,7 +437,7 @@ def work() -> int:
         if graph_only:
             g = run_build_ast_graph(
                 source_root=cfg.source_root,
-                kuzu_path=cfg.kuzu_path,
+                ladybug_path=cfg.ladybug_path,
                 verbose=verbose,
                 quiet=bool(args.quiet),
                 env=env,
@@ -509,7 +509,7 @@ def _cmd_erase(args: argparse.Namespace) -> int:
     cfg = _resolved_from_ns(args)
     _startup_hints(cfg)
     cfg.apply_to_os_environ()
-    to_describe: list[Path] = [cfg.kuzu_path, cfg.cocoindex_db]
+    to_describe: list[Path] = [cfg.ladybug_path, cfg.cocoindex_db]
     if cfg.index_dir.is_dir():
         try:
             import lancedb
@@ -546,8 +546,8 @@ def work() -> int:
             )
         elif drop.returncode != 0:
             print(clip(drop.stderr, 4000), file=sys.stderr)
-        if cfg.kuzu_path.exists():
-            shutil.rmtree(cfg.kuzu_path, ignore_errors=True)
+        if cfg.ladybug_path.exists():
+            shutil.rmtree(cfg.ladybug_path, ignore_errors=True)
         if cfg.cocoindex_db.exists():
             try:
                 cfg.cocoindex_db.unlink()
@@ -577,17 +577,17 @@ def _cmd_meta(args: argparse.Namespace) -> int:
     cfg = _resolved_from_ns(args)
     _startup_hints(cfg)
     cfg.apply_to_os_environ()
-    from kuzu_queries import KuzuGraph  # lazy
+    from ladybug_queries import LadybugGraph  # lazy
 
-    KuzuGraph._instance = None
-    KuzuGraph._instance_path = None
+    LadybugGraph._instance = None
+    LadybugGraph._instance_path = None
     payload = server._graph_meta_output().model_dump()
     payload["embedding_model"] = cfg.embedding_model
     payload["embedding_device"] = cfg.embedding_device
     payload["embedding_model_source"] = cfg.embedding_model_source
     payload["embedding_device_source"] = cfg.embedding_device_source
     payload["index_dir"] = str(cfg.index_dir.resolve())
-    payload["kuzu_path"] = str(cfg.kuzu_path.resolve())
+    payload["ladybug_path"] = str(cfg.ladybug_path.resolve())
     payload["index_dir_source"] = cfg.index_dir_source
     payload["hints_enabled"] = cfg.hints_enabled
     payload["hints_enabled_source"] = cfg.hints_enabled_source
@@ -637,12 +637,12 @@ def _cmd_unresolved_calls_list(args: argparse.Namespace) -> int:
     cfg = _resolved_from_ns(args)
     _startup_hints(cfg)
     cfg.apply_to_os_environ()
-    from kuzu_queries import KuzuGraph  # lazy
+    from ladybug_queries import LadybugGraph  # lazy
 
-    if not KuzuGraph.exists():
+    if not LadybugGraph.exists():
         _emit({"success": False, "message": "Kuzu graph not found"})
         return 1
-    graph = KuzuGraph.get()
+    graph = LadybugGraph.get()
     rows = graph.list_unresolved_call_sites(
         method_id=args.method_id,
         reason=args.reason,
@@ -658,12 +658,12 @@ def _cmd_unresolved_calls_stats(args: argparse.Namespace) -> int:
     cfg = _resolved_from_ns(args)
     _startup_hints(cfg)
     cfg.apply_to_os_environ()
-    from kuzu_queries import KuzuGraph  # lazy
+    from ladybug_queries import LadybugGraph  # lazy
 
-    if not KuzuGraph.exists():
+    if not LadybugGraph.exists():
         _emit({"success": False, "message": "Kuzu graph not found"})
         return 1
-    graph = KuzuGraph.get()
+    graph = LadybugGraph.get()
     buckets = graph.stats_unresolved_call_sites(by=args.by)
     total = sum(int(r.get("n") or 0) for r in buckets)
     _emit({"success": True, "total": total, "by": args.by, "buckets": buckets})
@@ -683,12 +683,12 @@ def _cmd_analyze_pr(args: argparse.Namespace) -> int:
         _emit({"success": False, "message": "Diff is empty"})
         return 1
     import pr_analysis  # lazy
-    from kuzu_queries import KuzuGraph  # lazy
+    from ladybug_queries import LadybugGraph  # lazy
 
-    if not KuzuGraph.exists():
+    if not LadybugGraph.exists():
         _emit({"success": False, "message": "Kuzu graph not found"})
         return 1
-    graph = KuzuGraph.get()
+    graph = LadybugGraph.get()
     report = pr_analysis.analyze_pr_pipeline(graph, diff_text)
     _emit(pr_analysis.pr_report_to_dict(report))
     return 0

java_codebase_rag/config.py

@@ -67,10 +67,9 @@ def resolved_sbert_model_for_process_env(import_time_default: str) -> str:
 # Legacy env keys: never honored; detection-only hints name the replacement (if any).
 _LEGACY_ENV_HINTS: tuple[tuple[str, str], ...] = (
     ("LANCEDB_URI", "JAVA_CODEBASE_RAG_INDEX_DIR"),
-    ("KUZU_DB_PATH", "JAVA_CODEBASE_RAG_INDEX_DIR (Kuzu lives at <index_dir>/code_graph.kuzu)"),
     ("LANCEDB_MCP_PROJECT_ROOT", "cwd or --source-root (no env replacement)"),
     ("LANCEDB_MCP_ALLOW_REFRESH", "(removed; use init / increment / reprocess / erase)"),
-    ("LANCEDB_MCP_GRAPH_ENABLED", "(removed; graph is used when code_graph.kuzu exists)"),
+    ("LANCEDB_MCP_GRAPH_ENABLED", "(removed; graph is used when code_graph.lbug exists)"),
     ("LANCEDB_MCP_MICROSERVICE_ROOTS", "microservice_roots: in .java-codebase-rag.yml"),
     ("LANCEDB_MCP_DEBUG_CONTEXT", ENV_DEBUG_CONTEXT),
     ("LANCEDB_MCP_RUN_HEAVY", ENV_RUN_HEAVY),
@@ -182,7 +181,7 @@ def load_yaml_mapping(source_root: Path) -> dict[str, Any]:
 class ResolvedOperatorConfig:
     source_root: Path
     index_dir: Path
-    kuzu_path: Path
+    ladybug_path: Path
     cocoindex_db: Path
     embedding_model: str
     embedding_device: str | None
@@ -193,7 +192,7 @@ class ResolvedOperatorConfig:
     hints_enabled_source: SettingSource
 
     def apply_to_os_environ(self) -> None:
-        """Make downstream modules (server, kuzu_queries, flows) see a consistent environment.
+        """Make downstream modules (server, ladybug_queries, flows) see a consistent environment.
 
         When ``embedding_device`` is unset, ``SBERT_DEVICE`` is not removed from ``os.environ`` so
         a long-lived host process is not mutated for unrelated callers; subprocesses still use
@@ -369,12 +368,12 @@ def resolve_operator_config(
         yaml_path=("hints", "enabled"),
         default=True,
     )
-    ku = index_dir / "code_graph.kuzu"
+    ku = index_dir / "code_graph.lbug"
     coco = index_dir / "cocoindex.db"
     return ResolvedOperatorConfig(
         source_root=root,
         index_dir=index_dir,
-        kuzu_path=ku,
+        ladybug_path=ku,
         cocoindex_db=coco,
         embedding_model=model,
         embedding_device=device,
@@ -387,9 +386,9 @@ def resolve_operator_config(
 
 
 def index_dir_has_existing_artifacts(index_dir: Path) -> tuple[bool, list[str]]:
-    """True if Kuzu graph dir or any Lance table already exists under index_dir."""
+    """True if graph dir or any Lance table already exists under index_dir."""
     paths: list[str] = []
-    ku = index_dir / "code_graph.kuzu"
+    ku = index_dir / "code_graph.lbug"
     if ku.exists():
         paths.append(str(ku.resolve()))
     if index_dir.is_dir():

java_codebase_rag/installer.py

@@ -791,7 +791,7 @@ def run_init_if_needed(
     # Run AST graph build
     g = run_build_ast_graph(
         source_root=cfg.source_root,
-        kuzu_path=cfg.kuzu_path,
+        ladybug_path=cfg.ladybug_path,
         verbose=not quiet,
         quiet=quiet,
         env=env,

java_codebase_rag/pipeline.py

@@ -201,7 +201,7 @@ def run_cocoindex_drop(env: dict[str, str], *, quiet: bool) -> subprocess.Comple
 def run_build_ast_graph(
     *,
     source_root: Path,
-    kuzu_path: Path,
+    ladybug_path: Path,
     verbose: bool,
     quiet: bool = False,
     env: dict[str, str] | None = None,
@@ -219,8 +219,8 @@ def run_build_ast_graph(
         str(builder),
         "--source-root",
         str(source_root),
-        "--kuzu-path",
-        str(kuzu_path),
+        "--ladybug-path",
+        str(ladybug_path),
     ]
     # Three-tier: --quiet (silent) / default (filtered progress) / --verbose (raw).
     # Default passes --verbose so the builder emits per-pass progress lines,
@@ -254,7 +254,7 @@ def run_build_ast_graph(
 def run_incremental_graph(
     *,
     source_root: Path,
-    kuzu_path: Path,
+    ladybug_path: Path,
     verbose: bool,
     quiet: bool = False,
     env: dict[str, str] | None = None,
@@ -273,8 +273,8 @@ def run_incremental_graph(
         str(builder),
         "--source-root",
         str(source_root),
-        "--kuzu-path",
-        str(kuzu_path),
+        "--ladybug-path",
+        str(ladybug_path),
         "--incremental",
     ]
     # Three-tier: --quiet (silent) / default (filtered progress) / --verbose (raw).

java_index_flow_lancedb.py

@@ -4,7 +4,7 @@
 LanceDB requires a single primary key per table; each chunk gets a UUID `id`.
 
 Environment:
-  JAVA_CODEBASE_RAG_INDEX_DIR — Lance tables + Kuzu + cocoindex state (default: ./.java-codebase-rag)
+  JAVA_CODEBASE_RAG_INDEX_DIR — Lance tables + LadybugDB + cocoindex state (default: ./.java-codebase-rag)
   JAVA_CODEBASE_RAG_SOURCE_ROOT — Java repo root for indexing (optional; else cocoindex cwd)
   SBERT_MODEL / SBERT_DEVICE — embedding (optional; YAML also supported via java-codebase-rag CLI)
 

java_ontology.py

@@ -141,7 +141,7 @@
 @dataclass(frozen=True)
 class EdgeAttr:
     name: str
-    kuzu_type: str
+    graph_type: str
     purpose: str