java-codebase-rag 0.5.2__py3-none-any.whl → 0.6.0__py3-none-any.whl

mcp_v2.py

@@ -30,7 +30,7 @@ from sentence_transformers import SentenceTransformer
 from index_common import SBERT_MODEL
 from java_codebase_rag.config import resolved_sbert_model_for_process_env
 from java_ontology import EDGE_SCHEMA, ResolveReason
-from kuzu_queries import KuzuGraph, OVERRIDE_AXIS_COMPOSED_EDGE_TYPES
+from ladybug_queries import LadybugGraph, OVERRIDE_AXIS_COMPOSED_EDGE_TYPES
 from mcp_hints import generate_hints, MCP_HINTS_STRUCTURED_FIELD_DESCRIPTION
 from search_lancedb import TABLES, run_search
 
@@ -604,7 +604,7 @@ def _node_kind_from_id(
 
 
 def _resolve_node_kind(
-    graph: KuzuGraph,
+    graph: LadybugGraph,
     node_id: str,
 ) -> Literal["symbol", "route", "client", "producer", "unresolved_call_site"]:
     try:
@@ -733,7 +733,7 @@ def _node_ref_from_row(kind: Literal["symbol", "route", "client", "producer"], r
 
 
 def _load_node_record(
-    graph: KuzuGraph, node_id: str, kind: Literal["symbol", "route", "client", "producer"],
+    graph: LadybugGraph, node_id: str, kind: Literal["symbol", "route", "client", "producer"],
 ) -> dict[str, Any] | None:
     if kind == "symbol":
         projection = (
@@ -807,7 +807,7 @@ def _merge_overrides_edge_summary(
 
 
 def _edge_summary_for_node(
-    graph: KuzuGraph, node_id: str, *, kind: str, row: dict[str, Any]
+    graph: LadybugGraph, node_id: str, *, kind: str, row: dict[str, Any]
 ) -> dict[str, dict[str, int]]:
     summary = dict(graph.edge_counts_for(node_id))
     sym_kind = str(row.get("kind") or "")
@@ -887,7 +887,7 @@ def search_v2(
     offset: int = 0,
     path_contains: str | None = None,
     filter: NodeFilter | dict[str, Any] | str | None = None,
-    graph: KuzuGraph | None = None,
+    graph: LadybugGraph | None = None,
 ) -> SearchOutput:
     try:
         raw_filter = _coerce_filter(filter)
@@ -967,10 +967,10 @@ def find_v2(
     filter: NodeFilter | dict[str, Any] | str,
     limit: int = 25,
     offset: int = 0,
-    graph: KuzuGraph | None = None,
+    graph: LadybugGraph | None = None,
 ) -> FindOutput:
     try:
-        g = graph or KuzuGraph.get()
+        g = graph or LadybugGraph.get()
         raw_filter = _coerce_filter(filter)
         if raw_filter is None:
             raw_filter = {}
@@ -1063,10 +1063,10 @@ _DESCRIBE_UCS_ID_MESSAGE = (
 def describe_v2(
     id: str | None = None,
     fqn: str | None = None,
-    graph: KuzuGraph | None = None,
+    graph: LadybugGraph | None = None,
 ) -> DescribeOutput:
     try:
-        g = graph or KuzuGraph.get()
+        g = graph or LadybugGraph.get()
         has_id = bool(id and str(id).strip())
         has_fqn = bool(fqn and str(fqn).strip())
         if not has_id and not has_fqn:
@@ -1171,7 +1171,7 @@ def _resolve_parse_microservice_route(identifier: str) -> tuple[str, str, str] |
 
 
 def _resolve_symbol_candidates(
-    g: KuzuGraph,
+    g: LadybugGraph,
     identifier: str,
 ) -> list[tuple[NodeRef, ResolveReason, int]]:
     out: list[tuple[NodeRef, ResolveReason, int]] = []
@@ -1213,7 +1213,7 @@ def _resolve_symbol_candidates(
 
 
 def _resolve_route_candidates(
-    g: KuzuGraph,
+    g: LadybugGraph,
     identifier: str,
 ) -> list[tuple[NodeRef, ResolveReason, int]]:
     out: list[tuple[NodeRef, ResolveReason, int]] = []
@@ -1265,7 +1265,7 @@ def _resolve_route_candidates(
 
 
 def _resolve_client_candidates(
-    g: KuzuGraph,
+    g: LadybugGraph,
     identifier: str,
 ) -> list[tuple[NodeRef, ResolveReason, int]]:
     out: list[tuple[NodeRef, ResolveReason, int]] = []
@@ -1304,7 +1304,7 @@ def _resolve_client_candidates(
 
 
 def _resolve_producer_candidates(
-    g: KuzuGraph,
+    g: LadybugGraph,
     identifier: str,
 ) -> list[tuple[NodeRef, ResolveReason, int]]:
     out: list[tuple[NodeRef, ResolveReason, int]] = []
@@ -1462,7 +1462,7 @@ def _resolve_finalize_success(
 def resolve_v2(
     identifier: str,
     hint_kind: Literal["symbol", "route", "client", "producer"] | None = None,
-    graph: KuzuGraph | None = None,
+    graph: LadybugGraph | None = None,
 ) -> ResolveOutput:
     try:
         trimmed, err = _resolve_validate_identifier(identifier)
@@ -1481,7 +1481,7 @@ def resolve_v2(
         if "*" in trimmed or "?" in trimmed:
             return _resolve_finalize_success(trimmed, hint_kind, [])
 
-        g = graph or KuzuGraph.get()
+        g = graph or LadybugGraph.get()
         raw: list[tuple[NodeRef, ResolveReason, int]] = []
         for kind in _resolve_kinds_to_search(hint_kind):
             if kind == "symbol":
@@ -1726,7 +1726,7 @@ def neighbors_v2(
         declares_composed = [k for k in composed_keys if k in _MEMBER_COMPOSED_EDGE_TYPES]
         override_composed = [k for k in composed_keys if k in _OVERRIDE_COMPOSED_EDGE_TYPES]
         ordered_composed = declares_composed + override_composed
-        g = graph or KuzuGraph.get()
+        g = graph or LadybugGraph.get()
         try:
             raw_filter = _coerce_filter(filter)
             nf = (

pr_analysis.py

@@ -12,7 +12,7 @@ from typing import Any
 from unidiff import PatchSet
 from unidiff.errors import UnidiffParseError
 
-from kuzu_queries import SymbolHit, find_symbols_in_file_range, _row_to_symbol
+from ladybug_queries import SymbolHit, find_symbols_in_file_range, _row_to_symbol
 
 
 @dataclass

search_lancedb.py

@@ -675,16 +675,16 @@ def _graph_expand_merge(
     limit: int,
     extra_predicates: list[str],
     expand_depth: int,
-    kuzu_path: str | None,
+    ladybug_path: str | None,
 ) -> list[dict]:
     """Expand vector top-k through the Kuzu graph and fuse (RRF) with the original list."""
     # Lazy import so the module works without kuzu installed when graph_expand=False.
     try:
-        from kuzu_queries import KuzuGraph
+        from ladybug_queries import LadybugGraph
     except Exception:
         return vector_rows
 
-    if not KuzuGraph.exists(kuzu_path):
+    if not LadybugGraph.exists(ladybug_path):
         return vector_rows
 
     seed_fqns = sorted({r.get("primary_type_fqn") for r in vector_rows if r.get("primary_type_fqn")})
@@ -692,7 +692,7 @@ def _graph_expand_merge(
         return vector_rows
 
     try:
-        graph = KuzuGraph.get(kuzu_path)
+        graph = LadybugGraph.get(ladybug_path)
         structural = graph.expand_fqns(seed_fqns, depth=expand_depth)
         method_pairs = graph.expand_methods(
             seed_fqns, depth=expand_depth, exclude_external=True,
@@ -804,7 +804,7 @@ def run_search(
     package_prefix: str | None = None,
     graph_expand: bool = False,
     expand_depth: int = 1,
-    kuzu_path: str | None = None,
+    ladybug_path: str | None = None,
     context_neighbors: int = 0,
     role_in: list[str] | None = None,
     exclude_roles: list[str] | None = None,
@@ -890,7 +890,7 @@ def run_search(
                 limit=need,
                 extra_predicates=extra_java,
                 expand_depth=expand_depth,
-                kuzu_path=kuzu_path,
+                ladybug_path=ladybug_path,
             )
 
         window = rows[offset : offset + limit]
@@ -966,7 +966,7 @@ def main() -> None:
     parser.add_argument("--package-prefix", default=None)
     parser.add_argument("--graph-expand", action="store_true")
     parser.add_argument("--expand-depth", type=int, default=1)
-    parser.add_argument("--kuzu-path", default=None)
+    parser.add_argument("--ladybug-path", default=None)
     parser.add_argument(
         "--context-neighbors", type=int, default=0,
         help="Attach N adjacent chunks per hit as surrounding context (Java only).",
@@ -1010,7 +1010,7 @@ def main() -> None:
             package_prefix=args.package_prefix,
             graph_expand=args.graph_expand,
             expand_depth=args.expand_depth,
-            kuzu_path=args.kuzu_path,
+            ladybug_path=args.ladybug_path,
             context_neighbors=args.context_neighbors,
         )
     except Exception as e:

server.py

@@ -16,23 +16,25 @@ from java_codebase_rag.cli_progress import (
     emit_vectors_finish,
     emit_vectors_start,
 )
+from java_codebase_rag._fdlimit import raise_fd_limit
 from java_codebase_rag.config import (
+    cocoindex_subprocess_env_defaults,
     discover_project_root,
     emit_legacy_env_hints_if_present,
     resolved_sbert_model_for_process_env,
     resolve_operator_config,
 )
-from kuzu_queries import KuzuGraph, resolve_kuzu_path
+from ladybug_queries import LadybugGraph, resolve_ladybug_path
 from mcp.server.fastmcp import FastMCP
 from pydantic import BaseModel, Field
 from search_lancedb import TABLES
 
 _COCOINDEX_TARGET = "java_index_flow_lancedb.py:JavaCodeIndexLance"
 _INSTRUCTIONS = (
-    "Java codebase graph navigator (LanceDB + Kuzu). "
+    "Java codebase graph navigator (LanceDB + Ladybug). "
     "Tools: search (NL/code locate), find (structured NodeFilter), describe (one node + edge_summary: stored edge-label counts and optional composed keys for type Symbols and override-axis virtual keys for method Symbols), "
     "neighbors (one hop; you MUST pass direction in|out AND edge_types list — no defaults), "
-    "resolve (identifier-shaped lookup for symbol/route/client/producer — three statuses one|many|none). "
+    "resolve (identifier-shaped lookup for symbol/route/client/producer — three statuses one|many/none). "
     "NodeFilter `filter` is a JSON object (preferred); a JSON-encoded string is also accepted as a fallback. "
     "Unknown filter keys and populated fields not applicable to the effective node kind fail with success=false and message. "
     "Edge labels: EXTENDS, IMPLEMENTS, INJECTS, OVERRIDES, DECLARES, DECLARES_CLIENT, DECLARES_PRODUCER, CALLS, EXPOSES, HTTP_CALLS, ASYNC_CALLS; "
@@ -85,6 +87,7 @@ class RefreshIndexOutput(BaseModel):
     graph_stdout: str = ""
     graph_stderr: str = ""
     phases_run: list[Literal["vectors", "graph"]] = Field(default_factory=list)
+    optimize_error: str | None = None
 
 
 class IndexInfoOutput(BaseModel):
@@ -162,36 +165,40 @@ def _cocoindex_subprocess_env(project_root: Path) -> dict[str, str]:
     idx = os.environ.get("JAVA_CODEBASE_RAG_INDEX_DIR", "").strip()
     if idx:
         sub_env["JAVA_CODEBASE_RAG_INDEX_DIR"] = str(Path(idx).expanduser().resolve())
+    # Cap CocoIndex concurrency to avoid EMFILE ("too many open files") under
+    # default OS fd limits. See: https://github.com/HumanBean17/java-codebase-rag/issues/306
+    for _k, _v in cocoindex_subprocess_env_defaults().items():
+        sub_env.setdefault(_k, _v)
     return sub_env
 
 
 def _graph_enabled() -> bool:
-    return KuzuGraph.exists()
+    return LadybugGraph.exists()
 
 
 def _graph_meta_output() -> GraphMetaOutput:
-    if not KuzuGraph.exists():
+    if not LadybugGraph.exists():
         return GraphMetaOutput(
             success=True,
             enabled=False,
-            db_path=resolve_kuzu_path(),
-            message="Kuzu graph not present; run java-codebase-rag reprocess or build_ast_graph.py",
+            db_path=resolve_ladybug_path(),
+            message="Ladybug graph not present; run java-codebase-rag reprocess or build_ast_graph.py",
         )
     try:
-        graph = KuzuGraph.get()
+        graph = LadybugGraph.get()
         meta = graph.meta()
     except Exception as e:
         return GraphMetaOutput(
             success=False,
             enabled=_graph_enabled(),
-            db_path=resolve_kuzu_path(),
-            message=f"Kuzu open failed: {e}",
+            db_path=resolve_ladybug_path(),
+            message=f"Ladybug open failed: {e}",
         )
     if "error" in meta:
         return GraphMetaOutput(
             success=False,
             enabled=_graph_enabled(),
-            db_path=meta.get("db_path", resolve_kuzu_path()),
+            db_path=meta.get("db_path", resolve_ladybug_path()),
             message=str(meta["error"]),
         )
     try:
@@ -209,7 +216,7 @@ def _graph_meta_output() -> GraphMetaOutput:
     return GraphMetaOutput(
         success=True,
         enabled=_graph_enabled(),
-        db_path=meta.get("db_path", resolve_kuzu_path()),
+        db_path=meta.get("db_path", resolve_ladybug_path()),
         ontology_version=int(meta.get("ontology_version") or 0),
         built_at=int(meta.get("built_at") or 0),
         source_root=str(meta.get("source_root") or ""),
@@ -323,9 +330,29 @@ async def run_refresh_pipeline(*, quiet: bool = False, verbose: bool = True) ->
     graph_code: int | None = None
     graph_out = ""
     graph_err = ""
+    optimize_error: str | None = None
     if ok:
         if not quiet:
             print(file=sys.stderr, flush=True)
+        # Serialized post-flow Lance optimize: the flow disabled its background
+        # optimize, so with cocoindex returned exit 0 there are no concurrent
+        # writers — this is the safe window to compact. An optimize failure is
+        # surfaced via optimize_error / stderr and must NOT flip the success of
+        # a vectors phase that succeeded; the index is still searchable.
+        try:
+            from java_codebase_rag.lance_optimize import optimize_lance_tables
+
+            idx_raw = os.environ.get("JAVA_CODEBASE_RAG_INDEX_DIR", "").strip()
+            if idx_raw and not idx_raw.startswith(("s3://", "gs://", "az://")):
+                idx_dir = Path(idx_raw).expanduser().resolve()
+            elif idx_raw:
+                idx_dir = Path(idx_raw)
+            else:
+                idx_dir = (root / ".java-codebase-rag").resolve()
+            await optimize_lance_tables(idx_dir, quiet=quiet)
+        except Exception as exc:
+            optimize_error = f"lance optimize failed: {exc}"
+            print(f"java-codebase-rag: {optimize_error}", file=sys.stderr)
         builder = Path(__file__).resolve().parent / "build_ast_graph.py"
         if builder.is_file():
             try:
@@ -334,8 +361,8 @@ async def run_refresh_pipeline(*, quiet: bool = False, verbose: bool = True) ->
                     str(builder),
                     "--source-root",
                     str(root),
-                    "--kuzu-path",
-                    resolve_kuzu_path(),
+                    "--ladybug-path",
+                    resolve_ladybug_path(),
                 ]
                 if not quiet:
                     graph_args.append("--verbose")
@@ -362,6 +389,10 @@ async def run_refresh_pipeline(*, quiet: bool = False, verbose: bool = True) ->
         message = f"cocoindex exit {proc.returncode}"
     elif graph_code is not None and graph_code != 0:
         message = f"graph builder exit {graph_code}"
+    # Surface a post-flow optimize failure in the message too (success is not
+    # flipped — the vectors phase succeeded and the index is still usable).
+    if optimize_error is not None:
+        message = optimize_error if message is None else f"{message}; {optimize_error}"
     return RefreshIndexOutput(
         success=ok and (graph_code is None or graph_code == 0),
         exit_code=proc.returncode,
@@ -372,6 +403,7 @@ async def run_refresh_pipeline(*, quiet: bool = False, verbose: bool = True) ->
         graph_stdout=graph_out[-4000:] if len(graph_out) > 4000 else graph_out,
         graph_stderr=graph_err[-4000:] if len(graph_err) > 4000 else graph_err,
         phases_run=phases_run,
+        optimize_error=optimize_error,
     )
 
 
@@ -619,6 +651,7 @@ def create_mcp_server() -> FastMCP:
 
 
 def main() -> None:
+    raise_fd_limit()
     emit_legacy_env_hints_if_present()
 
     # Load YAML config and apply embedding settings to environment

java_codebase_rag-0.5.2.dist-info/RECORD

@@ -1,31 +0,0 @@
-ast_java.py,sha256=OKoH7oX6L7AEEd6UY-spK8BPtWYY1T_4esrTC5VtoK8,98881
-brownfield_events.py,sha256=yxXkKDgMb3VPtaiakGzncHM_EGnda8xIue6w90yYp8s,2055
-build_ast_graph.py,sha256=1uqgFK2ebBdEc2QcAYK5vU4afOb95jU3zht5FracCkI,148683
-chunk_heuristics.py,sha256=aQk2NOKxzUdqoUAJUO3G3LE0MN_bYZWNLQ0tkmj5uts,1813
-graph_enrich.py,sha256=m3cksCHLqLHhA0Y-TLodbm09YfSJZjlTDN0Z51DiP2c,63317
-index_common.py,sha256=HT6FKHFJ084eFvd3fR1j8z8gf4eWoPHVW8GXLpw464I,285
-java_index_flow_lancedb.py,sha256=LMmfMSdE2d-ujxuJ2-hss7BhkrUMxHNyZuqsiGITuAI,12057
-java_index_v1_common.py,sha256=nF1KrSqboF_RRvWerG9knRRFmWwsrG_CvhgnsoZ8KqA,1154
-java_ontology.py,sha256=nM-oY8_91rmUudv9hAss1AMus9BFY9s5tTpAWjlCz00,16424
-kuzu_queries.py,sha256=9bQzrU311AOw_BcUp_KSGiZgPVSaLSU7y63XfcT_vqI,90137
-mcp_hints.py,sha256=3swh05LSiWur3tm3-yssndBsLxIxFhy501kBtJI8jJ0,42509
-mcp_v2.py,sha256=JFe62sYzJ2XiE6L3wAH8XG9_Ya2oOeJQ_hkiTmXFnSE,79065
-path_filtering.py,sha256=-oX16SYLWYwX9pcV1fu3vbVTIhY1GzFflT7J1E2tqPY,17122
-pr_analysis.py,sha256=Zaq90xYgMgrReV3vCGcFhOkK61gIRMAAIgs7ev-rJG4,18410
-search_lancedb.py,sha256=-XgtpbJ_3zDLiZ_vGKXjaLpl7RlvgyzUb7oAGoWkXO0,36754
-server.py,sha256=c4Bo0FXPoKP2AwIVP_wiv0XENkmKchLHf0QrQPUUgq4,28645
-java_codebase_rag/__init__.py,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
-java_codebase_rag/cli.py,sha256=WW-DsskSGr-d0JXBLkj4IdAa2OsAcLz5e54_DWvD9Sk,33872
-java_codebase_rag/cli_format.py,sha256=arU7P9W6Fvm7X_wzR1wJ8EfyxK1rDP_ESEhdA0ub4Mo,2579
-java_codebase_rag/cli_progress.py,sha256=9jCqEagYOXs32SYVA31_sOCrONvYy7cl1CrdBD2Pg44,3168
-java_codebase_rag/config.py,sha256=1BkRQsdY2ohZ8IWmbTG3WHgotVVUIrRTN537A1QAoCQ,15352
-java_codebase_rag/installer.py,sha256=flj330ZPSBrO2iw_yuNFBILHOTVbarMufYwqjZ8JzN0,42778
-java_codebase_rag/pipeline.py,sha256=nMXwX9r7HG9yPstrm7y_vfOMUZuDmw5_1lJTAfR-jwI,9488
-java_codebase_rag/install_data/agents/explorer-rag-enhanced.md,sha256=APl9d-No12qZNZLjU7mwNRwxHIgnT3ZtQZiD4clWlyU,14413
-java_codebase_rag/install_data/skills/explore-codebase/SKILL.md,sha256=pIM-Xdwq_fXkhhBJCdb-fA2nes5c_mMPcdUXb7Adyxo,12040
-java_codebase_rag-0.5.2.dist-info/licenses/LICENSE,sha256=gxvtiHtuviR_q8ZAjWw-QTcF3DyPzg6ZY-lQrr8OPpw,1068
-java_codebase_rag-0.5.2.dist-info/METADATA,sha256=LrlQYmOpTRNtx7q4-sr2yO8dCc5RHy4FgDY5GzvsoR8,16807
-java_codebase_rag-0.5.2.dist-info/WHEEL,sha256=aeYiig01lYGDzBgS8HxWXOg3uV61G9ijOsup-k9o1sk,91
-java_codebase_rag-0.5.2.dist-info/entry_points.txt,sha256=mVVQJa0n73OWfhHXYCDoPRrWin_LJhH2Rn0CkJ2iax4,101
-java_codebase_rag-0.5.2.dist-info/top_level.txt,sha256=5aIYoMkvJvvfXvf4iHn2OeSIM7PZXP-0j94eNESnwMw,242
-java_codebase_rag-0.5.2.dist-info/RECORD,,