java-codebase-rag 0.6.0__py3-none-any.whl → 0.6.1__py3-none-any.whl

ast_java.py

@@ -13,6 +13,7 @@ Python with no tree-sitter dependency.
 from __future__ import annotations
 
 import posixpath
+import sys
 from dataclasses import dataclass, field
 from functools import lru_cache
 from typing import Iterable
@@ -1642,9 +1643,17 @@ def _parse_codebase_http_client_annotation(
     pairs, _ = _annotation_kv_nodes(ann, src)
     client_kind = ""
     if "clientKind" in pairs:
-        val, _kind = _annotation_value(pairs["clientKind"], src)
-        if val and _kind == "enum":
-            client_kind = str(val)
+        val, vkind = _annotation_value(pairs["clientKind"], src)
+        if val and vkind == "enum":
+            kind_val = str(val)
+            from java_ontology import VALID_CLIENT_KINDS  # deferred: java_ontology imports ast_java
+            if kind_val in VALID_CLIENT_KINDS:
+                client_kind = kind_val
+            else:
+                print(
+                    f"[lancedb-mcp] CodebaseHttpClient: invalid clientKind {kind_val!r} — ignored",
+                    file=sys.stderr,
+                )
     target_service = ""
     if "targetService" in pairs:
         atoms = _string_value_atoms(pairs["targetService"], src, ctx)
@@ -1714,9 +1723,17 @@ def _parse_codebase_producer_annotation(
     client_kind = "kafka_send"
     kind_node = pairs.get("producerKind") or pairs.get("clientKind")
     if kind_node is not None:
-        val, _kind = _annotation_value(kind_node, src)
-        if val and _kind == "enum":
-            client_kind = str(val)
+        val, vkind = _annotation_value(kind_node, src)
+        if val and vkind == "enum":
+            kind_val = str(val)
+            from java_ontology import VALID_PRODUCER_KINDS  # deferred: java_ontology imports ast_java
+            if kind_val in VALID_PRODUCER_KINDS:
+                client_kind = kind_val
+            else:
+                print(
+                    f"[lancedb-mcp] CodebaseProducer: invalid producerKind {kind_val!r} — ignored",
+                    file=sys.stderr,
+                )
     topic = ""
     if "topic" in pairs:
         atoms = _string_value_atoms(pairs["topic"], src, ctx)

build_ast_graph.py

@@ -3668,10 +3668,17 @@ def incremental_rebuild(
 
 
 def _init_hash_tracker(source_root: Path, ladybug_path: Path) -> int:
-    """Initialize hash tracker for all Java files. Returns number of files hashed."""
+    """Initialize hash tracker for all Java files. Returns number of files hashed.
+
+    Called right after a full graph rebuild (``write_ladybug``), so the store must
+    mirror exactly the files that were just indexed. We deliberately do NOT
+    ``load()`` the existing store: ``update`` re-hashes every current file anyway,
+    and preserving old entries would leave stale hashes for files that no longer
+    exist (deleted or now-ignored). Those ghosts would be re-detected as "removed"
+    on every subsequent ``increment``, sustaining an endless full-rebuild loop.
+    """
     index_dir = ladybug_path.parent
     tracker = FileHashTracker(index_dir)
-    tracker.load()
     ignore = LayeredIgnore(source_root)
     all_files: set[str] = set()
     source_root_resolved = source_root.resolve()
@@ -3742,7 +3749,7 @@ def _write_clients_producers_and_calls(conn: ladybug.Connection, tables: GraphTa
 
     # Write declares_client edges
     for row in tables.declares_client_rows:
-        source_file = member_by_id.get(row.symbol_id, MemberEntry(kind="", decl=None, parent_id="", parent_fqn="", file_path="", module="", microservice="")).file_path
+        source_file = member_by_id.get(row.symbol_id, MemberEntry(kind="", decl=None, parent_id="", parent_fqn="", file_path="", module="", microservice="", node_id="")).file_path
         conn.execute(_CREATE_DECLARES_CLIENT, {
             "sid": row.symbol_id,
             "cid": row.client_id,
@@ -3753,7 +3760,7 @@ def _write_clients_producers_and_calls(conn: ladybug.Connection, tables: GraphTa
 
     # Write declares_producer edges
     for row in tables.declares_producer_rows:
-        source_file = member_by_id.get(row.symbol_id, MemberEntry(kind="", decl=None, parent_id="", parent_fqn="", file_path="", module="", microservice="")).file_path
+        source_file = member_by_id.get(row.symbol_id, MemberEntry(kind="", decl=None, parent_id="", parent_fqn="", file_path="", module="", microservice="", node_id="")).file_path
         conn.execute(_CREATE_DECLARES_PRODUCER, {
             "sid": row.symbol_id,
             "pid": row.producer_id,

java_codebase_rag/config.py

@@ -306,9 +306,19 @@ def _pick_bool(
 def _resolve_index_dir_path(
     *,
     source_root: Path,
+    config_dir: Path,
     cli_index_dir: str | None,
     yaml_dict: dict[str, Any],
 ) -> tuple[Path, SettingSource]:
+    # Bases for relative paths:
+    #   - YAML ``index_dir``  -> the config file's directory (``config_dir``),
+    #     the SAME base used for YAML ``source_root``. Paths written in the
+    #     config file are relative to the file, so both keys stay consistent.
+    #   - CLI / env ``index_dir`` -> ``source_root`` (unchanged). These are not
+    #     "in the config file"; preserving the existing base avoids a semantics
+    #     change for operators who pass ``--index-dir`` on the command line.
+    #   - Default ``./.java-codebase-rag`` -> ``source_root`` so the index sits
+    #     beside the Java tree (the layout ``discover_project_root`` anchors on).
     raw_cli = cli_index_dir.strip() if isinstance(cli_index_dir, str) else None
     if raw_cli:
         p = Path(raw_cli).expanduser()
@@ -324,7 +334,7 @@ def _resolve_index_dir_path(
     idx = yaml_dict.get("index_dir")
     if isinstance(idx, str) and idx.strip():
         p = Path(idx.strip()).expanduser()
-        out = p.resolve() if p.is_absolute() else (source_root / p).resolve()
+        out = p.resolve() if p.is_absolute() else (config_dir / p).resolve()
         return out, "yaml"
 
     return (source_root / ".java-codebase-rag").resolve(), "default"
@@ -368,7 +378,7 @@ def resolve_operator_config(
                 root = config_dir
 
     index_dir, index_src = _resolve_index_dir_path(
-        source_root=root, cli_index_dir=cli_index_dir, yaml_dict=yaml_dict
+        source_root=root, config_dir=config_dir, cli_index_dir=cli_index_dir, yaml_dict=yaml_dict
     )
     model, model_src = _pick_str(
         cli_val=cli_embedding_model,

{java_codebase_rag-0.6.0.dist-info → java_codebase_rag-0.6.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: java-codebase-rag
-Version: 0.6.0
+Version: 0.6.1
 Summary: MCP server for semantic + structural search over Java codebases
 Author: HumanBean17
 License-Expression: MIT
@@ -35,6 +35,7 @@ Requires-Dist: unidiff<1,>=0.7.3
 Provides-Extra: dev
 Requires-Dist: pytest>=7; extra == "dev"
 Requires-Dist: pytest-asyncio>=0.21; extra == "dev"
+Requires-Dist: pytest-xdist>=3; extra == "dev"
 Requires-Dist: ruff>=0.4; extra == "dev"
 Dynamic: license-file
 

{java_codebase_rag-0.6.0.dist-info → java_codebase_rag-0.6.1.dist-info}/RECORD

@@ -1,33 +1,36 @@
-ast_java.py,sha256=TMesuv4SYqzkwfKxf_Pps0KaPLZNZOrhU8mL20bwqeQ,98882
+ast_java.py,sha256=NQgZzstbsMq-PdowoD6r_ixJKxEEFzTP9xUzqDpiXeU,99661
 brownfield_events.py,sha256=yxXkKDgMb3VPtaiakGzncHM_EGnda8xIue6w90yYp8s,2055
-build_ast_graph.py,sha256=GNbjiIAwsXaJQ9Je6gbR-dB9SbnaLThya2pEw3tggQs,152396
+build_ast_graph.py,sha256=OKigswkUmWwUAKXXRNH4zplw2VonIdWUWzVjC-t5roo,152893
 chunk_heuristics.py,sha256=aQk2NOKxzUdqoUAJUO3G3LE0MN_bYZWNLQ0tkmj5uts,1813
 graph_enrich.py,sha256=POT4LwSkTsrjUmP67bsm2UezUam70cunuPDYDh-v1Bs,63332
 index_common.py,sha256=HT6FKHFJ084eFvd3fR1j8z8gf4eWoPHVW8GXLpw464I,285
 java_index_flow_lancedb.py,sha256=MH9iTNF6HDHDTt5Jn7TOVE5hQ4WUPNt7PlQoh1tuh9o,13212
 java_index_v1_common.py,sha256=nF1KrSqboF_RRvWerG9knRRFmWwsrG_CvhgnsoZ8KqA,1154
-java_ontology.py,sha256=FcnOq1XWhUP03OfnTkRStslqrNyukzUKH7VNuK6Bme4,16425
+java_ontology.py,sha256=71bCLDNvMy0SpZPzSR5apJ0qJXNd6y5ggkLdBEw_PFo,16682
+kuzu_queries.py,sha256=9bQzrU311AOw_BcUp_KSGiZgPVSaLSU7y63XfcT_vqI,90137
 ladybug_queries.py,sha256=912j9VAYDjcU4ReVorWQ6R4DZl0tteKic-Pqu0jyBS0,90837
 mcp_hints.py,sha256=3swh05LSiWur3tm3-yssndBsLxIxFhy501kBtJI8jJ0,42509
-mcp_v2.py,sha256=64UDrQ27hAQtlz3pFp9A3Xlk95bUjYZ4VBscsyAPCIY,79116
+mcp_v2.py,sha256=o94GJI7j6dLJDIA3R_1ZiQhjzQfMAEW3etdeZYnHOUc,80637
 path_filtering.py,sha256=-oX16SYLWYwX9pcV1fu3vbVTIhY1GzFflT7J1E2tqPY,17122
 pr_analysis.py,sha256=3-5L8_G5XupdJsl9RN73Lq-ejPoK11B3m_VzAx2fGG8,18413
 search_lancedb.py,sha256=scG6HBUrsgIeSWFrGcLcGdhWv1qODOx4JOBMAlLDY_E,36793
-server.py,sha256=uGKT0PdM-bVrzIsfbxF6ZuHGcuRMSSlvkJk0e7Ff43Y,30556
+server.py,sha256=Js3XDpV7ThAtj352StH6QdhHutf1D5qUkbR-8k3jO8g,31303
 java_codebase_rag/__init__.py,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
 java_codebase_rag/_fdlimit.py,sha256=WroFdfSNbcriKok6q8znTf74dqlznxea_1Fd5bHl_3o,1930
 java_codebase_rag/cli.py,sha256=a5IFLWAsh77mfLv1Z9OdpvLaYvj4i0KR3_kLtL-ans8,34156
 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=u4OomvglTWHUmMpcxN8wPRnRGfXVp3qK_GJ5pY96O98,16267
+java_codebase_rag/config.py,sha256=Gn3LgxkTOtAvsL-3U2Xn7atOIhyOT2aGmY8SBBTLoQg,16975
 java_codebase_rag/installer.py,sha256=DlBuVVWbHXgcjaQkuXUeT9fNdmk7XZefVT3zzw47k18,45965
 java_codebase_rag/lance_optimize.py,sha256=MzACYlgwxmkJCK64qQLyIAdizSq5BARqaMYSZONlc1I,6069
 java_codebase_rag/pipeline.py,sha256=UcgluFAW9Ghnas8u40x45bVic0mQv6rjzcliDKsnYJI,11936
 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-0.6.0.dist-info/licenses/LICENSE,sha256=gxvtiHtuviR_q8ZAjWw-QTcF3DyPzg6ZY-lQrr8OPpw,1068
-java_codebase_rag-0.6.0.dist-info/METADATA,sha256=GoMO3zFTb98w4rVV5SMXpcLK-irlDs7aUH0wBGlv5cQ,16887
-java_codebase_rag-0.6.0.dist-info/WHEEL,sha256=aeYiig01lYGDzBgS8HxWXOg3uV61G9ijOsup-k9o1sk,91
-java_codebase_rag-0.6.0.dist-info/entry_points.txt,sha256=mVVQJa0n73OWfhHXYCDoPRrWin_LJhH2Rn0CkJ2iax4,101
-java_codebase_rag-0.6.0.dist-info/top_level.txt,sha256=syQgi8XPBwY2ws_NZ1uRCxTf_s41NpshwEHNdcdnk3A,245
-java_codebase_rag-0.6.0.dist-info/RECORD,,
+java_codebase_rag-0.6.1.dist-info/licenses/LICENSE,sha256=gxvtiHtuviR_q8ZAjWw-QTcF3DyPzg6ZY-lQrr8OPpw,1068
+user_rag/__init__.py,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
+user_rag/cli.py,sha256=TVcyfzwvmdYXJW6KrEYTKMHm7z2JSXMmz2uB-8kkjxY,5604
+java_codebase_rag-0.6.1.dist-info/METADATA,sha256=aPiLbGD8xE-P3B_RI9gx7VuqrTd-VUriZ--ZPYNK02I,16934
+java_codebase_rag-0.6.1.dist-info/WHEEL,sha256=aeYiig01lYGDzBgS8HxWXOg3uV61G9ijOsup-k9o1sk,91
+java_codebase_rag-0.6.1.dist-info/entry_points.txt,sha256=mVVQJa0n73OWfhHXYCDoPRrWin_LJhH2Rn0CkJ2iax4,101
+java_codebase_rag-0.6.1.dist-info/top_level.txt,sha256=syQgi8XPBwY2ws_NZ1uRCxTf_s41NpshwEHNdcdnk3A,245
+java_codebase_rag-0.6.1.dist-info/RECORD,,

java_ontology.py

@@ -15,7 +15,10 @@ from ast_java import (
     _TYPE_ANN_TO_CAPABILITY,
 )
 
-# Roles: Spring stereotype values plus DTO from `infer_role_for_type`.
+# Roles assignable by indexing: Spring stereotype values plus DTO. ``OTHER`` is the
+# built-in inference fallback (ast_java.infer_role when nothing matches) and is
+# deliberately excluded here — it is a read-side value (the mcp_v2 ``Role`` enum
+# includes it) but not a role a user may set via @CodebaseRole / role_overrides.
 VALID_ROLES: frozenset[str] = frozenset((*ROLE_ANNOTATIONS.values(), "DTO"))
 
 VALID_CAPABILITIES: frozenset[str] = frozenset(