java-codebase-rag 0.6.2__py3-none-any.whl → 0.6.4__py3-none-any.whl

path_filtering.py

@@ -342,24 +342,22 @@ class LayeredIgnore:
         )
         return mega, GitIgnoreSpec.from_lines(mega), meta
 
-    def is_ignored(self, path: Path) -> tuple[bool, IgnoreLayer | None]:
-        """Return whether ``path`` is ignored and which layer last matched."""
+    def is_ignored(self, path: Path) -> bool:
+        """Return whether ``path`` is ignored by any configured layer.
+
+        Boolean-only fast path for the per-file index walk. It deliberately does
+        not compute *which* layer/source last matched: that attribution is
+        O(rules²) via :func:`_winning_row` (one ``GitIgnoreSpec`` rebuild per
+        rule prefix) and is only needed for ``diagnose-ignore``, so it lives in
+        :meth:`diagnose_dict` and is never paid on the hot path.
+        """
         rel = self._rel_project(path)
         if rel is None:
-            return False, None
-        mega, spec, meta = self._mega(rel)
+            return False
+        mega, spec, _ = self._mega(rel)
         if not mega:
-            return False, None
-        ignored = spec.match_file(rel)
-        if not ignored:
-            return False, None
-        src, fp, ln, _pat = _winning_row(rel, mega, meta)
-        return True, IgnoreLayer(
-            root=self.project_root,
-            spec=spec,
-            source=src,
-            ignore_file=fp,
-        )
+            return False
+        return spec.match_file(rel)
 
     def diagnose(self, path: Path) -> str:
         """Human-readable, multi-line explanation of the ignore decision."""
@@ -466,7 +464,6 @@ def iter_java_source_files(
             if not fn.endswith(".java"):
                 continue
             p = Path(dirpath) / fn
-            ign, _ = ignore_ctx.is_ignored(p)
-            if ign:
+            if ignore_ctx.is_ignored(p):
                 continue
             yield p

pr_analysis.py

@@ -152,7 +152,7 @@ def _resolve_graph_filename(
     *,
     ambiguity_notes: list[str] | None = None,
 ) -> str | None:
-    """Map a diff path to `Symbol.filename` values stored in Kuzu."""
+    """Map a diff path to `Symbol.filename` values stored in LadybugDB."""
     variants = {_strip_ab_prefix(path)}
     for v in list(variants):
         if v.startswith("./"):
@@ -362,7 +362,7 @@ def _is_public_interface_method(graph: Any, sym: SymbolHit) -> bool:
 
 
 def _route_ids_for_symbol(graph: Any, symbol_id: str) -> list[str]:
-    # Note: Kuzu rejects `ORDER BY r.id` together with `RETURN DISTINCT r.id` (binder loses `r`).
+    # Note: LadybugDB rejects `ORDER BY r.id` together with `RETURN DISTINCT r.id` (binder loses `r`).
     q = (
         "MATCH (s:Symbol)-[e:EXPOSES]->(r:Route) WHERE s.id = $sid "
         "RETURN r.id AS id ORDER BY id"
@@ -384,7 +384,6 @@ def compute_risk(graph: Any, changed: list[ChangedSymbol]) -> PrRiskReport:
     bump (up to +1.0) after normalization so they influence rank while
     preserving the public scalar contract.
     """
-    notes: list[str] = []
     blast_by: dict[str, int] = {}
     blast_total = 0
     routes: list[str] = []
@@ -495,7 +494,7 @@ def compute_risk(graph: Any, changed: list[ChangedSymbol]) -> PrRiskReport:
         routes_touched=routes,
         risk_score=score,
         risk_band=band,
-        notes=notes,
+        notes=[],
     )
 
 

search_lancedb.py

@@ -677,8 +677,8 @@ def _graph_expand_merge(
     expand_depth: int,
     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.
+    """Expand vector top-k through the LadybugDB graph and fuse (RRF) with the original list."""
+    # Lazy import so the module works without ladybug installed when graph_expand=False.
     try:
         from ladybug_queries import LadybugGraph
     except Exception:

server.py

@@ -13,9 +13,8 @@ import mcp_v2
 from index_common import SBERT_MODEL
 from java_codebase_rag.cli_progress import (
     accumulate_and_relay_subprocess_streams,
-    emit_vectors_finish,
-    emit_vectors_start,
 )
+from java_codebase_rag.progress import ProgressEvent
 from java_codebase_rag._fdlimit import raise_fd_limit
 from java_codebase_rag.config import (
     cocoindex_subprocess_env_defaults,
@@ -113,7 +112,44 @@ class ScopeManager:
 
     def _detect_scope(self) -> str | None:
         from graph_enrich import detect_microservice_from_path
-        return detect_microservice_from_path(Path.cwd(), self.source_root)
+
+        candidate = detect_microservice_from_path(Path.cwd(), self.source_root)
+        if candidate is None:
+            return None
+        # Only auto-scope to a microservice that actually has indexed code.
+        # detect_microservice_from_path can mislabel a non-microservice
+        # top-level child of source_root — most importantly the config/context
+        # directory the MCP server is launched from (no build marker, no
+        # source) — via its "first path segment under root" fallback. Scoping
+        # every query to such a name yields zero matches, so all tools return
+        # empty. A real microservice the operator is working in is, by
+        # definition, present in the index, so validating against the indexed
+        # set cannot suppress a legitimate scope. When the index is unreadable
+        # (empty known set) we keep the detected candidate rather than silently
+        # disabling auto-scope on a transient graph error.
+        known = self._indexed_microservices()
+        if known and candidate not in known:
+            return None
+        return candidate
+
+    def _indexed_microservices(self) -> set[str]:
+        """Microservice names that have indexed type symbols.
+
+        Graph-only source of truth: the graph is always built alongside Lance,
+        and a Lance-only index (no graph) is not a supported state. Any failure
+        (graph missing, open error, empty index) returns an empty set, which
+        ``_detect_scope`` treats as "cannot validate — keep detection".
+        """
+        try:
+            if not LadybugGraph.exists():
+                return set()
+            # LadybugGraph.get() opens the DB and runs meta(); it can raise
+            # (e.g. RuntimeError on ontology-version mismatch). Caught here ->
+            # empty set -> _detect_scope keeps the detected scope.
+            counts = LadybugGraph.get().microservice_counts()
+            return {name for name in counts if name}
+        except Exception:
+            return set()
 
     def _log_detection(self) -> None:
         if self.default_scope:
@@ -270,10 +306,20 @@ def list_code_index_tables_payload() -> IndexInfoOutput:
     )
 
 
-async def run_refresh_pipeline(*, quiet: bool = False, verbose: bool = True) -> RefreshIndexOutput:
+async def run_refresh_pipeline(
+    *,
+    quiet: bool = False,
+    verbose: bool = True,
+    on_progress=None,
+    on_progress_console: object | None = None,
+) -> RefreshIndexOutput:
     root = _project_root()
     cocoindex_bin = Path(sys.executable).parent / "cocoindex"
     if not cocoindex_bin.is_file():
+        # 127 pre-spawn: emit a terminal failed vectors event so the renderer's
+        # task doesn't hang at running (matches the sync pipeline path).
+        if on_progress is not None:
+            on_progress(ProgressEvent(kind="vectors", phase=None, pass_=None, done=None, total=None, status="failed", elapsed_s=None))
         return RefreshIndexOutput(
             success=False,
             message=f"cocoindex not found next to Python: {cocoindex_bin}",
@@ -286,6 +332,8 @@ async def run_refresh_pipeline(*, quiet: bool = False, verbose: bool = True) ->
         if fallback.is_file():
             flow_path = fallback
         else:
+            if on_progress is not None:
+                on_progress(ProgressEvent(kind="vectors", phase=None, pass_=None, done=None, total=None, status="failed", elapsed_s=None))
             return RefreshIndexOutput(
                 success=False,
                 message=f"java_index_flow_lancedb.py not found under {root} nor {bundle_dir}",
@@ -308,13 +356,14 @@ async def run_refresh_pipeline(*, quiet: bool = False, verbose: bool = True) ->
             )
             out_b, err_b = await proc.communicate()
         except Exception as exc:
+            if on_progress is not None:
+                on_progress(ProgressEvent(kind="vectors", phase=None, pass_=None, done=None, total=None, status="failed", elapsed_s=None))
             return RefreshIndexOutput(
                 success=False,
                 message=f"spawn failed: {exc!s}",
                 phases_run=[],
             )
     else:
-        emit_vectors_start()
         t0 = time.perf_counter()
         code_c = -1
         try:
@@ -329,16 +378,30 @@ async def run_refresh_pipeline(*, quiet: bool = False, verbose: bool = True) ->
                 stdout=asyncio.subprocess.PIPE,
                 stderr=asyncio.subprocess.PIPE,
             )
-            out_b, err_b = await accumulate_and_relay_subprocess_streams(proc, relay=True, verbose=verbose)
+            # The vectors task is fed by the child's per-file ticks + the
+            # approximate total line, parsed by the ProgressRelay inside the
+            # async drain and routed to on_progress.
+            out_b, err_b = await accumulate_and_relay_subprocess_streams(
+                proc, relay=True, verbose=verbose,
+                on_progress=on_progress, on_progress_console=on_progress_console,
+            )
             code_c = proc.returncode if proc.returncode is not None else -1
         except Exception as exc:
+            if on_progress is not None:
+                on_progress(ProgressEvent(kind="vectors", phase=None, pass_=None, done=None, total=None, status="failed", elapsed_s=None))
             return RefreshIndexOutput(
                 success=False,
                 message=f"spawn failed: {exc!s}",
                 phases_run=[],
             )
         finally:
-            emit_vectors_finish(elapsed_s=time.perf_counter() - t0, exit_code=code_c)
+            # The parent emits the terminal vectors event (the flow can't — no
+            # "all files done" hook). Drives clamp-on-completion + phase
+            # transition to Optimize.
+            if on_progress is not None:
+                elapsed = time.perf_counter() - t0
+                status = "done" if code_c == 0 else "failed"
+                on_progress(ProgressEvent(kind="vectors", phase=None, pass_=None, done=None, total=None, status=status, elapsed_s=elapsed))
     assert proc is not None
     out = out_b.decode(errors="replace")
     err = err_b.decode(errors="replace")
@@ -366,7 +429,7 @@ async def run_refresh_pipeline(*, quiet: bool = False, verbose: bool = True) ->
                 idx_dir = Path(idx_raw)
             else:
                 idx_dir = (root / ".java-codebase-rag").resolve()
-            await optimize_lance_tables(idx_dir, quiet=quiet)
+            await optimize_lance_tables(idx_dir, quiet=quiet, on_progress=on_progress)
         except Exception as exc:
             optimize_error = f"lance optimize failed: {exc}"
             print(f"java-codebase-rag: {optimize_error}", file=sys.stderr)
@@ -394,7 +457,10 @@ async def run_refresh_pipeline(*, quiet: bool = False, verbose: bool = True) ->
                 if quiet:
                     gout_b, gerr_b = await gproc.communicate()
                 else:
-                    gout_b, gerr_b = await accumulate_and_relay_subprocess_streams(gproc, relay=True, verbose=verbose)
+                    gout_b, gerr_b = await accumulate_and_relay_subprocess_streams(
+                        gproc, relay=True, verbose=verbose,
+                        on_progress=on_progress, on_progress_console=on_progress_console,
+                    )
                 graph_code = gproc.returncode
                 graph_out = gout_b.decode(errors="replace")
                 graph_err = gerr_b.decode(errors="replace")

java_codebase_rag-0.6.2.dist-info/RECORD

@@ -1,33 +0,0 @@
-ast_java.py,sha256=NQgZzstbsMq-PdowoD6r_ixJKxEEFzTP9xUzqDpiXeU,99661
-brownfield_events.py,sha256=yxXkKDgMb3VPtaiakGzncHM_EGnda8xIue6w90yYp8s,2055
-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=71bCLDNvMy0SpZPzSR5apJ0qJXNd6y5ggkLdBEw_PFo,16682
-ladybug_queries.py,sha256=912j9VAYDjcU4ReVorWQ6R4DZl0tteKic-Pqu0jyBS0,90837
-mcp_hints.py,sha256=3swh05LSiWur3tm3-yssndBsLxIxFhy501kBtJI8jJ0,42509
-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=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=HkzCP8-G3WlCzoXcVCI2K3forDOMpTmUSoxgx3jbKk4,34774
-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=bfwYI4R8PU9YV_M4r8-03iaUZ_0TW-qN_NuhIsDXy2M,18769
-java_codebase_rag/installer.py,sha256=sE0l85K_o291PdpF1vpesefR9VgdvvVeARXrpTxa30A,46689
-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-0.6.2.dist-info/licenses/LICENSE,sha256=gxvtiHtuviR_q8ZAjWw-QTcF3DyPzg6ZY-lQrr8OPpw,1068
-java_codebase_rag-0.6.2.dist-info/METADATA,sha256=X92kaZ5TbEacz0sznWtUtpYEJvBdWIXX0s8MlqOeRyg,16934
-java_codebase_rag-0.6.2.dist-info/WHEEL,sha256=aeYiig01lYGDzBgS8HxWXOg3uV61G9ijOsup-k9o1sk,91
-java_codebase_rag-0.6.2.dist-info/entry_points.txt,sha256=wsPZwot0Ui4JI3TIgW8LcbN8bNtKFbwQAlHAAJXfYgQ,117
-java_codebase_rag-0.6.2.dist-info/top_level.txt,sha256=syQgi8XPBwY2ws_NZ1uRCxTf_s41NpshwEHNdcdnk3A,245
-java_codebase_rag-0.6.2.dist-info/RECORD,,