sourcecode 1.31.5__py3-none-any.whl → 1.31.7__py3-none-any.whl

sourcecode/__init__.py

@@ -1,3 +1,3 @@
 """sourcecode — Deterministic codebase context maps for AI coding agents."""
 
-__version__ = "1.31.5"
+__version__ = "1.31.7"

sourcecode/cli.py

@@ -2495,23 +2495,81 @@ def _extract_java_endpoints(root: "Path") -> "dict[str, Any]":
 
     endpoints: list[dict] = []
     seen: set[tuple] = set()
+    # Fix 1: inheritance projection — tracks class data so controllers with ONLY
+    # inherited endpoints (no own @RequestMapping methods) are not silently dropped.
+    _class_info: dict[str, dict] = {}
+    _EXTENDS_RE_LOCAL = _re.compile(r'\bextends\s+(\w+)')
+
+    from sourcecode.path_filters import is_test_path as _is_test_path
 
     java_files = [
         p for p in root.rglob("*.java")
-        if "/test/" not in str(p).replace("\\", "/")
-        and "/tests/" not in str(p).replace("\\", "/")
+        if not _is_test_path(str(p).replace("\\", "/"))
         and "target/" not in str(p).replace("\\", "/")
     ]
 
+    # ── Meta-annotation index ─────────────────────────────────────────────────
+    # First pass: find @interface declarations and the annotations they carry.
+    # This lets us detect controllers annotated with framework wrappers like
+    # @FrameworkController that itself carries @Controller.
+    #
+    # Index maps annotation simple name → set of annotation names it is
+    # annotated with (one level; resolution is done recursively below).
+    _ANN_DECL_RE = _re.compile(r'public\s+@interface\s+(\w+)')
+    _ANN_USE_RE = _re.compile(r'@(\w+)')
+
+    _meta_index: dict[str, set[str]] = {}
+    for _jf in java_files:
+        _raw: str
+        try:
+            _raw = _jf.read_text(encoding="utf-8", errors="replace")
+        except OSError:
+            continue
+        if "@interface" not in _raw:
+            continue
+        _decl_m = _ANN_DECL_RE.search(_raw)
+        if not _decl_m:
+            continue
+        _ann_name = _decl_m.group(1)
+        # Collect annotations appearing in the header (before the @interface line)
+        _header = _raw[: _decl_m.start()]
+        _meta_anns: set[str] = {
+            m for m in _ANN_USE_RE.findall(_header)
+            if m not in ("interface", "interface")  # strip keywords; annotation names only
+        }
+        _meta_index[_ann_name] = _meta_anns
+
+    _CONTROLLER_CORE = frozenset({"Controller", "RestController"})
+
+    def _resolves_to_controller(name: str, visited: "set[str]") -> bool:
+        """Return True when annotation *name* transitively carries @Controller/@RestController."""
+        if name in _CONTROLLER_CORE:
+            return True
+        if name in visited:
+            return False
+        visited.add(name)
+        for parent in _meta_index.get(name, ()):
+            if _resolves_to_controller(parent, visited):
+                return True
+        return False
+
     for java_file in java_files:
         try:
             content = java_file.read_text(encoding="utf-8", errors="replace")
         except OSError:
             continue
 
-        # Only process files with REST controller or mapping annotations
-        if not any(x in content for x in ("@RestController", "@Controller", "@RequestMapping")):
-            continue
+        # Process files with direct controller/mapping annotations OR
+        # with custom annotations that transitively resolve to @Controller/@RestController.
+        _has_direct = any(x in content for x in ("@RestController", "@Controller", "@RequestMapping"))
+        if not _has_direct:
+            # Quick meta-annotation check: extract class-level annotation names
+            # (first 60 lines — before the class body opens) and resolve them.
+            _header_lines = content.splitlines()[:60]
+            _header_text = "\n".join(_header_lines)
+            _file_anns = set(_ANN_USE_RE.findall(_header_text))
+            if not any(_resolves_to_controller(a, set()) for a in _file_anns):
+                continue
 
         try:
             rel_path = str(java_file.relative_to(root)).replace("\\", "/")
@@ -2525,14 +2583,18 @@ def _extract_java_endpoints(root: "Path") -> "dict[str, Any]":
         # Extract class-level base path and locate class body start
         lines = content.splitlines()
 
-        # First pass: find class/interface declaration line index
+        # First pass: find class/interface declaration line index and extends clause.
         class_body_start = 0
+        extends_class_name: Optional[str] = None
         for i, line in enumerate(lines):
             stripped_l = line.strip()
             if (not stripped_l.startswith("//") and not stripped_l.startswith("*")
                     and ("class " in stripped_l or "interface " in stripped_l)
                     and _CLASS_RE.search(stripped_l)):
                 class_body_start = i + 1
+                _ext_m = _EXTENDS_RE_LOCAL.search(stripped_l)
+                if _ext_m:
+                    extends_class_name = _ext_m.group(1)
                 break
 
         # Second pass: extract class-level @RequestMapping path (only before class body).
@@ -2602,6 +2664,7 @@ def _extract_java_endpoints(root: "Path") -> "dict[str, Any]":
         pending_annotations: list[tuple[str, str]] = []  # (http_verb, path_suffix)
         pending_filtro: Optional[str] = None
         in_block_comment = False
+        file_own_endpoints: list[tuple] = []  # (http_verb, path_suffix, handler, filtro)
 
         for i in range(class_body_start, len(lines)):
             line = lines[i]
@@ -2654,6 +2717,7 @@ def _extract_java_endpoints(root: "Path") -> "dict[str, Any]":
                 handler = mm.group(1) if mm else ""
                 if handler and not handler.startswith("class"):
                     for http_verb, path_suffix in pending_annotations:
+                        file_own_endpoints.append((http_verb, path_suffix, handler, pending_filtro))
                         for _cb in class_bases:  # Bug #1B: one endpoint per class prefix
                             full_path = (_cb + "/" + path_suffix).replace("//", "/").rstrip("/") or "/"
                             if not full_path.startswith("/"):
@@ -2679,6 +2743,50 @@ def _extract_java_endpoints(root: "Path") -> "dict[str, Any]":
                 pending_annotations = []
                 pending_filtro = None
 
+        # Store per-class data for inheritance projection
+        _class_info[class_name] = {
+            "base_paths": class_bases,
+            "extends_class": extends_class_name,
+            "own_endpoints": file_own_endpoints,
+        }
+
+    # Fix 1: Inheritance projection — controllers whose methods are 100% inherited from a
+    # base class emit 0 endpoints above because no method-level annotations exist in their
+    # file. Walk the inheritance chain and project parent suffixes with the child's base path.
+    for _cls, _data in _class_info.items():
+        if _data["own_endpoints"]:
+            continue
+        _bases = _data["base_paths"]
+        if not _bases or _bases == [""]:
+            continue
+        _chain = _data.get("extends_class")
+        _visited: set[str] = {_cls}
+        while _chain and _chain not in _visited:
+            _visited.add(_chain)
+            _parent = _class_info.get(_chain)
+            if not _parent:
+                break
+            if _parent["own_endpoints"]:
+                for _verb, _suffix, _handler, _filtro in _parent["own_endpoints"]:
+                    for _cb in _bases:
+                        _fp = (_cb + "/" + _suffix).replace("//", "/").rstrip("/") or "/"
+                        if not _fp.startswith("/"):
+                            _fp = "/" + _fp
+                        _key = (_cls, _handler, _verb, _cb)
+                        if _key not in seen:
+                            seen.add(_key)
+                            _entry: dict[str, Any] = {
+                                "method": _verb,
+                                "path": _fp,
+                                "controller": _cls,
+                                "handler": _handler,
+                            }
+                            if _filtro:
+                                _entry["required_permission"] = _filtro
+                            endpoints.append(_entry)
+                break
+            _chain = _parent.get("extends_class")
+
     endpoints.sort(key=lambda e: (e.get("controller", ""), e.get("path", "")))
     undocumented = sum(1 for e in endpoints if "required_permission" not in e)
 

sourcecode/code_notes_analyzer.py

@@ -20,7 +20,8 @@ _SYMBOL_LOOKBACK = 25  # líneas hacia atrás para encontrar el símbolo envolve
 _SKIP_DIRS = {
     "node_modules", ".git", "__pycache__", ".venv", "venv",
     ".mypy_cache", "dist", "build", ".tox", ".eggs",
-    ".next", ".nuxt", ".output", "vendor", "coverage",
+    ".next", ".nuxt", ".output", "vendor", "vendors", "coverage",
+    "third_party", "thirdparty",
 }
 
 _CODE_EXTENSIONS = {
@@ -255,4 +256,6 @@ class CodeNotesAnalyzer:
                 # here was redundant and caused files to be silently skipped when
                 # traversal order varied (different files filled the quota first).
                 if suffix in _CODE_EXTENSIONS:
-                    _scan_source_file(entry, rel, notes, total_count)
+                    from sourcecode.path_filters import is_vendor_path as _is_vendor
+                    if not _is_vendor(rel):
+                        _scan_source_file(entry, rel, notes, total_count)

sourcecode/detectors/java.py

@@ -295,9 +295,13 @@ class JavaDetector(AbstractDetector):
         self._augment_deep_java_controllers(context, all_java)
 
         # 1. @SpringBootApplication entry: Application.java / Main.java by name
+        # Exclude test trees: test helpers like AdminApplication.java in
+        # integration/src/test/java/ must not be treated as production entrypoints.
+        from sourcecode.path_filters import is_test_path as _is_test_path
         app_candidates = [
             p for p in all_java
             if p.endswith(("Application.java", "Main.java"))
+            and not _is_test_path(p)
         ]
         entry_points: list[EntryPoint] = [
             EntryPoint(path=p, stack="java", kind="application", source="manifest")
@@ -307,7 +311,7 @@ class JavaDetector(AbstractDetector):
         # 2. Annotation-based scan: @RestController, @WebFilter, FilterRegistrationBean
         # Prioritize Controller-named files so all REST controllers are detected
         # even in large codebases where total files > _MAX_JAVA_ENTRY_SCAN.
-        _non_test = [p for p in all_java if "/test/" not in p and "/tests/" not in p]
+        _non_test = [p for p in all_java if not _is_test_path(p)]
         _ctrl_files = [p for p in _non_test if "Controller" in p]
         _other_files = [p for p in _non_test if "Controller" not in p]
         scan_candidates = _ctrl_files + _other_files[:max(0, _MAX_JAVA_ENTRY_SCAN - len(_ctrl_files))]

sourcecode/path_filters.py

@@ -0,0 +1,124 @@
+"""Shared path classification helpers used across all tools.
+
+Centralises test-path and vendor-path detection so each tool does not
+duplicate — and diverge — these heuristics.
+"""
+from __future__ import annotations
+
+_TEST_SEGMENTS = frozenset({
+    "test", "tests", "spec", "specs",
+    "test-helpers", "test_helpers", "testfixtures",
+    "it",          # integration-tests short name
+    "integrationtest", "integrationtests",
+})
+
+_VENDOR_SEGMENTS = frozenset({
+    "vendor", "vendors",
+    "third_party", "thirdparty",
+    "node_modules",
+    "external", "externals",
+    "contrib",
+})
+
+# lib/libs are vendor only for web-asset extensions.
+# Java/Kotlin/Python source in a package named "lib" is NOT vendor.
+_LIB_SEGMENTS = frozenset({"lib", "libs"})
+_WEB_ASSET_EXTS = frozenset({
+    ".js", ".jsx", ".mjs", ".cjs",
+    ".ts", ".tsx",
+    ".css", ".less", ".scss", ".sass",
+    ".json", ".map",
+})
+
+_VENDOR_PATH_FRAGMENTS = (
+    "/vendor/", "/vendors/",
+    "/third_party/", "/thirdparty/",
+    "/node_modules/",
+    "/external/", "/externals/",
+    "/contrib/",
+)
+
+_JAVA_TEST_ROOTS = (
+    "/src/test/",
+    "\\src\\test\\",
+)
+
+
+def is_test_path(path: str) -> bool:
+    """Return True when *path* is part of a test tree, not production code.
+
+    Handles:
+      - Standard Maven/Gradle layout  (src/test/java/…)
+      - Common naming conventions     (/tests/, /spec/, /it/)
+      - Java file name conventions    (FooTest.java, TestFoo.java)
+      - Python conventions            (test_foo.py, foo_test.py)
+      - JS/TS conventions             (foo.test.ts, foo.spec.ts)
+    """
+    norm = path.replace("\\", "/").lower()
+
+    # Maven/Gradle standard test root (fast path)
+    if "/src/test/" in norm:
+        return True
+
+    # Segment-based check – any directory component is a test segment
+    parts = norm.split("/")
+    for part in parts[:-1]:  # skip filename itself
+        bare = part.rstrip("/")
+        if bare in _TEST_SEGMENTS:
+            return True
+
+    # File-name conventions
+    name = parts[-1]
+    if (
+        name.startswith("test_")
+        or name.endswith("_test.py")
+        or name.endswith(".test.ts")
+        or name.endswith(".test.js")
+        or name.endswith(".spec.ts")
+        or name.endswith(".spec.js")
+        or (name.endswith("test.java") and name != "test.java")
+        or name.endswith("tests.java")
+        or (name.startswith("test") and name.endswith(".java") and len(name) > 9)
+    ):
+        return True
+
+    return False
+
+
+def is_vendor_path(path: str) -> bool:
+    """Return True when *path* is inside a vendored / third-party directory.
+
+    Handles:
+      - /vendor/, /vendors/, /third_party/, /node_modules/
+      - /lib/, /libs/ containing web assets (NOT JVM/Python source — those may
+        legitimately use "lib" as a package name)
+      - Minified JS/CSS files anywhere (*.min.js, *.min.css)
+    """
+    norm = path.replace("\\", "/").lower()
+
+    # Minified files are always vendor regardless of directory
+    if norm.endswith(".min.js") or norm.endswith(".min.css"):
+        return True
+
+    # Fast fragment check for unambiguous vendor directories
+    for frag in _VENDOR_PATH_FRAGMENTS:
+        if frag in norm:
+            return True
+
+    parts = norm.split("/")
+    dir_parts = parts[:-1]  # exclude filename
+
+    # Unambiguous vendor directory names
+    for part in dir_parts:
+        if part in _VENDOR_SEGMENTS:
+            return True
+
+    # lib/libs: vendor only for web-asset file types, not JVM/Python source
+    filename = parts[-1]
+    ext = "." + filename.rsplit(".", 1)[-1] if "." in filename else ""
+    if ext in _WEB_ASSET_EXTS:
+        for part in dir_parts:
+            if part in _LIB_SEGMENTS:
+                return True
+
+    return False

sourcecode/prepare_context.py

@@ -310,7 +310,8 @@ class RelevantFile:
     role: str    # entrypoint | source | test
     score: float
     reason: str
-    why: str = ""  # why this file matters for the specific task
+    why: str = ""                # why this file matters for the specific task
+    tier: Optional[str] = None   # fix-bug only: high | medium | low
 
 
 @dataclass
@@ -974,7 +975,33 @@ class TaskContextBuilder:
                 and (d.role or "unknown") in {"runtime", "parsing", "serialization", "observability", "infra"}
                 and d.scope not in {"dev"}
             ]
-            direct.sort(key=lambda d: (0 if d.ecosystem == primary_eco else 1, d.name.lower()))
+            # Rank by framework centrality: core infra (ORM, Spring) > serialization > other.
+            # Penalise vendored tooling (closure-compiler, shaded utilities) so that
+            # Hibernate/JPA/Solr appear before minor build-time dependencies.
+            _HIGH_SIGNAL_FRAGMENTS = (
+                "hibernate", "jpa", "spring-core", "spring-context", "spring-web",
+                "spring-boot", "spring-security", "spring-data",
+                "solr", "elasticsearch", "kafka", "redis",
+                "jackson", "gson",
+                "mybatis", "druid", "datasource",
+                "tomcat", "undertow", "netty",
+                "slf4j", "logback", "log4j",
+            )
+            _LOW_SIGNAL_FRAGMENTS = (
+                "closure-compiler", "closure-library",
+                "google-closure", "rhino",
+                "guava-gwt",
+            )
+
+            def _dep_rank(d: Any) -> tuple:
+                art = (d.name or "").lower()
+                eco_match = 0 if d.ecosystem == primary_eco else 1
+                is_high = any(frag in art for frag in _HIGH_SIGNAL_FRAGMENTS)
+                is_low = any(frag in art for frag in _LOW_SIGNAL_FRAGMENTS)
+                infra_score = 0 if is_high else (2 if is_low else 1)
+                return (eco_match, infra_score, art)
+
+            direct.sort(key=_dep_rank)
             _SKIP_DEP_KEYS = {"parent", "workspace", "resolved_version", "manifest_path"}
             key_dependencies = [
                 {k: v for k, v in asdict(d).items() if v is not None and k not in _SKIP_DEP_KEYS}
@@ -1182,6 +1209,7 @@ class TaskContextBuilder:
                 uncommitted_files=uncommitted_files,
                 code_notes=cn_notes_for_ranking if cn_notes_for_ranking else None,
                 delta_files=None,
+                symptom=symptom if task_name == "fix-bug" else None,
             )
 
         # ── 6b. review-pr: derive PR-specific impact sections from delta analysis ──
@@ -2025,6 +2053,7 @@ class TaskContextBuilder:
         uncommitted_files: Optional[set[str]] = None,
         code_notes: Optional[list] = None,
         delta_files: Optional[set[str]] = None,
+        symptom: Optional[str] = None,
     ) -> list[RelevantFile]:
         from sourcecode.ranking_engine import RankingEngine
         from sourcecode.file_classifier import FileClassifier
@@ -2043,6 +2072,11 @@ class TaskContextBuilder:
         _annotated_files: set[str] = set()
         _dominant_stack = ""
         _recently_changed_stacks: set[str] = set()
+        # Query-aware signals extracted from symptom (class names, exception types, tokens)
+        _symptom_class_names: set[str] = set()    # CamelCase class names
+        _symptom_exception_types: set[str] = set() # *Exception / *Error tokens
+        _symptom_tokens: set[str] = set()          # all lowercase tokens
+
         if task_name == "fix-bug":
             _bug_kinds = {"FIXME", "BUG", "HACK", "XXX"}
             for _n in (code_notes or []):
@@ -2068,6 +2102,19 @@ class TaskContextBuilder:
                 _dominant_stack = _stk_counts.most_common(1)[0][0]
                 _recently_changed_stacks = set(_stk_counts.keys())
 
+            # Extract structured signals from symptom text for AND-weighted ranking
+            if symptom:
+                import re as _re_bug
+                _camel_re = _re_bug.compile(r'\b([A-Z][a-zA-Z0-9]+)\b')
+                for _tok in _camel_re.findall(symptom):
+                    if _tok.endswith(("Exception", "Error", "Throwable")):
+                        _symptom_exception_types.add(_tok)
+                    else:
+                        _symptom_class_names.add(_tok)
+                _symptom_tokens = {
+                    w.lower() for w in _re_bug.split(r'[\s\W]+', symptom) if len(w) > 2
+                }
+
         scored: list[tuple[float, str, RelevantFile]] = []
 
         # For delta task, score only files changed in the specified git range.
@@ -2117,6 +2164,64 @@ class TaskContextBuilder:
             _fix_bug_why = ""
             if task_name == "fix-bug":
                 _why_parts: list[str] = []
+
+                # ── Query-aware AND-weighted signals (symptom-derived) ──
+                # These intentionally outweigh git-recency signals so that
+                # OrderServiceImpl.java ranks top-3 regardless of churn history.
+                if _symptom_class_names or _symptom_exception_types:
+                    _stem = Path(path).stem
+                    _stem_lower = _stem.lower()
+                    _matched_class = next(
+                        (c for c in _symptom_class_names if _stem_lower == c.lower()),
+                        None,
+                    )
+                    _matched_exc = next(
+                        (e for e in _symptom_exception_types if _stem_lower == e.lower()),
+                        None,
+                    )
+                    _impl_match = next(
+                        (c for c in _symptom_class_names
+                         if _stem_lower in (c.lower() + "impl", c.lower() + "service",
+                                            c.lower() + "serviceimpl", c.lower() + "helper")),
+                        None,
+                    )
+                    if _matched_class:
+                        content_boost += 3.0
+                        _why_parts.append(f"exact class match: {_stem} (+3.0)")
+                    elif _matched_exc:
+                        content_boost += 2.0
+                        _why_parts.append(f"exception class match: {_stem} (+2.0)")
+                    elif _impl_match:
+                        content_boost += 2.5
+                        _why_parts.append(f"class impl match: {_stem} (+2.5)")
+                    else:
+                        # Symbol appears anywhere in path (package adjacency)
+                        _path_class_hit = next(
+                            (c for c in _symptom_class_names if c.lower() in path_lower),
+                            None,
+                        )
+                        if _path_class_hit:
+                            content_boost += 1.0
+                            _why_parts.append(f"symbol in path: {_path_class_hit} (+1.0)")
+                        elif any(e.lower() in path_lower for e in _symptom_exception_types):
+                            content_boost += 0.8
+                            _why_parts.append("exception type in path (+0.8)")
+
+                # AND-weighted token intersection — multiple matching tokens >> single
+                if _symptom_tokens:
+                    _path_parts = set(path_lower.replace("/", " ").replace(".", " ").replace("_", " ").split())
+                    _intersection = _symptom_tokens & _path_parts
+                    _n_match = len(_intersection)
+                    if _n_match >= 3:
+                        _tok_boost = min(1.2, _n_match * 0.25)
+                        content_boost += _tok_boost
+                        _why_parts.append(f"token AND match ({_n_match} terms: {sorted(_intersection)[:3]}) (+{_tok_boost:.2f})")
+                    elif _n_match == 2:
+                        content_boost += 0.4
+                        _why_parts.append(f"token AND match (2 terms: {sorted(_intersection)}) (+0.40)")
+                    # Single-token match: no boost — avoids OR explosion
+
+                # ── Git / annotation signals ──
                 if path in _uncommitted:
                     content_boost += 0.40
                     _why_parts.append("uncommitted change (+0.40)")
@@ -2203,7 +2308,7 @@ class TaskContextBuilder:
             }
             _repo_size = len(all_paths)
             _task_budget = {
-                "fix-bug": max(20, min(40, _repo_size // 80)),
+                "fix-bug": 30,  # hard cap — prevents token explosion on large repos
                 "onboard": max(15, min(25, _repo_size // 150)),
                 "explain": max(10, min(20, _repo_size // 200)),
                 "generate-tests": max(20, min(35, _repo_size // 100)),
@@ -2271,7 +2376,21 @@ class TaskContextBuilder:
                             _covered.add(_layer)
                             _missing.discard(_layer)
 
-            return [_rf_map[p] for p in _selected if p in _rf_map]
+            result = [_rf_map[p] for p in _selected if p in _rf_map]
+
+            # Assign fix-bug tiers based on raw score (pre-normalised total)
+            if task_name == "fix-bug":
+                _score_lookup = {path: total for total, path, _ in scored}
+                for _rf in result:
+                    _s = _score_lookup.get(_rf.path, 0.0)
+                    if _s >= 4.0:
+                        _rf.tier = "high"
+                    elif _s >= 1.5:
+                        _rf.tier = "medium"
+                    else:
+                        _rf.tier = "low"
+
+            return result
         except Exception:
             return [f for _, _, f in scored[:15]]
 

sourcecode/repository_ir.py

@@ -178,8 +178,11 @@ _SPRING_OTHER: frozenset[str] = frozenset({
     "@PutMapping", "@DeleteMapping", "@PatchMapping", "@Autowired",
     "@Inject", "@Value", "@Qualifier", "@EnableWebSecurity",
     "@SpringBootApplication", "@EnableAutoConfiguration",
+    "@EventListener", "@Async", "@Scheduled", "@Cacheable", "@CacheEvict",
 })
 
+_PUBLISH_EVENT_RE = re.compile(r'\.publishEvent\s*\(\s*new\s+(\w+)\s*[(\{]')
+
 _HTTP_METHOD_MAP: dict[str, str] = {
     "@GetMapping": "GET",
     "@PostMapping": "POST",
@@ -787,6 +790,33 @@ def _build_relations(
                     evidence={"type": "structural", "value": f"member of {enclosing}"},
                 ))
 
+    # Fix 5: Event flow edges — publishEvent publishers and @EventListener subscribers.
+    # listens_to_event: method with @EventListener → resolved event parameter type(s).
+    for sym in symbols:
+        if sym.type == "method" and "@EventListener" in sym.annotations:
+            for imp_fqn in sym.imports_used:
+                edges.append(RelationEdge(
+                    from_symbol=sym.symbol,
+                    to_symbol=imp_fqn,
+                    type="listens_to_event",
+                    confidence="high",
+                    evidence={"type": "annotation", "value": "@EventListener"},
+                ))
+
+    # publishes_event: class that calls publishEvent(new XxxEvent(...)) → event type FQN.
+    _class_syms = [s for s in symbols if s.type in ("class", "interface") and "#" not in s.symbol]
+    for m in _PUBLISH_EVENT_RE.finditer(source):
+        event_simple = m.group(1)
+        event_fqn = import_map.get(event_simple, event_simple)
+        for cls_sym in _class_syms:
+            edges.append(RelationEdge(
+                from_symbol=cls_sym.symbol,
+                to_symbol=event_fqn,
+                type="publishes_event",
+                confidence="medium",
+                evidence={"type": "method_call", "value": f"publishEvent(new {event_simple})"},
+            ))
+
     seen: set[tuple[str, str, str]] = set()
     unique: list[RelationEdge] = []
     for e in edges:
@@ -1190,6 +1220,8 @@ _EDGE_REASON_TEMPLATES: dict[str, str] = {
     "contained_in": "{from_sym} is a member of {to_sym}",
     "annotated_with": "{from_sym} is annotated with {to_sym}",
     "mapped_to": "Route {to_sym} depends on {from_sym}",
+    "publishes_event": "{from_sym} publishes event {to_sym}",
+    "listens_to_event": "{from_sym} listens for event {to_sym}",
 }
 
 # Edge types to exclude from reverse impact traversal (too noisy / non-dependency semantics)
@@ -1526,13 +1558,43 @@ def _assemble(
         },
         "subsystems": subsystems,
         "change_set": change_set_out,
-        "route_surface": route_diffs or [],
+        "route_surface": _build_route_surface(sorted_syms, route_diffs),
         "audit": {
             "dropped_fields": dropped_fields,
         },
     }
 
 
+# ---------------------------------------------------------------------------
+# Route surface helper (Fix 4)
+# ---------------------------------------------------------------------------
+
+def _build_route_surface(symbols: list[SymbolRecord], route_diffs: Optional[list[dict]]) -> list[dict]:
+    """Return route surface: diffs when --since provided, else static snapshot of all endpoints."""
+    if route_diffs:
+        return route_diffs
+    routes: list[dict] = []
+    for sym in symbols:
+        if sym.symbol_kind != "endpoint":
+            continue
+        ann_name = next((a for a in sym.annotations if a in _ENDPOINT_ANNOTATIONS), None)
+        if not ann_name:
+            continue
+        args = sym.annotation_values.get(ann_name, "")
+        path = _parse_route_path(args)
+        method = _parse_route_http_method(ann_name, args)
+        if not path and not method:
+            continue
+        routes.append({
+            "symbol": sym.symbol,
+            "controller": _enclosing_class(sym.symbol),
+            "path": path,
+            "method": method or "GET",
+            "stable_id": sym.stable_id,
+        })
+    return sorted(routes, key=lambda r: (r["controller"], r["path"]))
+
+
 # ---------------------------------------------------------------------------
 # Public API
 # ---------------------------------------------------------------------------
@@ -1675,7 +1737,17 @@ def apply_ir_size_limits(
                 "remove --summary-only to restore full graph"
             ),
         }
-        out["reverse_graph"] = {}
+        # Fix 3: keep bounded reverse graph instead of wiping it.
+        full_rg: dict = ir.get("reverse_graph") or {}
+        if full_rg:
+            _rg_sorted = sorted(
+                full_rg.items(),
+                key=lambda x: sum(len(v) for v in x[1].values()),
+                reverse=True,
+            )
+            out["reverse_graph"] = dict(_rg_sorted[:50])
+        else:
+            out["reverse_graph"] = {}
         out["impact"] = {
             "global_score": (ir.get("impact") or {}).get("global_score", 0),
             "ranked_nodes": ranked[:20],
@@ -1703,10 +1775,19 @@ def apply_ir_size_limits(
 
     if kept_fqns is not None or max_edges is not None:
         if kept_fqns is not None:
-            # Priority: edges where both endpoints are kept nodes
-            priority = [e for e in edges if e["from"] in kept_fqns and e["to"] in kept_fqns]
-            rest = [e for e in edges if not (e["from"] in kept_fqns and e["to"] in kept_fqns)]
-            edges = priority + rest
+            # Fix 2: type-aware priority so semantic edges survive node truncation.
+            # Annotation strings (@Service etc.) and field FQNs are never in kept_fqns,
+            # so "both endpoints kept" drops all injects/annotated_with edges.
+            _SEMANTIC_TYPES = frozenset({"extends", "implements", "injects",
+                                         "publishes_event", "listens_to_event"})
+            _ANNOTATION_TYPES = frozenset({"annotated_with"})
+            tier1 = [e for e in edges if e["from"] in kept_fqns and e["type"] in _SEMANTIC_TYPES]
+            tier2 = [e for e in edges if e["from"] in kept_fqns and e["type"] in _ANNOTATION_TYPES]
+            tier3 = [e for e in edges
+                     if e["from"] in kept_fqns and e["to"] in kept_fqns and e["type"] == "imports"]
+            _seen_e = {(e["from"], e["to"], e["type"]) for e in tier1 + tier2 + tier3}
+            tier4 = [e for e in edges if (e["from"], e["to"], e["type"]) not in _seen_e]
+            edges = tier1 + tier2 + tier3 + tier4
         if max_edges is not None:
             edges = edges[:max_edges]
 

{sourcecode-1.31.5.dist-info → sourcecode-1.31.7.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sourcecode
-Version: 1.31.5
+Version: 1.31.7
 Summary: Deterministic codebase context for AI coding agents
 License:                                  Apache License
                                    Version 2.0, January 2004
@@ -225,7 +225,7 @@ Description-Content-Type: text/markdown
 
 **Deterministic, behavior-aware codebase context for AI agents and PR review.**
 
-![Version](https://img.shields.io/badge/version-1.31.5-blue)
+![Version](https://img.shields.io/badge/version-1.31.7-blue)
 ![Python](https://img.shields.io/badge/python-3.10%2B-green)
 
 ---
@@ -261,7 +261,7 @@ pipx install sourcecode
 
 ```bash
 sourcecode version
-# sourcecode 1.31.5
+# sourcecode 1.31.7
 ```
 
 ---

{sourcecode-1.31.5.dist-info → sourcecode-1.31.7.dist-info}/RECORD

@@ -1,11 +1,11 @@
-sourcecode/__init__.py,sha256=INTl4i-9L8tLlPUGq6eihBx0YEE8OQu6cHtsMNAu_vc,103
+sourcecode/__init__.py,sha256=8R-nJsTbK4mRxB0Ix8W0xEdhYCBpK8R-2dyS_nBxiDc,103
 sourcecode/adaptive_scanner.py,sha256=XffluXKzJUXrMtjEiAOnSNPZnztdIcts17T9ouHeID0,10521
 sourcecode/architecture_analyzer.py,sha256=MyBa0Hf5HmkudZQDLKrjcWDKETXETXl0mQX1swtTwAA,39091
 sourcecode/architecture_summary.py,sha256=z34_6v7cSwy98cof2UVciGho7SCrZ93tiqMmq5WNzRQ,20405
 sourcecode/ast_extractor.py,sha256=XgrZg2DcWcUm9r87cRG3KGO7IK2TIL_N-CvhSbUmmh4,49901
 sourcecode/classifier.py,sha256=-0t0HLc9L9UleMLfclfLM3AXhBjUb_AYyBPDbvgWtac,7755
-sourcecode/cli.py,sha256=IDe_6kp4hgqPX6YLA-Gu6-rlwbyT4iXfuYJhG4pmhe0,126525
-sourcecode/code_notes_analyzer.py,sha256=y1MJBnPZHYp4i6cQCXUb9ATIyifS_qMQWjw_8lPkpsU,9215
+sourcecode/cli.py,sha256=L0KDRruMogpw9kq7trZ7dmWQuFC746GIEN1oym6LmCE,131802
+sourcecode/code_notes_analyzer.py,sha256=EJemNCNc9Dn-1RZYu-aNbK0ELzmsyC4s6FdHi3XyNEI,9392
 sourcecode/confidence_analyzer.py,sha256=ZUn-Nywi5TEQcuozqK_vfOyPT-a1dYYO42elAtVFV-k,16412
 sourcecode/context_scorer.py,sha256=QpChSpsmaAYz91rXA4Ue5xzQmNz_ZboZN09YOHScq1U,14679
 sourcecode/context_summarizer.py,sha256=CiQrfBEzun949bWvmLabWoj2HhPn6Lw62ofqnsy0FlQ,6503
@@ -21,14 +21,15 @@ sourcecode/flow_analyzer.py,sha256=dSiuY4w49k29jW_EPXUOND9B5uVbuCA7kjnuHi-pIWA,2
 sourcecode/git_analyzer.py,sha256=0Gyj-vMpIIN4nfriKXVRouNYBeJ59s6pQDX2Xu9Pq-U,13177
 sourcecode/graph_analyzer.py,sha256=iUK-7pSV-cvGqqD2hENdYmhnm0wcXFEyK-xnu5ul8OU,62515
 sourcecode/metrics_analyzer.py,sha256=m0ENgtqKeBL17kUIK3fmGkgo7UfXBNHxCMj0H_Y5K7c,22750
+sourcecode/path_filters.py,sha256=ROFRQ8eSLBEMiixK9f45-RO7um4VEEcjoD5AA4I427I,3739
 sourcecode/pr_comment_renderer.py,sha256=smHslxiG14lrytCkq5nFrFu-qTHgA-t-LFYfdrfjz2o,14423
-sourcecode/prepare_context.py,sha256=Eid3Wmh4lmtr20pmobcCU0yzjRXhFql2rEmfGUgkFpU,183214
+sourcecode/prepare_context.py,sha256=-9kTYuPhwr79mF6lNe9tI7glGAU_d84GyVurHceGroo,189427
 sourcecode/progress.py,sha256=qn30sWaHOkjTgXsSBmiPkz7Rsbwc5oSlIe6JNEMYp_k,3149
 sourcecode/ranking_engine.py,sha256=ZAucq_YX2KkWUuAZf4P0lhtQ_38vEFnUhuGtSZd1S0E,12970
 sourcecode/redactor.py,sha256=xuGcadGEHaPw4qZXlMDvzMCsr4VOkdp3oBQptHyJk8c,2884
 sourcecode/relevance_scorer.py,sha256=MYF4FFkveAQps9SmTeTlh6ODiBz2F--_hWNeHMLtUHQ,8405
 sourcecode/repo_classifier.py,sha256=FG1vaWKdWXsWdl-S8hjVMiTqcwgaRXkDyvK4rPcOGtQ,22681
-sourcecode/repository_ir.py,sha256=KH5EehbjOh8ZwwTHcbzrAHiKDoquO49wgSvCX4bVq5k,64391
+sourcecode/repository_ir.py,sha256=fLt33wadCMY77jd34YwkYbUT5TyNu2G0LX0x76Krsvo,68348
 sourcecode/runtime_classifier.py,sha256=zWX3r3HCKHc-qtIobErOa8aKMmaoPYREtJKvPcBGPjQ,14792
 sourcecode/scanner.py,sha256=WdOQ78mMzjR1NjmKTlbxdgwinnCTfAhxCVLBEFQiFHU,8899
 sourcecode/schema.py,sha256=fj3BZ3IcnNV4j21BFIEvz8Qnw_vZoqIbzzRg-qQ-nd0,24530
@@ -46,7 +47,7 @@ sourcecode/detectors/elixir.py,sha256=jCpvt5Yi6jvplc80ovRtWh17q-11ZGo9qX7o8b57TJ
 sourcecode/detectors/go.py,sha256=2r66uRQfeTWsqxr4HDhT6vExZErby0t46QXLHVBRv9w,2782
 sourcecode/detectors/heuristic.py,sha256=7cRxrip4yIaggYzZJB6ef8yHKh-gHgiH_pXMFcjlyFU,3723
 sourcecode/detectors/hybrid.py,sha256=IGFRUVsAZ1ooRlFdznCeJAV6vy1yVDx-VyghvLtddXc,9101
-sourcecode/detectors/java.py,sha256=ldvaDZJADAKslOpS5qJ0bxexgeY6h_4Yx4NCtHio7J8,24203
+sourcecode/detectors/java.py,sha256=O2JdznVYv5364GSQExksYLsAi0pvDUW9ZowpsL0xLgM,24451
 sourcecode/detectors/jvm_ext.py,sha256=EgHJ5W8EE-ZTN9V607mVzohyKgZE8Mc2jCi-DF8RAZU,2616
 sourcecode/detectors/nodejs.py,sha256=Hg3Gmr7yIMJFiLoDwOTk2wtu00wxIs6kZf-oQujTFUA,13187
 sourcecode/detectors/parsers.py,sha256=ugPg8yNUf0Ai1gA7Fnn6wAkYGFjTxRodSP3IeViYJJ4,2290
@@ -72,8 +73,8 @@ sourcecode/telemetry/consent.py,sha256=wLMvGNJeSSyZoNkQXpoUioY6mMv4Qdvuw7S9jAEWn
 sourcecode/telemetry/events.py,sha256=oEvvulfsv5GIDWG2174gSS6tNB95w38AIYiYeifGKlE,2294
 sourcecode/telemetry/filters.py,sha256=Asa71oRl7q3Wt_FMwuufIZJFzSYdgRNKS8LHCIyFeYE,4805
 sourcecode/telemetry/transport.py,sha256=KJeIPCPWMdmbCP3ySGs2iUlia34U6vWne2dZsUezesw,1560
-sourcecode-1.31.5.dist-info/METADATA,sha256=UcfeOr7ZaZWAjC3WByNVOIoPw3VlJPF3pS9Fo7f-Ch0,29083
-sourcecode-1.31.5.dist-info/WHEEL,sha256=QccIxa26bgl1E6uMy58deGWi-0aeIkkangHcxk2kWfw,87
-sourcecode-1.31.5.dist-info/entry_points.txt,sha256=ex3F9rmbXeyDIoFQHtkEqTsKSaJow8F0LrVu8XfIktQ,57
-sourcecode-1.31.5.dist-info/licenses/LICENSE,sha256=7DdHrU9Z_3e7dSvq4ISijZNjnuHo5NIHNiHDouMQ9JU,10491
-sourcecode-1.31.5.dist-info/RECORD,,
+sourcecode-1.31.7.dist-info/METADATA,sha256=O08_eBOxRxeDd8UkUKyIrW6haJHuQvAddgiFAEma0dQ,29083
+sourcecode-1.31.7.dist-info/WHEEL,sha256=QccIxa26bgl1E6uMy58deGWi-0aeIkkangHcxk2kWfw,87
+sourcecode-1.31.7.dist-info/entry_points.txt,sha256=ex3F9rmbXeyDIoFQHtkEqTsKSaJow8F0LrVu8XfIktQ,57
+sourcecode-1.31.7.dist-info/licenses/LICENSE,sha256=7DdHrU9Z_3e7dSvq4ISijZNjnuHo5NIHNiHDouMQ9JU,10491
+sourcecode-1.31.7.dist-info/RECORD,,