cg-code-graph 0.17.0__tar.gz → 0.18.0__tar.gz

{cg_code_graph-0.17.0/cg_code_graph.egg-info → cg_code_graph-0.18.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cg-code-graph
-Version: 0.17.0
+Version: 0.18.0
 Summary: cg: a code graph of routes, calls, data access and platform boundaries for agents and humans (CLI + MCP server)
 Author: cyberchronos00
 License-Expression: MIT
@@ -17,6 +17,7 @@ Requires-Dist: tree-sitter-rust>=0.23
 Requires-Dist: tree-sitter-c>=0.23
 Requires-Dist: tree-sitter-cpp>=0.23
 Requires-Dist: tree-sitter-kotlin>=1.0
+Requires-Dist: tree-sitter-java>=0.23
 Requires-Dist: tree-sitter-swift<0.7.4,>=0.6
 Provides-Extra: dev
 Requires-Dist: pytest>=8; extra == "dev"
@@ -76,11 +77,12 @@ dependents: 18 nodes  (code: 6, entry points: 9, other: 3)
 | ![TypeScript](https://img.shields.io/badge/TypeScript-3178C6?logo=typescript&logoColor=white) ![JavaScript](https://img.shields.io/badge/JavaScript-F7DF1E?logo=javascript&logoColor=black) | Nuxt / Vue, NestJS, Next.js, Astro, Express, Fastify, Koa, Hono | exact (TypeScript compiler API) | [TS frameworks](https://code-graph.cyberchronos00.workers.dev/docs/ts-frameworks) |
 | ![PHP](https://img.shields.io/badge/PHP-777BB4?logo=php&logoColor=white) | Laravel (routes and middleware, queues, broadcasting) | exact (nikic/php-parser; needs PHP 8.2+) | [PHP](https://code-graph.cyberchronos00.workers.dev/docs/php) |
 | ![Dart](https://img.shields.io/badge/Dart-0175C2?logo=dart&logoColor=white) | Flutter | exact (Dart analyzer; needs the Dart SDK) | [Platforms](https://code-graph.cyberchronos00.workers.dev/docs/platforms) |
-| ![Kotlin](https://img.shields.io/badge/Kotlin-7F52FF?logo=kotlin&logoColor=white) | Android, Kotlin Multiplatform | heuristic; exact with scip-java (opt-in) | [Kotlin](https://code-graph.cyberchronos00.workers.dev/docs/kotlin) |
+| ![Kotlin](https://img.shields.io/badge/Kotlin-7F52FF?logo=kotlin&logoColor=white) | Android, Kotlin Multiplatform, Spring | heuristic; exact with scip-java (opt-in) | [Kotlin](https://code-graph.cyberchronos00.workers.dev/docs/kotlin) |
+| ![Java](https://img.shields.io/badge/Java-ED8B00?logo=openjdk&logoColor=white) | Spring (routes, guards, beans, tables, clients) | heuristic; exact with scip-java (opt-in) | [Java](https://code-graph.cyberchronos00.workers.dev/docs/java) |
 | ![Swift](https://img.shields.io/badge/Swift-F05138?logo=swift&logoColor=white) | iOS / SwiftPM | heuristic; exact with the Swift index store (opt-in) | [Swift](https://code-graph.cyberchronos00.workers.dev/docs/swift) |
 | ![Rust](https://img.shields.io/badge/Rust-000000?logo=rust&logoColor=white) ![C](https://img.shields.io/badge/C-A8B9CC?logo=c&logoColor=black) ![C++](https://img.shields.io/badge/C++-00599C?logo=cplusplus&logoColor=white) | crates / modules, entry points, FFI, `#[cfg]` / `#if` gates | heuristic; exact with rust-analyzer / scip-clang | [Rust, C and C++](https://code-graph.cyberchronos00.workers.dev/docs/native) |
 
-Go, Java and others: import a SCIP index with `cg index --scip FILE`. Run `cg doctor` to see exact / heuristic per language on your machine.
+Go and others: import a SCIP index with `cg index --scip FILE`. Run `cg doctor` to see exact / heuristic per language on your machine.
 
 ## What you get
 

{cg_code_graph-0.17.0 → cg_code_graph-0.18.0}/README.md

@@ -50,11 +50,12 @@ dependents: 18 nodes  (code: 6, entry points: 9, other: 3)
 | ![TypeScript](https://img.shields.io/badge/TypeScript-3178C6?logo=typescript&logoColor=white) ![JavaScript](https://img.shields.io/badge/JavaScript-F7DF1E?logo=javascript&logoColor=black) | Nuxt / Vue, NestJS, Next.js, Astro, Express, Fastify, Koa, Hono | exact (TypeScript compiler API) | [TS frameworks](https://code-graph.cyberchronos00.workers.dev/docs/ts-frameworks) |
 | ![PHP](https://img.shields.io/badge/PHP-777BB4?logo=php&logoColor=white) | Laravel (routes and middleware, queues, broadcasting) | exact (nikic/php-parser; needs PHP 8.2+) | [PHP](https://code-graph.cyberchronos00.workers.dev/docs/php) |
 | ![Dart](https://img.shields.io/badge/Dart-0175C2?logo=dart&logoColor=white) | Flutter | exact (Dart analyzer; needs the Dart SDK) | [Platforms](https://code-graph.cyberchronos00.workers.dev/docs/platforms) |
-| ![Kotlin](https://img.shields.io/badge/Kotlin-7F52FF?logo=kotlin&logoColor=white) | Android, Kotlin Multiplatform | heuristic; exact with scip-java (opt-in) | [Kotlin](https://code-graph.cyberchronos00.workers.dev/docs/kotlin) |
+| ![Kotlin](https://img.shields.io/badge/Kotlin-7F52FF?logo=kotlin&logoColor=white) | Android, Kotlin Multiplatform, Spring | heuristic; exact with scip-java (opt-in) | [Kotlin](https://code-graph.cyberchronos00.workers.dev/docs/kotlin) |
+| ![Java](https://img.shields.io/badge/Java-ED8B00?logo=openjdk&logoColor=white) | Spring (routes, guards, beans, tables, clients) | heuristic; exact with scip-java (opt-in) | [Java](https://code-graph.cyberchronos00.workers.dev/docs/java) |
 | ![Swift](https://img.shields.io/badge/Swift-F05138?logo=swift&logoColor=white) | iOS / SwiftPM | heuristic; exact with the Swift index store (opt-in) | [Swift](https://code-graph.cyberchronos00.workers.dev/docs/swift) |
 | ![Rust](https://img.shields.io/badge/Rust-000000?logo=rust&logoColor=white) ![C](https://img.shields.io/badge/C-A8B9CC?logo=c&logoColor=black) ![C++](https://img.shields.io/badge/C++-00599C?logo=cplusplus&logoColor=white) | crates / modules, entry points, FFI, `#[cfg]` / `#if` gates | heuristic; exact with rust-analyzer / scip-clang | [Rust, C and C++](https://code-graph.cyberchronos00.workers.dev/docs/native) |
 
-Go, Java and others: import a SCIP index with `cg index --scip FILE`. Run `cg doctor` to see exact / heuristic per language on your machine.
+Go and others: import a SCIP index with `cg index --scip FILE`. Run `cg doctor` to see exact / heuristic per language on your machine.
 
 ## What you get
 

{cg_code_graph-0.17.0 → cg_code_graph-0.18.0}/cg_code_graph/__init__.py

@@ -1,2 +1,2 @@
 """code-graph: deterministic code knowledge graph (beta)."""
-__version__ = "0.17.0"
+__version__ = "0.18.0"

{cg_code_graph-0.17.0 → cg_code_graph-0.18.0}/cg_code_graph/agent_rules.py

@@ -44,6 +44,8 @@ Pipe long output through `head` or `tail`:
     cg impact <symbol> --db out/graph.db | head -n 40
 While iterating, run one file: `pytest -q -x tests/test_x.py`.
 Run the full suite once at the end: `pytest -q 2>&1 | tail -n 30`.
+
+A new language or framework updates README's Supported languages table, its docs page, `docs-site/app/utils/docs-nav.ts`, and `docs-site/app/components/HomeStacks.vue` in the same change.
 """
 
 

{cg_code_graph-0.17.0 → cg_code_graph-0.18.0}/cg_code_graph/cli.py

@@ -118,9 +118,9 @@ def main(argv=None):
     p.add_argument("--dry-run", action="store_true", help="list what would be removed, delete nothing")
     p.add_argument("--json", action="store_true")
     p = sub.add_parser("setup", help="install the Node / PHP / Dart extractor dependencies (into the user cache; "
-                                     "otherwise done on the first index)")
+                                     "otherwise done on the first index). `java` prints the JDK 17+ / scip-java hint")
     p.add_argument("languages", nargs="*", metavar="LANG",
-                   help="typescript, php, dart (default: every one whose toolchain is installed)")
+                   help="typescript, php, dart, java (default: typescript, php, dart whose toolchain is installed)")
     p.add_argument("--quiet", action="store_true")
     p.add_argument("--prune", action="store_true",
                    help="remove extractor installs this version does not use (left by updates that changed a lock "
@@ -408,9 +408,9 @@ def main(argv=None):
         return 0
     if a.cmd == "setup":
         from .doctor import setup
-        bad = [x for x in a.languages if x not in ("typescript", "php", "dart")]
+        bad = [x for x in a.languages if x not in ("typescript", "php", "dart", "java")]
         if bad:
-            print(f"cg setup: unknown language {bad[0]!r} (typescript, php, dart)", file=sys.stderr)
+            print(f"cg setup: unknown language {bad[0]!r} (typescript, php, dart, java)", file=sys.stderr)
             return 2
         return setup(a.languages or None, quiet=a.quiet)
     if a.cmd == "coverage":

{cg_code_graph-0.17.0 → cg_code_graph-0.18.0}/cg_code_graph/core/cache.py

@@ -1,7 +1,7 @@
 """The per-user cache: where it lives, what is in it, and `cg clean`.
 
 Root (one lookup for every user: extractors, SCIP outputs, parse caches, the Swift build directory):
-$CG_CACHE (older names still work through 0.17.x; see core/env.py), else
+$CG_CACHE (older names still work through 0.18.x; see core/env.py), else
 %LOCALAPPDATA%\\cg on Windows, else $XDG_CACHE_HOME/cg, else ~/.cache/cg. The first use of that
 default moves a leftover `codegraph` directory to `cg` and drops `swift-build/` (SwiftPM build
 directories embed absolute paths). An explicit cache path is used as given and is not moved.

{cg_code_graph-0.17.0 → cg_code_graph-0.18.0}/cg_code_graph/core/detect.py

@@ -2,9 +2,12 @@
 from __future__ import annotations
 
 import json
+import os
 import re
 from pathlib import Path
 
+from ..presets import skip_dirs
+
 MARKERS = {
     # language: files/globs that indicate it
     "php": ["composer.json", "artisan"],
@@ -80,4 +83,34 @@ def detect(root: Path) -> dict:
                 fw[extra] = {}
     if "typescript" in deps:
         langs.setdefault("typescript", ["package.json:typescript"])
+    if _spring_build(root):
+        fw["spring"] = {}
     return {"languages": langs, "frameworks": fw}
+
+
+def _spring_build(root: Path) -> bool:
+    """Spring Boot Gradle plugin or a ``spring-boot-starter-*`` dependency in Gradle / Maven files.
+
+    The root build file and one module level (``app/build.gradle.kts``). Deeper trees stay out so a
+    monorepo's example apps do not mark the repository root.
+    """
+    names = {"build.gradle", "build.gradle.kts", "pom.xml"}
+    rx = re.compile(r"org\.springframework\.boot|spring-boot-starter(?:-[\w.-]+)?")
+    skip = skip_dirs("common") | skip_dirs("common", "scan_skip_dirs")
+    for dp, dn, fn in os.walk(root):
+        rel = os.path.relpath(dp, root)
+        depth = 0 if rel == "." else len(Path(rel).parts)
+        if depth > 1:
+            dn[:] = []
+            continue
+        dn[:] = [d for d in dn if d not in skip and not d.startswith(".")]
+        for name in fn:
+            if name not in names:
+                continue
+            try:
+                text = (Path(dp) / name).read_text(errors="replace")[:200_000]
+            except OSError:
+                continue
+            if rx.search(text):
+                return True
+    return False

{cg_code_graph-0.17.0 → cg_code_graph-0.18.0}/cg_code_graph/core/env.py

@@ -1,4 +1,4 @@
-"""Process environment: ``CG_*`` names, with ``CODEGRAPH_*`` aliases through 0.17.x.
+"""Process environment: ``CG_*`` names, with ``CODEGRAPH_*`` aliases through 0.18.x.
 
 ``get`` reads ``os.environ`` on every call. A legacy name that is set produces one stderr
 warning per name per process. ``CG_*`` wins when both are set. ``MCP_TOOLS`` has no legacy
@@ -13,7 +13,8 @@ import sys
 SUFFIXES = (
     "CACHE", "CACHE_DIR", "CARGO", "CFAMILY", "COMPDB", "C_MASK_ANNOTATIONS", "C_MAX_MACRO_REFS",
     "C_SCIP", "C_SCIP_FILE", "EXCLUDE_DIRS", "INCLUDE_DIRS", "INDEXER_TIMEOUT", "JOBS",
-    "KOTLIN_SCIP", "KOTLIN_SCIP_FILE", "LIBINDEXSTORE", "MAX_FILE_BYTES", "NODE", "NO_CACHE",
+    "JAVA_SCIP", "JAVA_SCIP_FILE", "KOTLIN_SCIP", "KOTLIN_SCIP_FILE", "LIBINDEXSTORE",
+    "MAX_FILE_BYTES", "NODE", "NO_CACHE",
     "NO_CARGO", "NO_HOOKS", "NO_STALE_CHECK", "RUST_ANALYZER", "RUST_BUILD_SCRIPTS", "RUST_SCIP",
     "RUST_SCIP_FILE", "RUST_TARGETS", "SCIP_CLANG", "SCIP_JAVA", "SWIFT", "SWIFT_INDEX",
     "SWIFT_INDEX_STORE",
@@ -34,9 +35,9 @@ def _warn(old: str, new: str, *, both: bool) -> None:
         return
     _warned.add(old)
     if both:
-        extra = f"{new} is set too and wins; the CODEGRAPH_* names are removed in 0.18.0"
+        extra = f"{new} is set too and wins; the CODEGRAPH_* names are removed in 0.19.0"
     else:
-        extra = "the CODEGRAPH_* names are removed in 0.18.0"
+        extra = "the CODEGRAPH_* names are removed in 0.19.0"
     print(f"cg: {old} is deprecated, use {new} ({extra})", file=sys.stderr)
 
 

{cg_code_graph-0.17.0 → cg_code_graph-0.18.0}/cg_code_graph/core/syntax_errors.py

@@ -20,6 +20,8 @@ _SWIFT_MODS = (r"(?:(?:public|private|fileprivate|internal|open|package|static|f
 DECL = {
     "swift": re.compile(r"^[ \t]*(?:@[\w.]+(?:\([^)\n]*\))?\s+)*" + _SWIFT_MODS +
                         r"(func|init|class|struct|enum|protocol|extension|actor)\b[ \t]*([A-Za-z_]\w*)?", re.M),
+    "java": re.compile(r"^[ \t]*(?:(?:public|private|protected|static|final|abstract|sealed|native|synchronized|"
+                       r"default|strictfp)\s+)*(class|interface|enum|record)\s+([A-Za-z_]\w*)", re.M),
     "kotlin": re.compile(r"^[ \t]*(?:@[\w.]+(?:\([^)\n]*\))?\s+)*(?:(?:public|private|internal|protected|open|abstract|"
                          r"override|suspend|inline|data|sealed|enum|annotation|inner|value|companion|actual|expect|"
                          r"operator|infix|tailrec|external)\s+)*(fun|class|interface|object)\b[ \t]*"

{cg_code_graph-0.17.0 → cg_code_graph-0.18.0}/cg_code_graph/coverage.py

@@ -22,13 +22,14 @@ SUPPORTED = {
     "rust": (".rs",),
     "c_cpp": (".c", ".h", ".cc", ".cpp", ".cxx", ".c++", ".hpp", ".hh", ".hxx", ".h++", ".ipp", ".inl"),
     "kotlin": (".kt", ".kts"),
+    "java": (".java",),
     "swift": (".swift",),
 }
-# source types without a native plugin (go / java can be imported from a SCIP index). A generic "looks like source" rule:
+# source types without a native plugin (go can be imported from a SCIP index). A generic "looks like source" rule:
 # text source extensions of programming / scripting languages, plus a shebang for extensionless scripts (SHEBANGS).
 # Data, markup, config and asset extensions are not listed, so they never count as unsupported source.
 UNSUPPORTED = {
-    ".go": "go", ".java": "java", ".rb": "ruby", ".cs": "csharp",
+    ".go": "go", ".rb": "ruby", ".cs": "csharp",
     ".scala": "scala", ".ex": "elixir", ".exs": "elixir", ".m": "objective-c", ".mm": "objective-c", ".lua": "lua",
     ".pl": "perl", ".pm": "perl", ".clj": "clojure", ".erl": "erlang", ".hrl": "erlang", ".hs": "haskell", ".fs": "fsharp",
     ".fsx": "fsharp", ".groovy": "groovy", ".r": "r", ".jl": "julia", ".zig": "zig", ".sol": "solidity",
@@ -53,7 +54,7 @@ BUCKET_TEXT = {"parse_failed": "parse failed", "skipped_oversize": "over size li
 BUCKET_SHORT = {"parse_failed": "parse failed", "skipped_oversize": "over size limit", "unmapped": "unmapped",
                 "excluded": "excluded"}
 LANG_LABEL = {"php": "PHP", "typescript": "TypeScript / JavaScript", "python": "Python", "dart": "Dart", "rust": "Rust",
-              "c_cpp": "C / C++", "kotlin": "Kotlin", "swift": "Swift"}
+              "c_cpp": "C / C++", "kotlin": "Kotlin", "java": "Java", "swift": "Swift"}
 MAX_PATHS = 500      # file paths stored per bucket in the index (counts are always exact)
 SHOW_PATHS = 5       # shown per bucket by default (`cg coverage --all-files` / coverage(all_files=true) for all)
 HINTS = {
@@ -73,11 +74,15 @@ HINTS = {
              "CG_SWIFT_INDEX=1 (SwiftPM: runs `swift build --enable-index-store`; needs a Swift toolchain) or "
              "CG_SWIFT_INDEX_STORE to an existing index store (docs/swift.md#exact-mode)",
     "go": "no native plugin: index with scip-go and pass `--scip index.scip`",
-    "java": "no native plugin: index with scip-java and pass `--scip index.scip`",
+    "java": "heuristic mode (tree-sitter syntax layer, receiver-type call resolution){layer}. "
+            "For compiler-resolved references set CG_JAVA_SCIP=1 (or CG_KOTLIN_SCIP=1; one scip-java run serves both) "
+            "or pass `--scip index.scip`. install.sh --with java installs scip-java (JDK 17, 21 or 25). "
+            "docs/java.md#exact-mode",
 }
 # tree-sitter modules of the syntax layer per language: the hint names the ones missing (none: no install hint, #75)
 LAYER_MODULES = {"rust": ("tree_sitter", "tree_sitter_rust"), "c_cpp": ("tree_sitter", "tree_sitter_c", "tree_sitter_cpp"),
-                 "kotlin": ("tree_sitter", "tree_sitter_kotlin"), "swift": ("tree_sitter", "tree_sitter_swift")}
+                 "kotlin": ("tree_sitter", "tree_sitter_kotlin"), "java": ("tree_sitter", "tree_sitter_java"),
+                 "swift": ("tree_sitter", "tree_sitter_swift")}
 
 
 def hint(lang: str) -> str | None:
@@ -111,7 +116,7 @@ class Scan:
         self.counts: Counter = Counter()
         self.paths: dict[str, list[str]] = {}
         self.scripts: Counter = Counter()
-        self.bridge_paths: list[str] = []      # Java / ObjC files: no language plugin, scanned for bridge receivers
+        self.bridge_paths: list[str] = []      # Java / ObjC files scanned for bridge receivers (Java is also a language plugin)
 
     def files(self, exts) -> list[str]:
         return [f for e in exts for f in self.paths.get(e, [])]
@@ -175,7 +180,9 @@ def scan_tree(root: str | Path, rules=None, classifier=None) -> Scan:
                 sc.counts[ext] += 1
                 if ext in _SUPPORTED_EXTS:
                     sc.paths.setdefault(ext, []).append(rel_dir + fn)
-                elif ext in (".java", ".m", ".mm"):
+                # .java is a language plugin and still a bridge receiver source (Android
+                # @CapacitorPlugin / @ReactMethod). .m / .mm stay bridge-only until they have one.
+                if ext == ".java" or (ext in (".m", ".mm") and ext not in _SUPPORTED_EXTS):
                     sc.bridge_paths.append(rel_dir + fn)
             elif not fn.startswith(".") and is_real_file(p):
                 lang = _shebang(p)
@@ -271,6 +278,15 @@ def _status(lang: str, st: dict | None) -> tuple[str, str | None]:
             part += ("; skipped modules: " + ", ".join(m["module"] for m in sc["skipped_modules"][:5])
                      + f" ({sc['skipped_modules'][0]['reason']})")
         return "exact", f"scip-java index ({sc.get('source', 'scip')}){part}"
+    if lang == "java" and mode == "heuristic":
+        why = (st.get("scip") or {}).get("status") if isinstance(st.get("scip"), dict) else None
+        return "heuristic", ("tree-sitter syntax layer with receiver-type call resolution; no compiler index"
+                             + (f": {why}" if why else ""))
+    if lang == "java" and mode == "scip":
+        sc = st.get("scip") if isinstance(st.get("scip"), dict) else {}
+        n, k = st.get("scip_files"), st.get("files")
+        part = f"; {k - n} of {k} Java files not in the index keep heuristic calls" if n is not None and k and n < k else ""
+        return "exact", f"scip-java index ({sc.get('source', 'scip')}){part}"
     if lang == "typescript" and st.get("program_files") == 0 and not st.get("nodes"):
         return "not_indexed", "the TypeScript plugin ran but found no source files (tsconfig include / source dirs)"
     return "exact", None
@@ -349,10 +365,6 @@ def compute(root: str | Path, plugins: dict, scip_imported: bool = False, report
                 other[lang]["reason"] = st.get("reason")
             elif "status" not in st:
                 other[lang]["status"] = "exact"  # a SCIP indexer ran
-    kj = ((plugins.get("kotlin") or {}).get("java") or {})
-    if kj.get("documents") and "java" in other and other["java"]["status"] == "unsupported":
-        other["java"].update(status="scip", reason=f"{kj['documents']} Java file(s) imported from the Kotlin build's "
-                                                   "scip-java index (Kotlin exact mode)")
     for o in other.values():
         if o["status"] == "unsupported":
             if scip_imported and o["language"] in ("go", "java"):
@@ -632,7 +644,7 @@ def render_summary(covs: dict[str, dict | None], db: str | None = None) -> str:
                 line += f"; fix: {e['hint']}"
             out.append(line)
         for e in (cov or {}).get("languages", []):        # the mode an exact-capable language ran in, and why
-            if e["language"] in ("kotlin", "swift") and not _is_gap(e) and e.get("reason"):
+            if e["language"] in ("kotlin", "swift", "java") and not _is_gap(e) and e.get("reason"):
                 out.append(f"  {e['language']} {e['files']} {e['status']}: {e['reason']}")
         se = [e for e in (cov or {}).get("languages", []) if e.get("syntax_errors")]
         if se:
@@ -691,7 +703,7 @@ def render(covs: dict[str, dict | None], all_files: bool = False) -> str:
         for e in (cov or {}).get("languages", []):
             out += syntax_error_lines(e, all_files)
         for e in (cov or {}).get("languages", []):     # which mode an exact-capable language ran in, and why
-            if e["language"] in ("kotlin", "swift") and not _is_gap(e) and e.get("reason"):
+            if e["language"] in ("kotlin", "swift", "java") and not _is_gap(e) and e.get("reason"):
                 out.append(f"  {e['language']}: {e['files']} files {e['status']}: {e['reason']}")
         if all_files:
             for e in (cov or {}).get("languages", []):

{cg_code_graph-0.17.0 → cg_code_graph-0.18.0}/cg_code_graph/doctor.py

@@ -24,13 +24,15 @@ INSTALL_DOC = "docs/install.md"
 REPO = "https://github.com/cyberchronos00/code-graph"
 GRAMMARS = {"rust": "tree_sitter_rust", "c_cpp": "tree_sitter_c", "kotlin": "tree_sitter_kotlin", "swift": "tree_sitter_swift"}
 PIP_NAMES = {"tree_sitter": "tree-sitter", "tree_sitter_rust": "tree-sitter-rust", "tree_sitter_c": "tree-sitter-c",
-             "tree_sitter_cpp": "tree-sitter-cpp", "tree_sitter_kotlin": "tree-sitter-kotlin", "tree_sitter_swift": "tree-sitter-swift"}
+             "tree_sitter_cpp": "tree-sitter-cpp", "tree_sitter_kotlin": "tree-sitter-kotlin",
+             "tree_sitter_java": "tree-sitter-java", "tree_sitter_swift": "tree-sitter-swift"}
 
 
 # language -> the plugin packages its `cg index` run loads (besides the indexer, which loads all of them)
 PLUGIN_PACKAGES = {"python": ("python", "pyweb", "django"), "typescript": ("ts", "tsweb", "nuxt", "nest", "nextjs", "express"),
                    "php": ("php", "laravel"), "dart": ("dart", "flutter"), "rust": ("rust", "native", "scip"),
-                   "c_cpp": ("cfamily", "native", "scip"), "kotlin": ("kotlin", "scip"), "swift": ("swift",)}
+                   "c_cpp": ("cfamily", "native", "scip"), "kotlin": ("kotlin", "scip"), "java": ("java",),
+                   "swift": ("swift",)}
 
 
 def _run(tool: str | None, args=("--version",)) -> tuple[str | None, bool]:
@@ -222,6 +224,34 @@ def _languages(tools: dict, root: Path | None, cfg: dict) -> list[dict]:
             "set CG_KOTLIN_SCIP=1 (or pass --scip index.scip)")
     else:
         add("kotlin", "exact", "scip-java on the Gradle / Maven build")
+    # Java exact mode is the same scip-java run as Kotlin (CG_JAVA_SCIP or CG_KOTLIN_SCIP).
+    jdk_major = None
+    if has("java"):
+        from .plugins.kotlin.exact import java_major
+        jdk_major = java_major(tools["java"]["path"])
+    jdk_note = f"; JDK {jdk_major}" if jdk_major else ""
+    if not ts_ok or not _module("tree_sitter_java"):
+        add("java", "unavailable", "tree-sitter grammar missing", _pip("tree_sitter", "tree_sitter_java"))
+    elif env.get("JAVA_SCIP_FILE") or env.get("KOTLIN_SCIP_FILE"):
+        which = "CG_JAVA_SCIP_FILE" if env.get("JAVA_SCIP_FILE") else "CG_KOTLIN_SCIP_FILE (shared with Kotlin)"
+        add("java", "exact", which + jdk_note)
+    elif not has("scip-java") or not has("java"):
+        miss = " and ".join(t for t in ("scip-java", "java") if not has(t))
+        why = f"tree-sitter-java; {miss} not found{jdk_note}"
+        if jdk_major is not None and jdk_major < 17:
+            why += " (scip-java needs JDK 17, 21 or 25)"
+        add("java", "heuristic", why,
+            "install.sh --with java  (JDK 17+ and scip-java via coursier), then CG_JAVA_SCIP=1")
+    elif jdk_major is not None and jdk_major < 17:
+        extra = "; unsupported Java 8/11 build JDK" if jdk_major in (8, 11) else ""
+        add("java", "heuristic", f"JDK too old (found {jdk_major}; scip-java needs JDK 17, 21 or 25){extra}",
+            "install JDK 17, 21 or 25, then CG_JAVA_SCIP=1")
+    elif env.get("JAVA_SCIP") != "1" and env.get("KOTLIN_SCIP") != "1":
+        add("java", "heuristic", "scip-java found but not run: it runs the Gradle / Maven build (the project's build scripts)"
+            + jdk_note,
+            "set CG_JAVA_SCIP=1 (or CG_KOTLIN_SCIP=1; one run serves both) or pass --scip index.scip")
+    else:
+        add("java", "exact", "scip-java on the Gradle / Maven build" + jdk_note)
     # Swift
     if not ts_ok or not _module("tree_sitter_swift"):
         add("swift", "unavailable", "tree-sitter grammar missing", _pip("tree_sitter", "tree_sitter_swift"))
@@ -441,7 +471,7 @@ def _legacy_lines(r: dict) -> list[str]:
     names = r.get("legacy_env") or []
     if names:
         shown = ", ".join(names)
-        lines.append(f"note: {shown} is set; use the CG_* name (removed in 0.18.0)")
+        lines.append(f"note: {shown} is set; use the CG_* name (removed in 0.19.0)")
     leg = r.get("legacy_cache") or None
     if leg:
         if leg.get("path"):
@@ -469,12 +499,16 @@ def _mcp(rootp: Path | None) -> list:
 
 
 def setup(langs: list[str] | None = None, quiet: bool = False) -> int:
-    """`cg setup [typescript php dart]`: install the extractor dependencies for the languages whose toolchain is
-    present (all three by default); returns 1 if an explicitly named one failed."""
+    """`cg setup [typescript php dart java]`: install the extractor dependencies for the languages whose toolchain is
+    present (typescript, php, dart by default). `java` prints the JDK 17+ hint and reuses the Kotlin scip-java
+    install path; it does not run a build. Returns 1 if an explicitly named one failed."""
     from .plugins.dart.plugin import DartPlugin, find_dart
     want = langs or list(extractors.SPECS)
     rc = 0
     for lang in want:
+        if lang == "java":
+            rc = rc or _setup_java(quiet)
+            continue
         tool = {"typescript": "npm", "php": "composer", "dart": None}[lang]
         need = {"typescript": "node", "php": "php", "dart": None}[lang]
         dart = find_dart() if lang == "dart" else None
@@ -503,3 +537,22 @@ def setup(langs: list[str] | None = None, quiet: bool = False) -> int:
             print(f"{lang}: failed: {e}", file=sys.stderr)
             rc = 1
     return rc
+
+
+def _setup_java(quiet: bool) -> int:
+    """`cg setup java`: the same scip-java / JDK hint as `install.sh --with java`. Heuristic mode needs neither."""
+    from .plugins.kotlin.exact import find_java, scip_java_candidates
+    jdk = find_java()
+    tools = scip_java_candidates()
+    if jdk:
+        print(f"java: JDK found ({jdk}). scip-java needs JDK 17+ (17, 21 or 25).")
+    else:
+        print("java: no JDK found. Install JDK 17+ (17, 21 or 25; https://adoptium.net).", file=sys.stderr)
+    if tools:
+        print(f"java: scip-java found ({tools[0]}). Exact mode stays opt-in, because it runs the build: CG_JAVA_SCIP=1")
+    else:
+        print("java: scip-java not found. install.sh --with java installs it (the same path as --with kotlin).",
+              file=sys.stderr)
+    if not quiet:
+        print("java: heuristic indexing uses tree-sitter-java and does not need a JDK.")
+    return 0

{cg_code_graph-0.17.0 → cg_code_graph-0.18.0}/cg_code_graph/indexer.py

@@ -28,9 +28,11 @@ from .plugins.nextjs.plugin import NextPlugin
 from .plugins.express.plugin import ExpressPlugin
 from .plugins.astro.plugin import AstroPlugin
 from .plugins.kotlin.plugin import KotlinPlugin
+from .plugins.java.plugin import JavaPlugin
 from .plugins.swift.plugin import SwiftPlugin
 
-LANGUAGE_PLUGINS = [PhpPlugin(), TypeScriptPlugin(), PythonPlugin(), DartPlugin(), RustPlugin(), CFamilyPlugin(), KotlinPlugin(), SwiftPlugin(),
+LANGUAGE_PLUGINS = [PhpPlugin(), TypeScriptPlugin(), PythonPlugin(), DartPlugin(), RustPlugin(), CFamilyPlugin(),
+                    KotlinPlugin(), JavaPlugin(), SwiftPlugin(),
                     *SCIP_PLUGINS]
 FRAMEWORK_PLUGINS = [LaravelPlugin(), NuxtPlugin(), DjangoPlugin(), FastAPIPlugin(), FlaskPlugin(), FlutterPlugin(), NestPlugin(), NextPlugin(), ExpressPlugin(), AstroPlugin()]
 
@@ -206,11 +208,18 @@ def index_project(root: str | Path, db_path: str | Path, name: str | None = None
     for fw in frameworks:
         if f"{fw.language}/{fw.name}" not in stats["plugins"]:
             stats["plugins"][f"{fw.language}/{fw.name}"] = {"status": "detected; language plugin not active"}
+    kp = project.options.get("kotlin_plugin")
+    if kp is not None and not getattr(kp, "_exact_done", False):
+        kp._finish_exact(project, project.options.get("kotlin_files") or [], kp._exported_stats)
     if scip:
         from .plugins.scip.importer import import_scip
         for s in scip:
             if s in (project.options.get("scip_consumed") or []):
-                stats["plugins"][f"scip:{s}"] = {"status": "imported by the Kotlin plugin (exact layer)"}
+                who = (project.options.get("scip_consumed_by") or {}).get(s) or ["kotlin"]
+                names = [w.capitalize() for w in who]
+                word = "plugin" if len(names) == 1 else "plugins"
+                stats["plugins"][f"scip:{s}"] = {
+                    "status": f"imported by the {' and '.join(names)} {word} (exact layer)"}
                 continue
             stats["plugins"][f"scip:{s}"] = import_scip(s, builder)
     # generated / copied / vendored files: out of the graph (default) or labelled (attrs.generated)

{cg_code_graph-0.17.0 → cg_code_graph-0.18.0}/cg_code_graph/mcp_server.py

@@ -946,7 +946,7 @@ def coverage(path: str | None = None, all_files: bool = False, json_output: bool
     exact-mode indexer such as rust-analyzer or scip-clang is missing, skipped when the toolchain is missing, with the
     install hint) and file completeness (discovered vs indexed, with parse failures, files over the size limit,
     unmapped files outside the source roots, excluded files); unsupported source types by extension or shebang
-    (.go .java .kt .swift .qml .sh ...); blind spots: route / handler registrations cg does not model, with file:line.
+    (.go .qml .sh ...); blind spots: route / handler registrations cg does not model, with file:line.
     path (optional): a file or directory; says whether cg has it in the graph. all_files: list every file per bucket
     (default: the first 5). json_output: the completeness object as JSON. Not covered, heuristic or a blind spot
     means: use your normal search and file reading for that part."""