cg-code-graph 0.21.1__tar.gz → 0.22.0__tar.gz

{cg_code_graph-0.21.1/cg_code_graph.egg-info → cg_code_graph-0.22.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cg-code-graph
-Version: 0.21.1
+Version: 0.22.0
 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
@@ -33,6 +33,8 @@ Dynamic: license-file
 
 For you and your AI agent (CLI + MCP). No LLM in the graph.
 
+<!-- mcp-name: io.github.cyberchronos00/cg-code-graph -->
+
 [![PyPI](https://img.shields.io/pypi/v/cg-code-graph)](https://pypi.org/project/cg-code-graph/) [![Downloads](https://img.shields.io/pepy/dt/cg-code-graph)](https://pepy.tech/project/cg-code-graph) [![Python 3.11+](https://img.shields.io/badge/python-3.11%2B-3776AB)](https://pypi.org/project/cg-code-graph/) [![License: MIT](https://img.shields.io/badge/license-MIT-blue)](https://github.com/cyberchronos00/code-graph/blob/main/LICENSE) [![Docs](https://img.shields.io/badge/docs-code--graph-93b600)](https://code-graph.cyberchronos00.workers.dev/)
 
 [Docs](https://code-graph.cyberchronos00.workers.dev/) · [Quick start](https://code-graph.cyberchronos00.workers.dev/docs/quickstart) · [CLI](https://code-graph.cyberchronos00.workers.dev/docs/cli) · [MCP](https://code-graph.cyberchronos00.workers.dev/docs/mcp) · [Changelog](https://github.com/cyberchronos00/code-graph/blob/main/CHANGELOG.md)
@@ -91,7 +93,7 @@ Go and others: import a SCIP index with `cg index --scip FILE`. Run `cg doctor`
 - Across repos: `cg link` matches client HTTP calls (frontends and Laravel / Guzzle backends) to server routes, and checks request keys (Laravel FormRequest, Elysia schemas, Nuxt / TypeScript bodies).
 - Route guards: `cg routes --writes --unguarded` lists routes that write with no auth guard. Laravel checks inside the action count as inline guards unless `--strict`.
 - Security views: `cg surface` lists the inbound / outbound attack surface and findings (hardcoded credentials, plaintext protocols, unverified webhooks, unguarded handlers, listeners on all interfaces, disabled TLS / SSH verification, plaintext gRPC, exposed IPC, outbound URLs built from request input) as text, JSON or SARIF, with `--fail-on` for CI.
-- MCP server (`cg-mcp`) so agents query the graph instead of grepping.
+- MCP server (`cg-mcp`) so agents query the graph instead of grepping. One server can serve many graphs (`graphs` plus an optional `graph` argument); a single `--db` stays one graph. See [MCP](https://code-graph.cyberchronos00.workers.dev/docs/mcp).
 - Visual view (`cg serve`) and plan checks (`cg plan check`) for planned changes.
 
 ## Docs

{cg_code_graph-0.21.1 → cg_code_graph-0.22.0}/README.md

@@ -6,6 +6,8 @@
 
 For you and your AI agent (CLI + MCP). No LLM in the graph.
 
+<!-- mcp-name: io.github.cyberchronos00/cg-code-graph -->
+
 [![PyPI](https://img.shields.io/pypi/v/cg-code-graph)](https://pypi.org/project/cg-code-graph/) [![Downloads](https://img.shields.io/pepy/dt/cg-code-graph)](https://pepy.tech/project/cg-code-graph) [![Python 3.11+](https://img.shields.io/badge/python-3.11%2B-3776AB)](https://pypi.org/project/cg-code-graph/) [![License: MIT](https://img.shields.io/badge/license-MIT-blue)](https://github.com/cyberchronos00/code-graph/blob/main/LICENSE) [![Docs](https://img.shields.io/badge/docs-code--graph-93b600)](https://code-graph.cyberchronos00.workers.dev/)
 
 [Docs](https://code-graph.cyberchronos00.workers.dev/) · [Quick start](https://code-graph.cyberchronos00.workers.dev/docs/quickstart) · [CLI](https://code-graph.cyberchronos00.workers.dev/docs/cli) · [MCP](https://code-graph.cyberchronos00.workers.dev/docs/mcp) · [Changelog](https://github.com/cyberchronos00/code-graph/blob/main/CHANGELOG.md)
@@ -64,7 +66,7 @@ Go and others: import a SCIP index with `cg index --scip FILE`. Run `cg doctor`
 - Across repos: `cg link` matches client HTTP calls (frontends and Laravel / Guzzle backends) to server routes, and checks request keys (Laravel FormRequest, Elysia schemas, Nuxt / TypeScript bodies).
 - Route guards: `cg routes --writes --unguarded` lists routes that write with no auth guard. Laravel checks inside the action count as inline guards unless `--strict`.
 - Security views: `cg surface` lists the inbound / outbound attack surface and findings (hardcoded credentials, plaintext protocols, unverified webhooks, unguarded handlers, listeners on all interfaces, disabled TLS / SSH verification, plaintext gRPC, exposed IPC, outbound URLs built from request input) as text, JSON or SARIF, with `--fail-on` for CI.
-- MCP server (`cg-mcp`) so agents query the graph instead of grepping.
+- MCP server (`cg-mcp`) so agents query the graph instead of grepping. One server can serve many graphs (`graphs` plus an optional `graph` argument); a single `--db` stays one graph. See [MCP](https://code-graph.cyberchronos00.workers.dev/docs/mcp).
 - Visual view (`cg serve`) and plan checks (`cg plan check`) for planned changes.
 
 ## Docs

{cg_code_graph-0.21.1 → cg_code_graph-0.22.0}/cg_code_graph/__init__.py

@@ -1,2 +1,2 @@
 """code-graph: deterministic code knowledge graph (beta)."""
-__version__ = "0.21.1"
+__version__ = "0.22.0"

{cg_code_graph-0.21.1 → cg_code_graph-0.22.0}/cg_code_graph/cli.py

@@ -73,6 +73,27 @@ def _add_install_parser(sub, uninstall: bool):
     p.add_argument("--uninstall", action="store_true", help="remove the entry instead of writing it")
     p.add_argument("--dry-run", action="store_true", help="print the exact changes and write nothing")
     p.add_argument("--yes", action="store_true", help="skip the confirmation prompt")
+    if not uninstall:
+        p.add_argument("--registry", action="store_true",
+                       help="global single entry: cg-mcp with no --db; graphs come from the registry")
+        p.add_argument("--migrate", action="store_true",
+                       help="fold per-repo cg-* host entries into the graph registry (dry run unless --apply)")
+        p.add_argument("--apply", action="store_true", help="with --migrate, write the registry and the host config")
+        p.add_argument("--graphs", help="registry file for --migrate (default: ~/.config/cg/graphs.toml)")
+
+
+def _impact_labels(rows: list[dict], name_of, group_of) -> list[str]:
+    """Display names for one impact listing. Same-name rows in one group gain a file suffix."""
+    from collections import defaultdict
+    groups = defaultdict(list)
+    for i, row in enumerate(rows):
+        groups[group_of(row)].append(i)
+    labels = [name_of(row) for row in rows]
+    for idxs in groups.values():
+        named = Q.disambiguated([labels[i] for i in idxs], [rows[i].get("file") for i in idxs])
+        for i, lab in zip(idxs, named):
+            labels[i] = lab
+    return labels
 
 
 def main(argv=None):
@@ -321,8 +342,8 @@ def main(argv=None):
     p.add_argument("--quiet", action="store_true", help="pass --quiet to each refresh")
     helps = {"reaches": "everything that depends on the targets, grouped by entry classification",
              "siblings": "code related to a symbol: class hierarchy, the same method in sibling classes, shared resources, co-callers",
-             "writers": "code that writes a table (or column), or a stored property `Type.prop`",
-             "readers": "code that reads a stored property `Type.prop` (READS_PROP)",
+             "writers": "code that writes a table, a column (`table.column`) or a stored property `Type.prop`",
+             "readers": "code that reads a table, a column (`table.column`) or a stored property `Type.prop`",
              "impact": "callers of a method up to their entry points (reverse walk), overrides listed apart",
              "stats": "node / edge counts of a graph DB",
              "node": "one node's details and its incoming / outgoing edges",
@@ -342,6 +363,9 @@ def main(argv=None):
         p.add_argument("--no-paths", action="store_true")
         p.add_argument("--max-depth", type=int, default=30)
         p.add_argument("--gate", default="auto", help="gate scenario for live/gated split (default: the one indexed; 'none' to disable)")
+        if name in ("reaches", "writers", "readers"):
+            p.add_argument("--connection", help="only the table / column edges that go to this database connection "
+                                                "(a dynamic name by its pattern, e.g. 'legacy_{store.id}'; docs/cli.md)")
         if name in ("reaches", "impact", "downstream"):
             p.add_argument("--platform", help="only code built for this target (windows, linux, macos, ios, android, web; see docs/platforms.md)")
         if name == "impact":
@@ -489,9 +513,14 @@ def main(argv=None):
                       mcp_file=a.mcp_file, dry_run=a.dry_run, assume_yes=a.yes)
     if a.cmd in ("install", "uninstall"):
         from .mcp_install import run as run_install
+        registry = bool(getattr(a, "registry", False))
+        migrate = bool(getattr(a, "migrate", False))
+        if (registry or migrate) and not a.project:
+            a.glob = True
         return run_install(a.host, scope="global" if a.glob else "project", root=a.dir, db=a.db,
                            tools=a.tools, portable=a.portable, remove=a.cmd == "uninstall" or a.uninstall,
-                           dry_run=a.dry_run, assume_yes=a.yes)
+                           dry_run=a.dry_run, assume_yes=a.yes, registry=registry, migrate=migrate,
+                           apply=bool(getattr(a, "apply", False)), graphs=getattr(a, "graphs", None))
     if a.cmd == "hooks":
         from . import hooks as HK
         return HK.run(a.action, root=a.dir, db=a.db, dry_run=a.dry_run, assume_yes=a.yes)
@@ -646,7 +675,7 @@ def main(argv=None):
         return
     if a.cmd == "reaches":
         res = Q.reaches(st, a.specs, min_conf=a.min_confidence, max_depth=a.max_depth, gate=None if a.gate == "none" else a.gate,
-                        platform=a.platform)
+                        platform=a.platform, connection=a.connection)
         res["completeness"] = _completeness(st, [x for t in res["targets"].values() for x in t] + [i["id"] for i in res["items"]])
         print(json.dumps(res, indent=1, default=str) if a.json else Q.render_reaches(res, show_paths=not a.no_paths))
         _note(res["completeness"], a.json)
@@ -658,26 +687,35 @@ def main(argv=None):
         if res.get("platform"):
             from .platforms import render_filter
             print(render_filter(res["platform"]))
-            if res["platform"].get("targets_not_built"):
-                print(f"not built for {a.platform}: {', '.join(res['platform']['targets_not_built'][:6])}")
+            dropped = res["platform"].get("targets_not_built") or []
+            if dropped and res["targets"]:
+                print(f"not built for {a.platform}: {', '.join(dropped[:6])}")
+            if dropped and not res["targets"]:
+                msg = f"{a.spec} is not built for {a.platform}: nothing calls it there"
+                if Q.divergence_lists(st, a.platform, dropped):
+                    msg += (f" (`cg platforms divergence --target {a.platform}` lists references to it that would not build)")
+                print(msg); return
         print(f"targets: {res['targets'][:5]}")
         if not res["targets"]:
             print(f"no method matches {a.spec!r}; try `search` with part of the name"); return
-        if a.platform and res["targets"] and set(res["platform"]["targets_not_built"]) == set(res["targets"]):
-            print(f"{a.spec} is not built for {a.platform}: nothing calls it there (`cg platforms divergence --target "
-                  f"{a.platform}` lists references to it that would not build)"); return
+        for d in res.get("default_arguments") or []:
+            print(f"via default argument: {d['src']} @ {d['file']}:{d['line']}")
         for line in Q.override_lines(res):
             print(line)
+        if res.get("connections"):
+            print(Q.connections_line(res["connections"]))
         if not res["callers"] and not res["entry_points"]:
             print(Q.explain_no_callers(st, a.spec, res["targets"], a.min_confidence)); return
         print(f"callers (transitive): {len(res['callers'])}")
-        for c in res["callers"]:
+        caller_names = _impact_labels(res["callers"], lambda c: c.get("fqn") or c["id"], lambda c: c.get("module") or c.get("kind"))
+        for c, nm in zip(res["callers"], caller_names):
             loc = f"  @ {c['file']}:{c['line']}" if Q.NATIVE_FILE_RE.search(c.get("file") or "") else ""
-            print(f"  d={c['depth']} [{c['module'] or c['kind']}] {c['fqn']}{loc}{Q.caller_label(c)}")
+            print(f"  d={c['depth']} [{c['module'] or c['kind']}] {nm}{loc}{Q.caller_label(c)}")
         print(f"entry points: {len(res['entry_points'])}")
-        for e in res["entry_points"]:
+        entry_names = _impact_labels(res["entry_points"], lambda e: e["name"], lambda e: e.get("entry_kind"))
+        for e, nm0 in zip(res["entry_points"], entry_names):
             native = Q.NATIVE_FILE_RE.search(e.get("file") or "")
-            nm = f"{e.get('fqn') or e['name']}  @ {e['file']}:{e['line']}" if native else e["name"]
+            nm = f"{e.get('fqn') or e['name']}  @ {e['file']}:{e['line']}" if native else nm0
             print(f"  {e['entry_kind']:16} {nm}  conf={e['path_confidence']}{Q.CANDIDATE_LABEL if e.get('candidate') else ''}"
                   f"{Q.generated_label(e)}{Q.platform_label(e)}")
             if not a.no_paths:
@@ -702,17 +740,14 @@ def main(argv=None):
         res = RT.roundtrip(st, a.spec, include_tests=a.tests)
         print(RT.to_json(res) if a.json else RT.render(res))
     elif a.cmd in ("writers", "readers"):
-        rows = Q.writers(st, a.spec) if a.cmd == "writers" else Q.readers(st, a.spec)
+        groups = Q.access(st, a.spec, a.cmd, a.connection)
+        rows = [r for g in groups for r in g["rows"]]
         if a.json:
             print(json.dumps(rows, indent=1)); return
         if not rows:
-            print(Q.explain_no_writers(st, a.spec, a.cmd)); return
-        for r in rows:
-            ek = ",".join(sorted(r["entry_kinds"]))
-            ex = {k: v for k, v in (r.get("attrs") or {}).items() if k in ("receiver", "accessor", "storage")}
-            extra = ("  " + " ".join(f"{k}={v}" for k, v in ex.items())) if ex else ""
-            kind = f"{r['kind']}({r['attrs']['orig']})" if r.get("test") else r['kind']
-            print(f"[{r['module']}] {r['fqn']}  {kind} {r['dst']}  @{r['file']}:{r['line']} ({r['confidence']}){extra}  entries: {ek}")
+            print(Q.explain_no_connection(st, a.spec, a.connection, a.cmd) if a.connection else Q.explain_no_writers(st, a.spec, a.cmd))
+            return
+        print("\n".join(Q.render_access(groups)))
         verb = "write" if a.cmd == "writers" else "read"
         nt = sum(1 for r in rows if r.get("test"))
         print(f"{len(rows)} {verb} edges, {len({r['src'] for r in rows})} {a.cmd}" + (f" ({nt} from test code)" if nt else ""))
@@ -773,6 +808,14 @@ def node_doc(st, nid: str) -> dict:
     return {"node": n, "out": edges("src", "dst"), "in": edges("dst", "src")}
 
 
+def _mass_note(e: dict) -> str:
+    """`  via update(validated()) keys from UpdateBookRequest::rules ∩ Book::$fillable` for a mass-assignment column write (#177)."""
+    at = e.get("attrs") or {}
+    if e["kind"] != "WRITES_COLUMN" or not at.get("mass_assignment"):
+        return ""
+    return f"  via {at.get('via')}" + (f" keys from {' ∩ '.join(at['keys_from'])}" if at.get("keys_from") else "")
+
+
 def render_node(d: dict) -> str:
     """`cg node` text: the node's location, kind and attrs, then its edges; the same edge from one site (one row per
     gate scenario or accessor) is listed once with a count."""
@@ -801,10 +844,10 @@ def render_node(d: dict) -> str:
         out.append(f"  {'outgoing' if key == 'out' else 'incoming'} ({len(es)}):" if es else f"  {'outgoing' if key == 'out' else 'incoming'}: none")
         seen: dict = {}
         for e in es:
-            k = (e["kind"], e[other], e["file"], e["line"], e["confidence"])
+            k = (e["kind"], e[other], e["file"], e["line"], e["confidence"], _mass_note(e))
             seen[k] = seen.get(k, 0) + 1
-        for (kind, o, f, line, conf), c in seen.items():
-            out.append(f"    {arrow} {kind} {o}  @{f}:{line} {conf}" + (f"  ×{c}" if c > 1 else ""))
+        for (kind, o, f, line, conf, note), c in seen.items():
+            out.append(f"    {arrow} {kind} {o}  @{f}:{line} {conf}{note}" + (f"  ×{c}" if c > 1 else ""))
     return "\n".join(out)
 
 

{cg_code_graph-0.21.1 → cg_code_graph-0.22.0}/cg_code_graph/coverage.py

@@ -233,9 +233,20 @@ def file_completeness(discovered: list[str], report: dict) -> dict:
     for b in BUCKETS:
         res[b] = len(out[b])
     res["paths"] = {b: v[:MAX_PATHS] for b, v in out.items() if v}
+    if out["excluded"]:
+        res["exclude_reasons"] = {f: _exclude_reason(f) for f in out["excluded"]}
     return res
 
 
+def _exclude_reason(path: str) -> str:
+    """Why a discovered file was left out of the graph: tool config, a test file, or a `.cg.yaml` exclude."""
+    if _TOOL_FILE.search(path):
+        return "tool config"
+    if _TEST_NAME.search(path):
+        return "test file"
+    return ".cg.yaml exclude"
+
+
 def missing_files(e: dict) -> int:
     """Discovered files of a language entry that are not in the graph for a reason other than a deliberate exclusion."""
     return sum(e.get(b, 0) for b in ("parse_failed", "skipped_oversize", "unmapped"))
@@ -308,6 +319,8 @@ def compute(root: str | Path, plugins: dict, scip_imported: bool = False, report
         status, reason = _status(lang, st)
         e = {"language": lang, "files": n, "status": status,
              "by_ext": {x: files[x] for x in exts if files[x]}}
+        if lang == "typescript" and not any(x in e["by_ext"] for x in (".ts", ".tsx", ".mts", ".cts")):
+            e["label"] = "javascript"
         if reason:
             e["reason"] = reason
         if status in ("skipped", "heuristic", "not_indexed"):
@@ -495,16 +508,22 @@ def blind_spots(cov: dict | None) -> list[dict]:
     return list((cov or {}).get("blind_spots") or [])
 
 
+def _shown_lang(e: dict) -> str:
+    """Display name. A TypeScript-plugin entry with no .ts/.tsx/.mts/.cts files reads as javascript; JSON keeps `language`."""
+    return e.get("label") or e["language"]
+
+
 def _entry_text(e: dict) -> str:
     """'php 18 exact' / 'python 4 discovered, 2 indexed (exact parser): 1 parse failed, 1 unmapped' /
-    'swift 101 heuristic, 8 parsed with syntax errors'."""
+    'javascript 12 discovered, 2 indexed (exact parser): 9 unmapped, 1 excluded'."""
     pe = e.get("parsed_with_errors")
     errs = f", {pe} parsed with syntax errors" if pe else ""
+    name = _shown_lang(e)
     if e.get("files_complete") is False:
-        parts = [f"{e[b]} {BUCKET_SHORT[b]}" for b in ("parse_failed", "skipped_oversize", "unmapped") if e.get(b)]
-        return (f"{e['language']} {e['files']} discovered, {e['indexed']} indexed ({e['status'].replace('_', ' ')} parser)"
+        parts = [f"{e[b]} {BUCKET_SHORT[b]}" for b in BUCKETS if e.get(b)]
+        return (f"{name} {e['files']} discovered, {e['indexed']} indexed ({e['status'].replace('_', ' ')} parser)"
                 + (": " + ", ".join(parts) if parts else "") + errs)
-    return f"{e['language']} {e['files']} {e['status'].replace('_', ' ')}{errs}"
+    return f"{name} {e['files']} {e['status'].replace('_', ' ')}{errs}"
 
 
 def realtime_unresolved_lines(e: dict, indent: str = "  ") -> list[str]:
@@ -579,11 +598,16 @@ def _paths_lines(e: dict, all_files: bool, indent: str = "    ") -> list[str]:
     out = []
     for b in BUCKETS:
         ps = (e.get("paths") or {}).get(b) or []
-        if not ps or (b == "excluded" and not all_files):
+        if not ps:
             continue
         show = ps if all_files else ps[:SHOW_PATHS]
         more = e.get(b, len(ps)) - len(show)
-        out.append(f"{indent}{BUCKET_SHORT[b]}: " + ", ".join(show) + (f" … +{more} more (--all-files)" if more > 0 else ""))
+        if b == "excluded":
+            reasons = e.get("exclude_reasons") or {}
+            body = ", ".join(f"{p} ({reasons.get(p) or _exclude_reason(p)})" for p in show)
+        else:
+            body = ", ".join(show)
+        out.append(f"{indent}{BUCKET_SHORT[b]}: " + body + (f" … +{more} more (--all-files)" if more > 0 else ""))
     return out
 
 
@@ -720,11 +744,12 @@ def render(covs: dict[str, dict | None], all_files: bool = False) -> str:
         for e in gaps(cov):
             any_gap = True
             exts = ", ".join(f"{k} {v}" for k, v in sorted(e["by_ext"].items()))
+            shown = _shown_lang(e)
             if e.get("files_complete") is False:
                 parts = [f"{e[b]} {BUCKET_SHORT[b]}" for b in BUCKETS if e.get(b)]
-                line = f"  {e['language']}: {e['files']} files ({exts}) {e['indexed']} indexed, " + ", ".join(parts)
+                line = f"  {shown}: {e['files']} files ({exts}) {e['indexed']} indexed, " + ", ".join(parts)
             else:
-                line = f"  {e['language']}: {e['files']} files ({exts}) {e['status'].replace('_', ' ')}"
+                line = f"  {shown}: {e['files']} files ({exts}) {e['status'].replace('_', ' ')}"
             if e.get("reason"):
                 line += f": {e['reason']}"
             out.append(line)
@@ -795,7 +820,7 @@ def note(covs: dict[str, dict | None]) -> str:
         for e in gaps(c):
             what = (f"{e['indexed']} of {e['files']} files indexed" if e.get("files_complete") is False
                     else f"{e['files']} files, {e['status'].replace('_', ' ')}")
-            bad.append(f"{e['language']} ({what}{', ' + r if len(covs) > 1 else ''})")
+            bad.append(f"{_shown_lang(e)} ({what}{', ' + r if len(covs) > 1 else ''})")
     bs = [b for c in covs.values() for b in blind_spots(c)]
     bs_txt = f"; blind spots: {_bs_count(bs)} (see coverage)" if bs else ""
     nerr = sum(e.get("syntax_error_files", 0) for c in covs.values() for e in (c or {}).get("languages", []))
@@ -956,6 +981,18 @@ def completeness_for(store, node_ids=None, categories=("route", "handler"), whol
                         categories=categories, ids=set(node_ids), files=sc["files"], **kw)
 
 
+def _unsupported_text(uns: dict) -> str:
+    """`unsupported: sh 3`. A single-repo DB drops a `shop/sh` prefix; a combined graph keeps `api/sh`."""
+    keys = list(uns)
+    prefixes = {k.partition("/")[0] for k in keys if "/" in k}
+    drop = len(prefixes) <= 1
+    parts = []
+    for k, v in sorted(uns.items()):
+        label = k.partition("/")[2] if drop and "/" in k else k
+        parts.append(f"{label} {v}")
+    return "unsupported: " + ", ".join(parts)
+
+
 def possibly_more(comp: dict) -> str:
     """'1 unmodelled route registration, 2 Python files not indexed' for an incomplete answer, '' when complete."""
     if comp.get("complete"):
@@ -990,7 +1027,7 @@ def possibly_more(comp: dict) -> str:
         parts.append(f"{len(se)} file{'s' if len(se) != 1 else ''} with syntax errors ({x['file']}{pos}"
                      + (f" +{len(se) - 1}" if len(se) > 1 else "") + ")")
     if comp.get("unsupported"):
-        parts.append("unsupported: " + ", ".join(f"{k} {v}" for k, v in sorted(comp["unsupported"].items())))
+        parts.append(_unsupported_text(comp["unsupported"]))
     if comp.get("recorded") is False:
         parts.append("coverage not recorded")
     return ", ".join(parts)

{cg_code_graph-0.21.1 → cg_code_graph-0.22.0}/cg_code_graph/external.py

@@ -292,8 +292,35 @@ def read_spring_datasources(root: Path) -> list:
         out.append(fact)
     return out
 
+
+CONNECTION_EDGE_KINDS = ("READS_TABLE", "WRITES_TABLE", "READS_COLUMN", "WRITES_COLUMN", "MENTIONS_COLUMN")
+
+
+def table_connections(builder) -> int:
+    """`attrs.connections` of a table node: `{connection: edge count}` over its table / column edges that carry
+    `attrs.connection` (most used first). One table name used on several connections stays one node. Not tied to an ORM:
+    a plugin only has to set `attrs.connection` on its table / column edges."""
+    counts: dict[str, dict[str, int]] = {}
+    for e in builder.edges.values():
+        conn = (e.attrs or {}).get("connection")
+        if e.kind not in CONNECTION_EDGE_KINDS or not conn:
+            continue
+        key = e.dst.split(":", 1)[1]
+        tid = f"table:{key}" if e.dst.startswith("table:") else f"table:{key.rsplit('.', 1)[0]}"
+        counts.setdefault(tid, {})
+        counts[tid][conn] = counts[tid].get(conn, 0) + 1
+    n = 0
+    for tid, c in counts.items():
+        node = builder.nodes.get(tid)
+        if node is not None:
+            node.attrs["connections"] = dict(sorted(c.items(), key=lambda kv: (-kv[1], kv[0])))
+            n += 1
+    return n
+
+
 def attach(builder, root: Path) -> dict:
     root = Path(root)
+    table_connections(builder)
     envs = {n.id[len("env:"):]: n for n in builder.nodes.values() if n.kind == "env"}
     conns = [n for n in builder.nodes.values() if n.kind == "connection"]
     # Spring datasources before the empty-check so a repo with only application*.properties still indexes (#41)
@@ -646,6 +673,16 @@ def attach(builder, root: Path) -> dict:
                          (builder.nodes[nid].attrs or {}).get("confidence") or "resolved",
                          op="table", via="laravel connection")
         st["tables"] = st.get("tables", 0) + 1
+    # a table used on several connections: one CONNECTS_TO per connection that maps to a system, not only the first
+    for n in builder.nodes.values():
+        if n.kind != "table":
+            continue
+        for cname in (n.attrs or {}).get("connections") or {}:
+            nid = conn_ext.get(f"connection:{cname}")
+            if nid:
+                builder.add_edge(n.id, nid, "CONNECTS_TO", n.file, n.line,
+                                 (builder.nodes[nid].attrs or {}).get("confidence") or "resolved",
+                                 op="table", via=f"connection {cname}")
 
     # ---- ORM tables without a datasource block: attach to the sole SQL system of the project (#41 step 2)
     # Immich (Kysely + DB_URL), Nest TypeORM (@Entity + DATABASE_URL) when no DataSource fact named them.

{cg_code_graph-0.21.1 → cg_code_graph-0.22.0}/cg_code_graph/indexer.py

@@ -28,6 +28,19 @@ from .plugins.flutter.plugin import FlutterPlugin
 from .plugins.nest.plugin import NestPlugin
 from .plugins.nextjs.plugin import NextPlugin
 from .plugins.express.plugin import ExpressPlugin
+import subprocess
+
+
+def _git_head(root) -> str | None:
+    """Commit indexed for this root, when it is a git work tree."""
+    try:
+        proc = subprocess.run(["git", "-C", str(root), "rev-parse", "HEAD"],
+                              capture_output=True, text=True, timeout=8)
+    except (OSError, subprocess.TimeoutExpired):
+        return None
+    if proc.returncode != 0:
+        return None
+    return (proc.stdout or "").strip() or None
 from .plugins.astro.plugin import AstroPlugin
 from .plugins.kotlin.plugin import KotlinPlugin
 from .plugins.java.plugin import JavaPlugin
@@ -464,6 +477,9 @@ def index_project(root: str | Path, db_path: str | Path, name: str | None = None
     stats["nodes"] = len(builder.nodes)
     stats["edges"] = len(builder.edges)
     stats["index_seconds"] = round(time.time() - t0, 2)
-    store.set_meta(project=project.name, root=str(project.root), stats=stats, indexed_at=time.strftime("%Y-%m-%d %H:%M:%S UTC", time.gmtime()))
+    head = _git_head(project.root)
+    extra = {"commit": head} if head else {}
+    store.set_meta(project=project.name, root=str(project.root), stats=stats,
+                   indexed_at=time.strftime("%Y-%m-%d %H:%M:%S UTC", time.gmtime()), **extra)
     store.db.close()
     return stats

{cg_code_graph-0.21.1 → cg_code_graph-0.22.0}/cg_code_graph/link.py

@@ -497,10 +497,23 @@ def link_many(repos: list[tuple[str, str, str]], out_db: str, *, allow: dict | N
     for name, path, _role in repos:
         got = for_graph(GraphStore(path))
         cov.update({name: next(iter(got.values()))} if len(got) == 1 else got)
+    commits = {}
+    for name, path, _role in repos:
+        try:
+            member_st = GraphStore(path)
+            try:
+                member = member_st.meta()
+            finally:
+                member_st.db.close()
+        except Exception:
+            member = {}
+        if member.get("commit"):
+            commits[name] = member["commit"]
     store.set_meta(project="+".join(names), stats=stats, repos=names, coverage=cov,
                    sources={name: str(Path(path).resolve()) for name, path, _role in repos},
                    repo_roles={name: role for name, _path, role in repos}, link_clients=clients, link_allow=allow,
-                   indexed_at=time.strftime("%Y-%m-%d %H:%M:%S UTC", time.gmtime()))
+                   indexed_at=time.strftime("%Y-%m-%d %H:%M:%S UTC", time.gmtime()),
+                   **({"commit": commits} if commits else {}))
     db.close()
     return {"stats": stats, "results": results, "payload_issues": issues, "uncalled_routes": uncalled,
             "webhook_subscriptions": subscriptions, "external_callbacks": callbacks}