code-context-control 2.43.0__tar.gz → 2.44.0__tar.gz

{code_context_control-2.43.0 → code_context_control-2.44.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: code-context-control
-Version: 2.43.0
+Version: 2.44.0
 Summary: Local code-intelligence layer for AI coding tools (Claude Code, Codex, Gemini, Copilot). Retrieve less, read less, edit safer.
 Author-email: Dimitri Tselenchuk <dtselenc@gmail.com>
 License-Expression: Apache-2.0
@@ -197,7 +197,7 @@ c3 ui                             # opens http://127.0.0.1:3333
   <img src="https://raw.githubusercontent.com/drknowhow/code-context-control/main/docs/screenshots/ui_dashboard.png" alt="C3 per-project dashboard" width="900">
 </p>
 
-Illustrative example from one project's dashboard (numbers vary by project): **448K tokens saved** (89.9% rate) — C3's estimate versus a full-file-read baseline — plus 208 files indexed, 20 sessions, codebase breakdown by language, current-session live counters (in/out tokens, cache reads, services online), and a stream of recent tool calls and file changes.
+Illustrative example from one project's dashboard (numbers vary by project): **448K tokens saved** (89.9% rate) — C3's estimate versus a full-file-read baseline — plus 208 files indexed, 20 sessions, codebase breakdown by language, current-session live counters (in/out tokens, cache reads, services online), and a stream of recent tool calls and file changes. Run `c3 bench session` on your own project to generate your own scorecard.
 
 ### 3. Edit Ledger — every AI-driven edit tracked
 
@@ -395,13 +395,20 @@ Applying or switching a tier **preserves your own `allow`/`deny` rules** (and ke
 
 ## Benchmarks
 
+Don't take our word for it — every number C3 advertises is reproducible on your own machine, against your own project:
+
 ```bash
-c3 benchmark /path/to/project
-c3 bench aider                    # Aider Polyglot suite
-c3 bench swe                      # SWE-bench Lite
+c3 bench session                  # six realistic workflow scenarios, A/B with vs without C3
+c3 benchmark /path/to/project     # per-operation micro-benchmarks (compression, retrieval, filtering, validation)
+c3 bench aider                    # Aider Polyglot suite (external; burns real API tokens)
+c3 bench swe                      # SWE-bench Lite (external)
 ```
 
-Real-world A/B tests: same task, with and without C3 mounted. Reports include token deltas, cost deltas, win rates, tool-usage analysis, and per-task breakdowns. See the **Benchmark Dashboard** under Settings → Background Agents in the Hub.
+The session benchmark's baseline models a *competent* agent working without C3 — one targeted search, each file read once — not a strawman that re-reads everything, and it scores answer quality alongside tokens. For reference, a run against C3's own repository (v2.43.0, 2026-07-02) measured **51.8% token savings (2.07×)** across the six scenarios at quality parity. Your numbers will differ with your project's shape — that's why the harness ships with the tool.
+
+Beyond synthetic scenarios, C3 records real per-tool usage to `.c3/tool_telemetry.jsonl`, so estimated savings can always be checked against what actually happened in your sessions.
+
+Reports include token deltas, cost deltas, win rates, tool-usage analysis, and per-task breakdowns. See the **Benchmark Dashboard** under Settings → Background Agents in the Hub.
 
 ---
 

{code_context_control-2.43.0 → code_context_control-2.44.0}/README.md

@@ -135,7 +135,7 @@ c3 ui                             # opens http://127.0.0.1:3333
   <img src="https://raw.githubusercontent.com/drknowhow/code-context-control/main/docs/screenshots/ui_dashboard.png" alt="C3 per-project dashboard" width="900">
 </p>
 
-Illustrative example from one project's dashboard (numbers vary by project): **448K tokens saved** (89.9% rate) — C3's estimate versus a full-file-read baseline — plus 208 files indexed, 20 sessions, codebase breakdown by language, current-session live counters (in/out tokens, cache reads, services online), and a stream of recent tool calls and file changes.
+Illustrative example from one project's dashboard (numbers vary by project): **448K tokens saved** (89.9% rate) — C3's estimate versus a full-file-read baseline — plus 208 files indexed, 20 sessions, codebase breakdown by language, current-session live counters (in/out tokens, cache reads, services online), and a stream of recent tool calls and file changes. Run `c3 bench session` on your own project to generate your own scorecard.
 
 ### 3. Edit Ledger — every AI-driven edit tracked
 
@@ -333,13 +333,20 @@ Applying or switching a tier **preserves your own `allow`/`deny` rules** (and ke
 
 ## Benchmarks
 
+Don't take our word for it — every number C3 advertises is reproducible on your own machine, against your own project:
+
 ```bash
-c3 benchmark /path/to/project
-c3 bench aider                    # Aider Polyglot suite
-c3 bench swe                      # SWE-bench Lite
+c3 bench session                  # six realistic workflow scenarios, A/B with vs without C3
+c3 benchmark /path/to/project     # per-operation micro-benchmarks (compression, retrieval, filtering, validation)
+c3 bench aider                    # Aider Polyglot suite (external; burns real API tokens)
+c3 bench swe                      # SWE-bench Lite (external)
 ```
 
-Real-world A/B tests: same task, with and without C3 mounted. Reports include token deltas, cost deltas, win rates, tool-usage analysis, and per-task breakdowns. See the **Benchmark Dashboard** under Settings → Background Agents in the Hub.
+The session benchmark's baseline models a *competent* agent working without C3 — one targeted search, each file read once — not a strawman that re-reads everything, and it scores answer quality alongside tokens. For reference, a run against C3's own repository (v2.43.0, 2026-07-02) measured **51.8% token savings (2.07×)** across the six scenarios at quality parity. Your numbers will differ with your project's shape — that's why the harness ships with the tool.
+
+Beyond synthetic scenarios, C3 records real per-tool usage to `.c3/tool_telemetry.jsonl`, so estimated savings can always be checked against what actually happened in your sessions.
+
+Reports include token deltas, cost deltas, win rates, tool-usage analysis, and per-task breakdowns. See the **Benchmark Dashboard** under Settings → Background Agents in the Hub.
 
 ---
 

{code_context_control-2.43.0 → code_context_control-2.44.0}/cli/c3.py

@@ -85,7 +85,7 @@ console = Console() if HAS_RICH else None
 # Config
 CONFIG_DIR = ".c3"
 CONFIG_FILE = ".c3/config.json"
-__version__ = "2.43.0"
+__version__ = "2.44.0"
 
 
 def _command_deps() -> CommandDeps:
@@ -1051,6 +1051,10 @@ def cmd_init(args):
     # ── Non-interactive (--clear) ──────────────────────────────
     if getattr(args, "clear", False):
         print("\n[--clear] Wiping C3 files...")
+        parent_link = (load_config(project_path) or {}).get("parent") or {}
+        if parent_link.get("path"):
+            print(f"  [!] This project is a sub-project of {parent_link['path']}")
+            print("      The parent still lists it -- run 'c3 sub check --fix' there.")
         _uninstall_mcp_all(project_path)
         if c3_dir.exists():
             shutil.rmtree(c3_dir)
@@ -5733,6 +5737,137 @@ def cmd_projects(args):
             print(f"  Port {s['port']:>5}  {s.get('project_name', '?'):<25}  {s.get('project_path', '')}")
 
 
+def cmd_sub(args):
+    """Manage sub-projects: designated sub-folders with linked .c3 branches."""
+    parent = str(Path(getattr(args, "parent", ".") or ".").resolve())
+    if not (Path(parent) / CONFIG_DIR).is_dir():
+        print(f"No .c3 found in {parent}. Run 'c3 init' there first.")
+        return
+    # Import after the cheap guard — the registry module needs a resolvable home.
+    from services.subprojects import VALID_CASCADE_OPS, SubprojectManager
+
+    sm = SubprojectManager(parent)
+    sub = getattr(args, "sub_cmd", "list") or "list"
+    target = getattr(args, "target", None)
+    as_json = getattr(args, "json", False)
+
+    if sub == "add":
+        if not target:
+            print("Usage: c3 sub add <folder> [--parent PATH] [--name NAME]")
+            return
+        result = sm.add(
+            target,
+            name=getattr(args, "name", None),
+            ide=getattr(args, "ide", None),
+            run_init=not getattr(args, "no_init", False),
+            reindex_parent=not getattr(args, "no_reindex_parent", False),
+        )
+        if as_json:
+            print(json.dumps(result, indent=2))
+            return
+        if not result.get("added"):
+            print(f"Failed: {result.get('error')}")
+            return
+        verb = "Adopted (existing .c3 kept)" if result.get("adopted") else "Initialized"
+        print(f"\n[OK] {verb}: {result['name']}  ({result['path']})")
+        code = (result.get("parent_reindex") or {}).get("code")
+        if code:
+            print(f"  Parent reindexed: {code.get('files_indexed', '?')} files, "
+                  f"{code.get('chunks_created', '?')} chunks (sub-project now excluded)")
+
+    elif sub == "list":
+        report = sm.reconcile(fix=False)  # report-only consistency pass
+        children = sm.list()
+        if as_json:
+            print(json.dumps({"children": children, "orphans": report.get("orphans", [])}, indent=2))
+            return
+        if not children:
+            print("No sub-projects designated. Use `c3 sub add <folder>`.")
+            return
+        fmt = "{:<22} {:<16} {:>6} {:>7}  {}"
+        print(fmt.format("NAME", "STATUS", "FACTS", "ALERTS", "REL PATH"))
+        print("-" * 76)
+        for c in children:
+            print(fmt.format(
+                (c.get("name") or "?")[:21],
+                c.get("status", "?"),
+                c.get("facts_count", 0),
+                c.get("notification_count", 0),
+                c.get("rel_path", ""),
+            ))
+        issues = sum(1 for c in children if c["status"] != "ok")
+        line = f"\n{len(children)} sub-project(s)"
+        if issues:
+            line += f" -- {issues} with issues (run `c3 sub check --fix`)"
+        if report.get("orphans"):
+            line += f" -- {len(report['orphans'])} registry orphan(s)"
+        print(line)
+
+    elif sub == "remove":
+        if not target:
+            print("Usage: c3 sub remove <name|path> [--clear] [--yes]")
+            return
+        mode = "clear" if getattr(args, "clear", False) else "unlink"
+        if mode == "clear" and not getattr(args, "yes", False):
+            print("This will DELETE the sub-project's .c3 directory and instruction docs.")
+            confirm = input("Type 'clear' to confirm: ").strip().lower()
+            if confirm != "clear":
+                print("Aborted.")
+                return
+        result = sm.remove(target, mode=mode,
+                           reindex_parent=not getattr(args, "no_reindex_parent", False))
+        if as_json:
+            print(json.dumps(result, indent=2))
+            return
+        if not result.get("removed"):
+            print(f"Failed: {result.get('error')}")
+            return
+        print(f"\n[OK] {'Cleared' if mode == 'clear' else 'Unlinked'}: "
+              f"{result.get('name')}  ({result.get('path')})")
+        for w in result.get("warnings", []):
+            print(f"  warning: {w}")
+
+    elif sub == "run":
+        if target not in VALID_CASCADE_OPS:
+            print(f"Usage: c3 sub run {{{'|'.join(VALID_CASCADE_OPS)}}} [--include-parent] [--json]")
+            return
+        result = sm.cascade(target,
+                            include_parent=getattr(args, "include_parent", False),
+                            mcp=getattr(args, "mcp", False))
+        if as_json:
+            print(json.dumps(result, indent=2))
+            return
+        for row in result["results"]:
+            mark = "OK  " if row["ok"] else "FAIL"
+            extra = f" -- {row.get('error')}" if row.get("error") else ""
+            print(f"  [{mark}] {row['name']:<22} {row['elapsed_ms']:>6}ms{extra}")
+        s = result["summary"]
+        print(f"\n{target}: {s['ok']}/{s['total']} ok, {s['failed']} failed")
+
+    elif sub == "check":
+        result = sm.reconcile(fix=getattr(args, "fix", False),
+                              prune=getattr(args, "prune", False))
+        if as_json:
+            print(json.dumps(result, indent=2))
+            return
+        if not result["children"] and not result["orphans"] and not result["pruned"]:
+            print("No sub-projects designated.")
+            return
+        for c in result["children"]:
+            print(f"  [{c['status']:<16}] {c.get('name') or '?':<22} {c.get('rel_path', '')}")
+        for o in result["orphans"]:
+            print(f"  [orphan_registry ] {o}")
+        for f in result.get("fixed", []):
+            print(f"  fixed: {f.get('action')} -> {f.get('path') or f.get('rel_path')}")
+        for p in result.get("pruned", []):
+            print(f"  pruned: {p.get('rel_path')}")
+        if result["ok"]:
+            print("\nAll links consistent.")
+        else:
+            hint = "" if getattr(args, "fix", False) else " Run `c3 sub check --fix` to repair."
+            print(f"\nIssues found.{hint}")
+
+
 def cmd_session_benchmark(args):
     """Run real-world session workflow benchmark."""
     if getattr(args, "command", "") == "session-benchmark":
@@ -6577,6 +6712,7 @@ def main():
         "terse": cmd_terse,
         "ui": cmd_ui,
         "projects": cmd_projects,
+        "sub": cmd_sub,
         "hub": cmd_hub,
         "bitbucket": cmd_bitbucket,
         "oracle": cmd_oracle,

{code_context_control-2.43.0 → code_context_control-2.44.0}/cli/commands/parser.py

@@ -138,6 +138,33 @@ def build_parser(version: str, parse_cli_ide_arg):
     )
     p_projects.add_argument("--name", default=None, help="Display name (for add)")
 
+    p_sub = subparsers.add_parser("sub", help="Manage sub-projects (linked child .c3 branches)")
+    p_sub.add_argument(
+        "sub_cmd",
+        nargs="?",
+        choices=["add", "list", "remove", "run", "check"],
+        default="list",
+        help="Sub-command (default: list)",
+    )
+    p_sub.add_argument(
+        "target",
+        nargs="?",
+        default=None,
+        help="Folder (add), sub-project name/path (remove), or operation update|reindex|health (run)",
+    )
+    p_sub.add_argument("--parent", default=".", help="Parent project path (default: current directory)")
+    p_sub.add_argument("--name", default=None, help="Display name for the sub-project (add)")
+    p_sub.add_argument("--ide", default=None, type=parse_cli_ide_arg, help="IDE for the sub-project init (add)")
+    p_sub.add_argument("--no-reindex-parent", action="store_true", help="Skip the parent reindex after add/remove")
+    p_sub.add_argument("--no-init", action="store_true", help="Link only; skip running init in the folder (add)")
+    p_sub.add_argument("--clear", action="store_true", help="Also wipe the sub-project's .c3 and unregister it (remove; default keeps .c3)")
+    p_sub.add_argument("--yes", action="store_true", help="Skip confirmation prompts")
+    p_sub.add_argument("--include-parent", action="store_true", help="Also run the operation on the parent (run)")
+    p_sub.add_argument("--mcp", action="store_true", help="Also reinstall MCP config on update (run update)")
+    p_sub.add_argument("--fix", action="store_true", help="Repair links from the parent config (check)")
+    p_sub.add_argument("--prune", action="store_true", help="With --fix: drop entries whose folder is gone (check)")
+    p_sub.add_argument("--json", action="store_true", help="Emit JSON output")
+
     p_perms = subparsers.add_parser("permissions",
         help="Manage Claude Code permissions — show | preview <tier> | diff | clean | <tier>")
     p_perms.add_argument("tier", nargs="?", default="show",