code-context-control 2.92.2__py3-none-any.whl → 2.92.3__py3-none-any.whl

cli/c3.py

@@ -92,7 +92,7 @@ console = Console() if HAS_RICH else None
 # Config
 CONFIG_DIR = ".c3"
 CONFIG_FILE = ".c3/config.json"
-__version__ = "2.92.2"
+__version__ = "2.92.3"
 
 
 def _compress_file_cli(compressor, path, mode="smart", **kw):

{code_context_control-2.92.2.dist-info → code_context_control-2.92.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: code-context-control
-Version: 2.92.2
+Version: 2.92.3
 Summary: Local MCP code-intelligence for AI coding tools: surgical search/read/edit, agent-config version history, path-level access + masking guards, and a multi-project hub.
 Author-email: Dimitri Tselenchuk <dtselenc@gmail.com>
 License-Expression: Apache-2.0

{code_context_control-2.92.2.dist-info → code_context_control-2.92.3.dist-info}/RECORD

@@ -1,7 +1,7 @@
 cli/__init__.py,sha256=ec66drCZGNMRU4V6ov0zVhYZph1us12Vn8OvG_LJyRY,22
 cli/_hook_utils.py,sha256=lxYVHSqK4FTiFU2avXW2hbOheMbaaZhnAZqqwqtrZHI,20862
 cli/_shell_writes.py,sha256=XzUNlUDljmg-wZxK03xHnRe36WeQxe4tpO9HKVFwP5o,5090
-cli/c3.py,sha256=G4_iBSVNkNBuPUgKBmyK1z3K84wm_j7n7pSgtgUqDGw,390431
+cli/c3.py,sha256=_DNAo4tAgIsq8d-XPXELnqBk9bL03qQaXa-_u8A8gnU,390431
 cli/docs.html,sha256=X9M7qC7LWi8pPJNF-4yEM57jhuZ3WeTESECbdDOeNaA,143172
 cli/edits.html,sha256=UjAhoCmBmQ89cklGvJqzC6eyNP2tc8H6T-e01DVkLvE,43418
 cli/hook_access_guard.py,sha256=4SIX-A4-tlXcOpzV4k5PZvyXibRvw9buqAV1-_doJpg,17200
@@ -115,7 +115,7 @@ cli/ui/components/sessions.js,sha256=FIKtil76B8tCkAmcFV7hlj6GQ_DCJK2jCzvEmdK7NBE
 cli/ui/components/settings.js,sha256=ATbAjBlVIwCNpxq7s191b49a_INQV38iwmySqtJLYwY,79066
 cli/ui/components/sidebar.js,sha256=K2ym2kUgpbyG-EBA_wBIIiqQY8qTatsiBZwzseMWuIQ,9939
 cli/ui/components/tasks.js,sha256=vyKQ3uwoppMwvdEaHlhWXW4oWcAisx4NveqzMhsYqHo,38438
-code_context_control-2.92.2.dist-info/licenses/LICENSE,sha256=l8Kh5QCNWNvR6kIt8L0BUZvc2LAFiHv2c-FnsGnUZf4,11301
+code_context_control-2.92.3.dist-info/licenses/LICENSE,sha256=l8Kh5QCNWNvR6kIt8L0BUZvc2LAFiHv2c-FnsGnUZf4,11301
 core/__init__.py,sha256=TSDCEcM4V7gcZVM3w2ykJaqEUch4Dkon-rivV17T73s,2501
 core/config.py,sha256=YmkcZwedz_lfDM0ZuI_f4xpe98ixPLcSQFcTCHgRiRg,19206
 core/ide.py,sha256=V6VVMVsFdmmcsMyxikjQp7z9xa42CWiHKS-ya-MAcG4,6172
@@ -206,7 +206,7 @@ services/error_reporting.py,sha256=HZ3ru8i5RLf8nq2R4iRnTs5sm1blUxknSbv5hdxuxs0,4
 services/file_memory.py,sha256=vUvgBBdFC_rvCp9GXfvfojQIr6iSHWTRFrzWG5e0kHs,35048
 services/git_context.py,sha256=FFdA8Y46op4DYsUnZsXLTqxIjgxb0q_d-hkFTbosC8Q,9518
 services/hub_service.py,sha256=z2MZ2Rg9NLNXxQ17e05XGF_f6qSTDWLYtoC3WpuQfGU,2634
-services/indexer.py,sha256=Dme1bqX7NDZjmLDWcc1pbnIeLC5jaKUVQ7k-uMJy-ls,31216
+services/indexer.py,sha256=sODQ7kV1dmLAuMVJTnd9JpSaPJsq2eNPIqGziM9sr3g,32963
 services/jira_client.py,sha256=blh2LhjHrWL3OUHMS_ab7JguQqtBbx66m5UYZz7sbXQ,18935
 services/jira_cloud.py,sha256=vOMQxwEuPa-FcLgmch-to7feiPkGrXXjUf9mNgpV4VE,12656
 services/jira_credentials.py,sha256=QmODZa2ivZ1DwsxSpnmM1XhPxNBiTJCR1LrdCPh7Nuk,9124
@@ -281,8 +281,8 @@ tui/screens/search_view.py,sha256=MMHjVdlk3HZSuDBSvq8IGrqv_Mh5Us6YqXQ80bcWSMk,19
 tui/screens/session_view.py,sha256=eZ1eDwHTvPOck1wCCviixtOaCxIkBT_95ytNNNriGNA,5991
 tui/screens/stats.py,sha256=p81PjzdaIv7hllb8f45-rlVe4lJZwSdIMqu7e86_u5s,6223
 tui/screens/ui_view.py,sha256=1QJCgLh2YfgWIpvzRG1KOGXYEaOYX6ojN61Azjf2oX0,2125
-code_context_control-2.92.2.dist-info/METADATA,sha256=c9FOYVgbXn3kqLd4hyYYcFs5oz_ZaLYIhL8MHBZQvNc,26368
-code_context_control-2.92.2.dist-info/WHEEL,sha256=YVMoNqKzERt-wjUZwJ33xBGAwnFl-4cqbYkTtWa4itE,91
-code_context_control-2.92.2.dist-info/entry_points.txt,sha256=7kX_WUsDCF2hbXzvbNyscyaBb9AeA-DJY5v_5hN0DlU,93
-code_context_control-2.92.2.dist-info/top_level.txt,sha256=wRt41zBybVF3qAiNXHz9BURbkKvUvfhmWWtKMhaw6eE,29
-code_context_control-2.92.2.dist-info/RECORD,,
+code_context_control-2.92.3.dist-info/METADATA,sha256=69oubv8g3Xd1SiRr5Dcqn5RFP_ot3WvUF1FucgTOwp4,26368
+code_context_control-2.92.3.dist-info/WHEEL,sha256=YVMoNqKzERt-wjUZwJ33xBGAwnFl-4cqbYkTtWa4itE,91
+code_context_control-2.92.3.dist-info/entry_points.txt,sha256=7kX_WUsDCF2hbXzvbNyscyaBb9AeA-DJY5v_5hN0DlU,93
+code_context_control-2.92.3.dist-info/top_level.txt,sha256=wRt41zBybVF3qAiNXHz9BURbkKvUvfhmWWtKMhaw6eE,29
+code_context_control-2.92.3.dist-info/RECORD,,

services/indexer.py

@@ -37,6 +37,20 @@ def _masked_content(fpath, project_path):
         return None
 
 
+# Co-occurrence is O(unique_tokens^2) per chunk, so one minified vendor
+# bundle can dominate a whole index build. Both bounds below are deliberate
+# and their effect is REPORTED in build_index()'s stats rather than applied
+# silently — a capped build must never look like a complete one.
+#
+# 200: chunks above this are minified/bundled/data blobs, not code. The
+# pruning step keeps only the top-5 co-occurring terms per token with at
+# least 3 hits, so such a chunk contributes noise that is discarded anyway.
+_COOC_MAX_CHUNK_TOKENS = 200
+# 20M: whole-pass ceiling on pair updates, so no repository shape can make
+# this step unbounded. ~10s of CPU on the boxes this runs on.
+_COOC_MAX_PAIR_UPDATES = 20_000_000
+
+
 class CodeIndex:
     """TF-IDF based code search index with structural awareness."""
 
@@ -240,6 +254,7 @@ class CodeIndex:
             "entries_scanned": scan_stats["entries"],
             "files_capped": files_capped,
             "max_files": max_files,
+            "cooccurrence": getattr(self, "_cooccurrence_stats", {}),
         }
 
     def _chunk_file(self, content: str, ext: str, doc_id: str) -> list:
@@ -625,16 +640,39 @@ class CodeIndex:
                 self.chunk_tfidf[chunk_id][term] = normalized_tf * self.idf.get(term, 0)
 
     def _build_cooccurrence(self):
-        """Build lightweight co-occurrence map from indexed chunks for auto-synonyms."""
+        """Build lightweight co-occurrence map from indexed chunks for auto-synonyms.
+
+        Bounded on purpose: the pair loop is quadratic in a chunk's unique
+        token count, so a single minified bundle (798 unique tokens x 4091
+        chunks = 1.9e9 pair updates, measured on a real project) can hang a
+        build that has nothing else wrong with it. Skipped chunks and an
+        exhausted budget are recorded in ``self._cooccurrence_stats`` and
+        returned by :meth:`build_index`.
+        """
         self._cooccurrence = {}
+        skipped = 0
+        updates = 0
+        budget_hit = False
         for chunk in self.chunks.values():
             tokens = set(self._tokenize(chunk["content"]))
+            if len(tokens) > _COOC_MAX_CHUNK_TOKENS:
+                skipped += 1
+                continue
+            if updates + len(tokens) * (len(tokens) - 1) > _COOC_MAX_PAIR_UPDATES:
+                budget_hit = True
+                break
             for t in tokens:
                 if t not in self._cooccurrence:
                     self._cooccurrence[t] = Counter()
                 for t2 in tokens:
                     if t != t2:
                         self._cooccurrence[t][t2] += 1
+            updates += len(tokens) * (len(tokens) - 1)
+        self._cooccurrence_stats = {
+            "chunks_skipped": skipped,
+            "pair_updates": updates,
+            "budget_exhausted": budget_hit,
+        }
         # Prune: keep only top-5 co-occurring terms per token (minimum 3 co-occurrences)
         pruned = {}
         for term, counts in self._cooccurrence.items():