sourcecode 0.47.0__py3-none-any.whl → 0.49.0__py3-none-any.whl

sourcecode/__init__.py

@@ -1,3 +1,3 @@
 """sourcecode — Deterministic codebase context maps for AI coding agents."""
 
-__version__ = "0.47.0"
+__version__ = "0.49.0"

sourcecode/cli.py

@@ -1445,8 +1445,10 @@ def main(
         data = agent_view(sm)
         # When contract pipeline ran (mode=contract, no legacy flags), include
         # per-file contracts in agent output so agents get structural context.
+        # Remove file_relevance — contracts cover this signal with more detail.
         if _is_contract_mode and sm.file_contracts:
             from sourcecode.serializer import _serialize_contract_minimal
+            data.pop("file_relevance", None)
             data["contracts"] = [_serialize_contract_minimal(c) for c in sm.file_contracts]
             if sm.contract_summary is not None:
                 cs = sm.contract_summary

sourcecode/context_scorer.py

@@ -177,8 +177,8 @@ class ContextScorer:
         node_scores: dict[str, NodeScore],
         contracts: list[Any],
         *,
-        budget: int = 30,
-        min_score: float = 0.05,
+        budget: int = 25,
+        min_score: float = 0.15,
     ) -> list[str]:
         """Greedy minimum-sufficient subgraph selection with diversity re-ranking.
 

sourcecode/contract_pipeline.py

@@ -17,6 +17,7 @@ from typing import Any, Literal, Optional
 
 from sourcecode.ast_extractor import AstExtractor, _LANGUAGE_MAP
 from sourcecode.contract_model import ContractSummary, FileContract
+from sourcecode.entrypoint_classifier import is_production_entry_point
 from sourcecode.ranking_engine import RankingEngine
 from sourcecode.relevance_scorer import RelevanceScorer
 from sourcecode.schema import EntryPoint, MonorepoPackageInfo
@@ -25,7 +26,9 @@ from sourcecode.schema import EntryPoint, MonorepoPackageInfo
 # Constants
 # ---------------------------------------------------------------------------
 
-_MAX_FILES = 500      # hard cap on files extracted per run
+_MAX_FILES = 500          # hard cap on files extracted per run
+_MAX_CONTRACTS = 15       # default top-N output cap — omit rather than flood
+_MIN_CONTRACT_SCORE = 0.15  # drop contracts below this relevance threshold
 _SRC_EXTENSIONS: frozenset[str] = frozenset(_LANGUAGE_MAP.keys())
 
 
@@ -178,12 +181,20 @@ class ContractPipeline:
         max_importers: int = 50,
         semantic_calls: Optional[list] = None,
         code_notes: Optional[list] = None,
+        max_contracts: Optional[int] = _MAX_CONTRACTS,
     ) -> tuple[list[FileContract], ContractSummary]:
         """Run the full extraction pipeline.
 
         Returns (ranked_contracts, summary).
         """
-        entry_paths = {ep.path.replace("\\", "/") for ep in (entry_points or [])}
+        # Only treat production-classified entrypoints as anchors.
+        # Convention/heuristic entrypoints (index.ts barrels, plugin examples)
+        # must not bubble to the top via is_entrypoint=True ranking boost.
+        entry_paths = {
+            ep.path.replace("\\", "/")
+            for ep in (entry_points or [])
+            if is_production_entry_point(ep)
+        }
         scorer = RelevanceScorer(monorepo_packages)
         engine = RankingEngine(monorepo_packages)
 
@@ -299,16 +310,25 @@ class ContractPipeline:
                     max_importers=max_importers,
                 )
 
-        # 9. Entrypoints-only filter
+        # 9. Entrypoints-only filter — production entry points only, no export spill
         if entrypoints_only and not symbol:
-            contracts = [c for c in contracts if c.is_entrypoint or c.exports]
-
-        # 10. Compress types if requested
+            contracts = [c for c in contracts if c.is_entrypoint]
+
+        # 10. Top-N cap — enforce max_contracts when not in symbol-search mode.
+        # Symbol searches must return all matching files; budget applies only to
+        # the default architectural briefing use case.
+        if symbol is None and max_contracts is not None:
+            contracts = [
+                c for c in contracts
+                if c.relevance_score >= _MIN_CONTRACT_SCORE or c.is_entrypoint
+            ][:max_contracts]
+
+        # 11. Compress types if requested
         if compress_types:
             for c in contracts:
                 _compress_contract_types(c)
 
-        # 11. Apply max_symbols limit (limits total exports across all contracts)
+        # 12. Apply max_symbols limit (limits total exports across all contracts)
         if max_symbols is not None and max_symbols > 0:
             contracts = _limit_symbols(contracts, max_symbols)
 

sourcecode/prepare_context.py

@@ -758,7 +758,7 @@ class TaskContextBuilder:
                 )
                 for total, path, rf in scored
             }
-            _selected = _ctx.select_subgraph(_ns, contracts=[], budget=15, min_score=0.05)
+            _selected = _ctx.select_subgraph(_ns, contracts=[], budget=15, min_score=0.15)
             _rf_map = {path: rf for _, path, rf in scored}
             return [_rf_map[p] for p in _selected if p in _rf_map]
         except Exception:

sourcecode/relevance_scorer.py

@@ -124,8 +124,9 @@ _HIGH_VALUE_SUFFIXES: frozenset[str] = frozenset({
 })
 
 _ENTRYPOINT_STEMS: frozenset[str] = frozenset({
-    "main", "cli", "app", "server", "index", "__main__",
+    "main", "cli", "app", "server", "__main__",
     "application", "bootstrap", "entry",
+    # "index" removed — barrel export convention, not a runtime entrypoint signal
 })
 
 

sourcecode/serializer.py

@@ -31,7 +31,7 @@ from sourcecode.schema import (
 _EP_PRODUCTION_CAP = 5       # max production entry points in default output
 _EP_DEV_CAP = 3              # max development entry points in default output
 _FILE_RELEVANCE_LIMIT = 10   # max files in file_relevance section
-_FILE_RELEVANCE_MIN_COMBINED = 0.20  # minimum combined score to appear (out of ~2.0 max)
+_FILE_RELEVANCE_MIN_COMBINED = 0.40  # minimum combined score — must earn inclusion
 _PROD_DEPS_CAP = 10          # max production dependencies shown
 _SECONDARY_DEPS_CAP = 5      # max per dev/test/build dependency group
 _MONOREPO_PKGS_CAP = 8       # max workspace/runtime packages shown
@@ -1130,9 +1130,6 @@ def _contract_view_minimal(
         "project": project,
     }
 
-    if sm.metadata.traversal_topology:
-        result["traversal"] = sm.metadata.traversal_topology
-
     # Per-file contracts
     if contracts:
         serialized: list[dict[str, Any]] = []
@@ -1310,6 +1307,10 @@ def _compress_sig(name: str, sig: str, max_len: int = 100) -> str:
     return full
 
 
+_MAX_FN_PER_CONTRACT = 5   # max function signatures per contract (token budget)
+_MAX_SIG_LEN = 60          # max chars per compressed signature
+
+
 def _serialize_contract_minimal(c: Any) -> dict[str, Any]:
     """Serialize one FileContract to minimal format."""
     item: dict[str, Any] = {"path": c.path, "role": c.role}
@@ -1317,38 +1318,47 @@ def _serialize_contract_minimal(c: Any) -> dict[str, Any]:
     if c.is_changed:
         item["changed"] = True
 
-    # Exports: flat string for functions/unknown, {name,k} for others
-    # When all exports are same non-function kind, group them
-    if c.exports:
-        exs: list[Any] = []
-        kinds = {e.kind for e in c.exports}
-        if len(kinds) == 1 and "function" not in kinds and "unknown" not in kinds:
-            # All same non-function kind — compact: {"k": "class", "names": [...]}
-            only_kind = next(iter(kinds))
-            exs = [{"k": only_kind, "names": sorted(e.name for e in c.exports)}]
-        else:
-            for e in sorted(c.exports, key=lambda e: e.name):
-                if e.kind in ("function", "unknown"):
-                    exs.append(e.name)
-                else:
-                    exs.append({"name": e.name, "k": e.kind})
-        item["exports"] = exs
-
-    # External deps (non-stdlib already filtered in extractor)
-    if c.dependencies:
-        item["deps"] = sorted(c.dependencies)
-
-    # Exported function signatures — compressed
+    # Exported function signatures — compressed, capped
     exported_names = {e.name for e in c.exports}
+    fn_names_in_sigs: set[str] = set()
     if c.functions:
         fns = []
         for f in sorted(c.functions, key=lambda f: f.name):
             if not (f.exported or f.name in exported_names):
                 continue
-            fns.append(_compress_sig(f.name, f.signature))
+            fns.append(_compress_sig(f.name, f.signature, max_len=_MAX_SIG_LEN))
+            fn_names_in_sigs.add(f.name)
+            if len(fns) >= _MAX_FN_PER_CONTRACT:
+                break
         if fns:
             item["fn"] = fns
 
+    # Exports: omit function names already shown in fn; keep non-function exports
+    if c.exports:
+        exs: list[Any] = []
+        non_fn_exports = [e for e in c.exports if e.kind not in ("function", "unknown")]
+        fn_exports_not_in_sig = [
+            e for e in c.exports
+            if e.kind in ("function", "unknown") and e.name not in fn_names_in_sigs
+        ]
+        remaining = non_fn_exports + fn_exports_not_in_sig
+        if remaining:
+            kinds = {e.kind for e in remaining}
+            if len(kinds) == 1 and "function" not in kinds and "unknown" not in kinds:
+                only_kind = next(iter(kinds))
+                exs = [{"k": only_kind, "names": sorted(e.name for e in remaining)}]
+            else:
+                for e in sorted(remaining, key=lambda e: e.name):
+                    if e.kind in ("function", "unknown"):
+                        exs.append(e.name)
+                    else:
+                        exs.append({"name": e.name, "k": e.kind})
+            item["exports"] = exs
+
+    # External deps (non-stdlib already filtered in extractor)
+    if c.dependencies:
+        item["deps"] = sorted(c.dependencies)
+
     # Types: skip if fully covered by exports (avoids duplication in model files)
     if c.types:
         export_names_set = {e.name for e in c.exports}
@@ -1363,12 +1373,6 @@ def _serialize_contract_minimal(c: Any) -> dict[str, Any]:
     if c.hooks_used:
         item["hooks"] = c.hooks_used
 
-    # Ranking signals: why this file was ranked here
-    if getattr(c, "ranking_reasons", None):
-        non_trivial = [r for r in c.ranking_reasons if r not in ("source file", "noise")]
-        if non_trivial:
-            item["why"] = non_trivial
-
     return item
 
 
@@ -1409,12 +1413,12 @@ def _contract_view_standard(
              "package_manager": s.package_manager}
             for s in sm.stacks
         ],
-        "entry_points": ep_groups["production"],
+        "entry_points": ep_groups["production"][:_EP_PRODUCTION_CAP],
     }
     if sm.metadata.traversal_topology:
         result["traversal"] = sm.metadata.traversal_topology
     if ep_groups["development"]:
-        result["development_entry_points"] = ep_groups["development"]
+        result["development_entry_points"] = ep_groups["development"][:_EP_DEV_CAP]
 
     if sm.confidence_summary is not None:
         result["confidence"] = {

{sourcecode-0.47.0.dist-info → sourcecode-0.49.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sourcecode
-Version: 0.47.0
+Version: 0.49.0
 Summary: Deterministic codebase context for AI coding agents
 License:                                  Apache License
                                    Version 2.0, January 2004

{sourcecode-0.47.0.dist-info → sourcecode-0.49.0.dist-info}/RECORD

@@ -1,16 +1,16 @@
-sourcecode/__init__.py,sha256=d1ZO7BWBwtFn6pK50mntp9_F0e-laNVU7vBBdeZXfU0,103
+sourcecode/__init__.py,sha256=W3DJGnBZMJZBnvn9pO7FSLfHppERKWNuRgtqy1X-umM,103
 sourcecode/adaptive_scanner.py,sha256=6dh34C2qZXyRbw-8xBhbEwDdXanM6CRFRWayVoYITnA,10190
 sourcecode/architecture_analyzer.py,sha256=O4AXc7l_WTzIXrcAzstqZy-TGKNaFa6p3MzpgVjaO8g,27749
 sourcecode/architecture_summary.py,sha256=rSY5MRiaz4N1YdG0pqDTDuFjSN7PO_Zplx-dtNzv2Yo,19985
 sourcecode/ast_extractor.py,sha256=0OHQwTUBBc9lmqPLryVeB1z8dGIC6NhLlar800CD9oI,41129
 sourcecode/classifier.py,sha256=GKTMN8qKZX7ponSwDJfN08RrasI4CVpq1_gFBgEopps,7093
-sourcecode/cli.py,sha256=BC83AWW3oSIWcQM8IPN0tU-1bSb2NyhkMS_Fyi3bRRc,71887
+sourcecode/cli.py,sha256=YusMOF5OfihL3nBw66LcANRFSiVHugPrXE0vPIycjLQ,72016
 sourcecode/code_notes_analyzer.py,sha256=rRd8bFYV0krjlxxQV0wenwE9K7pVpUQSR7KvSvUQKw4,9226
 sourcecode/confidence_analyzer.py,sha256=HxJMPLI5ulqtkncnv98W4iVO6yMbpQo87VuxiuNbDmY,12167
-sourcecode/context_scorer.py,sha256=nhppAo80fblAqcB9Ns0iQd21TZUrl2mQMo_xzPgavRE,14679
+sourcecode/context_scorer.py,sha256=QpChSpsmaAYz91rXA4Ue5xzQmNz_ZboZN09YOHScq1U,14679
 sourcecode/context_summarizer.py,sha256=CiQrfBEzun949bWvmLabWoj2HhPn6Lw62ofqnsy0FlQ,6503
 sourcecode/contract_model.py,sha256=gCf9-Kj0G7l0lvRTAcRfFAfMgs1Rpizv4mKovQLYUkw,3434
-sourcecode/contract_pipeline.py,sha256=wWkm0_zScUy-Q-mrAG57qysYxjNPOybjObxpwp859e0,24615
+sourcecode/contract_pipeline.py,sha256=eann6PZJ8Bg2zavaOhmfN_oy9rMWNSkKPDW68rRZV8E,25686
 sourcecode/coverage_parser.py,sha256=q0LeZJaX1bnntLu-ImksdBsMlpsVmk_iUfSaB4eaJGo,19702
 sourcecode/dependency_analyzer.py,sha256=Exq0BfInvfS5iAg9xAr6WI2uPNuotkIudTKcYJcRhB8,52757
 sourcecode/doc_analyzer.py,sha256=TttdS7mndKQhyJCfJnnAsyGCJrf-TIL7oXxDlTLUFKE,21248
@@ -20,16 +20,16 @@ sourcecode/file_classifier.py,sha256=_KfFIIolharaIxbSTrCkaWauQIqNHCyor_n47RGyDh8
 sourcecode/git_analyzer.py,sha256=PD3eNWydznQ6KLNpxGzBqizIHoPIKevfwz9Xyf_pDt4,11600
 sourcecode/graph_analyzer.py,sha256=hMOsLLz9B0UnQ4xwbHdgr3bFvqpw0bQ8kN-xmEn3Krk,64156
 sourcecode/metrics_analyzer.py,sha256=e2cFwB9XubFq_dIVsP2PLjpr4wX0N6ulb3ol3sGDUeo,20777
-sourcecode/prepare_context.py,sha256=pSVti4gjfBQM24bBCnEW74uZ_c6XijX6z4f8Bt-DkSY,33943
+sourcecode/prepare_context.py,sha256=n7NghZJt8zPt7bzMVpk6gvHlQfhwDYjuLJjgHSOTfD4,33943
 sourcecode/ranking_engine.py,sha256=virVglafZufioHpZpwktjMvUiL0TZELWQCQnQNV8dFo,9360
 sourcecode/redactor.py,sha256=xuGcadGEHaPw4qZXlMDvzMCsr4VOkdp3oBQptHyJk8c,2884
-sourcecode/relevance_scorer.py,sha256=E74w7nlsNVobO3LqKHiMtBd84ONwGp8uDpwXJEjRtLA,8330
+sourcecode/relevance_scorer.py,sha256=MYF4FFkveAQps9SmTeTlh6ODiBz2F--_hWNeHMLtUHQ,8405
 sourcecode/repo_classifier.py,sha256=FG1vaWKdWXsWdl-S8hjVMiTqcwgaRXkDyvK4rPcOGtQ,22681
 sourcecode/runtime_classifier.py,sha256=zWX3r3HCKHc-qtIobErOa8aKMmaoPYREtJKvPcBGPjQ,14792
 sourcecode/scanner.py,sha256=aM3h9-DCQ3xKpeHpHYdo2vX6T5P95HA_YwZbkAVNwmo,8288
 sourcecode/schema.py,sha256=ofEge9hTWHOTjeWt7ceCDQWzP-uhhenrYX2usjW2KVU,22759
 sourcecode/semantic_analyzer.py,sha256=16EFTgM7ooW0m5gNUKOlTSn7IEMLSzKmzQn-cWaSqjs,82604
-sourcecode/serializer.py,sha256=Bt7HLDu6qt6NllsYhoPEeA1Nrk8nDkwZByymwuF3NKs,62595
+sourcecode/serializer.py,sha256=nh8DNGVPVszy60YnWGVH_sLyskgDN973glPIMzNeFWA,62843
 sourcecode/summarizer.py,sha256=ZuzIdm3t8A-d5MuQL0TSNLrd-L0IQIuguIxeNXMNJf8,16070
 sourcecode/tree_utils.py,sha256=Fj9OIuUksBvgibNd3feog0sMDjVypJzPexp5lvMoYWI,1424
 sourcecode/workspace.py,sha256=X_6NmNnitvT3_38V-JDChydo_sR68s249hLFlrQskU0,8271
@@ -60,8 +60,8 @@ 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-0.47.0.dist-info/METADATA,sha256=5jHdy2fsqrC8KEFvA7Z_ZcunHN_zUyDcYafe0X1tKWs,25209
-sourcecode-0.47.0.dist-info/WHEEL,sha256=QccIxa26bgl1E6uMy58deGWi-0aeIkkangHcxk2kWfw,87
-sourcecode-0.47.0.dist-info/entry_points.txt,sha256=ex3F9rmbXeyDIoFQHtkEqTsKSaJow8F0LrVu8XfIktQ,57
-sourcecode-0.47.0.dist-info/licenses/LICENSE,sha256=7DdHrU9Z_3e7dSvq4ISijZNjnuHo5NIHNiHDouMQ9JU,10491
-sourcecode-0.47.0.dist-info/RECORD,,
+sourcecode-0.49.0.dist-info/METADATA,sha256=5FVQYOuzhccMc8oiJ-tPJPr3XJqrdDzWRWf32W8HqWk,25209
+sourcecode-0.49.0.dist-info/WHEEL,sha256=QccIxa26bgl1E6uMy58deGWi-0aeIkkangHcxk2kWfw,87
+sourcecode-0.49.0.dist-info/entry_points.txt,sha256=ex3F9rmbXeyDIoFQHtkEqTsKSaJow8F0LrVu8XfIktQ,57
+sourcecode-0.49.0.dist-info/licenses/LICENSE,sha256=7DdHrU9Z_3e7dSvq4ISijZNjnuHo5NIHNiHDouMQ9JU,10491
+sourcecode-0.49.0.dist-info/RECORD,,