refactor(p1): one source-facts module owns function ranges, callers, declaration tables and language aliases

src/openultrasast/source.py replaces the copies in semantic/functions.py,
plane/tasks/repo_facts.py, learn/features.py, learn/excerpt.py and
plane/tasks/verify.py. Two engines share one sentinel (<module>), one
DECLARATION/KEYWORDS/RESERVED set and one alias map: the parser engine
(ranges, enclosing, bounds, call_sites, callers_of, facts_for) for the
scorer, the overlay and the hunter's annotations; the line-pattern engine
(declared_functions, enclosing_at, function_span, function_bounds, the brace
matcher) that labels, features, roles, excerpts and plane facts were built
on. learn no longer imports plane.tasks. Frozen records that store the old
sentinel are mapped on read (module_function).

Gate: labels rebuilt before and after are identical, 7,921 rows, after
sentinel normalisation (benchmarks/measurements/2026-10-13-source-module-
label-diff/record.json). 490 targeted tests pass; 3,312 collect. Net: 657
lines deleted, 212 added. benchmarks/unseen keeps <global> for its own
frozen manifests.

Co-Authored-By: Claude Fable 5.1 <[email protected]>

benchmarks/pairs/advisory-fixes/materialize.py

@@ -106,10 +106,10 @@ def _body(lines: list[str], span: tuple[int, int]) -> str:
 
 
 def extract(source: str, language: str, function: str) -> tuple[list[str], tuple[int, int]] | None:
-    from openultrasast.learn.excerpt import LANGUAGE_ALIASES, function_bounds
+    from openultrasast.source import canonical_language, function_bounds
 
     lines = source.splitlines()
-    span = function_bounds(lines, LANGUAGE_ALIASES.get(language, language), function)
+    span = function_bounds(lines, canonical_language(language), function)
     return (lines, span) if span else None
 
 
@@ -159,14 +159,14 @@ def materialise(pair: dict[str, Any], *, force: bool = False) -> dict[str, Any]:
 
 def verify(pair: dict[str, Any]) -> dict[str, Any]:
     """The label builder's check, from the cached excerpts: declared on both sides, body changed, a real fix."""
-    from openultrasast.learn.excerpt import LANGUAGE_ALIASES, function_bounds
+    from openultrasast.source import canonical_language, function_bounds
 
     result: dict[str, Any] = {"name": pair["name"], "ok": False, "bytes_read": 0}
     commit, parent = str(pair.get("commit", "")), str(pair.get("parent", ""))
     if not (_SHA.fullmatch(commit) and _SHA.fullmatch(parent)) or commit == parent:
         result["reason"] = "fix commit must be a full SHA distinct from its parent"
         return result
-    language = LANGUAGE_ALIASES.get(str(pair["language"]), str(pair["language"]))
+    language = canonical_language(str(pair["language"]))
     spans = {}
     for side, path in zip(("vuln", "fixed"), excerpt_paths(pair), strict=True):
         if not path.is_file():

benchmarks/pairs/catalog_gen.py

@@ -15,7 +15,8 @@ import tomllib
 from pathlib import Path
 
 try:
-    from openultrasast.semantic.functions import ANON_FUNCTION, named_function_ranges
+    from openultrasast.source import ANON_FUNCTION
+    from openultrasast.source import ranges as source_ranges
 except ImportError as exc:  # never degrade into "no named function": that path deletes corpora under --prune
     raise SystemExit(f"catalog_gen needs the openultrasast package (run it with the project venv): {exc}") from exc
 
@@ -268,7 +269,7 @@ def _language(recipe: dict[str, object]) -> str:
 
 
 def _ranges(path: Path, language: str) -> tuple[tuple[str, int, int], ...] | None:
-    ranges = named_function_ranges(str(path), path.read_text(errors="ignore"), language)
+    ranges = source_ranges(str(path), path.read_text(errors="ignore"), language)
     if ranges is None:
         return None  # parser rejected the excerpt; a regex label would not scope
     return tuple(r for r in ranges if r[0] != ANON_FUNCTION)