fix: cross-reference checks data/ directories for false positives

_find_in_repo and _name_in_index now scan data/ in addition to
bios/ via database.json. Eliminates 129 false positives from
game data migrated to data/ (OpenTyrian, ScummVM, SDLPAL, Cave
Story, Syobon Action). True missing: 782 -> 653.

scripts/cross_reference.py

@@ -50,7 +50,21 @@ def load_platform_files(platforms_dir: str) -> tuple[dict[str, set[str]], dict[s
     return declared, platform_data_dirs
 
 
-def _find_in_repo(fname: str, by_name: dict[str, list], by_name_lower: dict[str, str]) -> bool:
+def _build_data_dir_index(data_root: str = "data") -> set[str]:
+    """Build a set of filenames available in data/ directories."""
+    names: set[str] = set()
+    root_path = Path(data_root)
+    if not root_path.is_dir():
+        return names
+    for fpath in root_path.rglob("*"):
+        if fpath.is_file() and not fpath.name.startswith("."):
+            names.add(fpath.name)
+            names.add(fpath.name.lower())
+    return names
+
+
+def _find_in_repo(fname: str, by_name: dict[str, list], by_name_lower: dict[str, str],
+                  data_names: set[str] | None = None) -> bool:
     if fname in by_name:
         return True
     basename = fname.rsplit("/", 1)[-1] if "/" in fname else None
@@ -63,6 +77,11 @@ def _find_in_repo(fname: str, by_name: dict[str, list], by_name_lower: dict[str,
         key = basename.lower()
         if key in by_name_lower:
             return True
+    if data_names:
+        if fname in data_names or key in data_names:
+            return True
+        if basename and (basename in data_names or basename.lower() in data_names):
+            return True
     return False
 
 
@@ -71,12 +90,14 @@ def cross_reference(
     declared: dict[str, set[str]],
     db: dict,
     platform_data_dirs: dict[str, set[str]] | None = None,
+    data_names: set[str] | None = None,
 ) -> dict:
     """Compare emulator profiles against platform declarations.
 
     Returns a report with gaps (files emulators need but platforms don't list)
     and coverage stats. Files covered by matching data_directories between
     emulator profile and platform config are not reported as gaps.
+    Checks both bios/ (via database) and data/ (via data_names index).
     """
     platform_data_dirs = platform_data_dirs or {}
     by_name = db.get("indexes", {}).get("by_name", {})
@@ -106,7 +127,7 @@ def cross_reference(
                 continue
 
             in_platform = fname in platform_names
-            in_repo = _find_in_repo(fname, by_name, by_name_lower)
+            in_repo = _find_in_repo(fname, by_name, by_name_lower, data_names)
 
             entry = {
                 "name": fname,
@@ -205,7 +226,8 @@ def main():
 
     declared, plat_data_dirs = load_platform_files(args.platforms_dir)
     db = load_database(args.db)
-    report = cross_reference(profiles, declared, db, plat_data_dirs)
+    data_names = _build_data_dir_index()
+    report = cross_reference(profiles, declared, db, plat_data_dirs, data_names)
 
     if args.json:
         print(json.dumps(report, indent=2))

scripts/verify.py

@@ -246,8 +246,9 @@ def _build_expected(file_entry: dict, checks: list[str]) -> dict:
         expected["adler32"] = adler_val
     return expected
 
-def _name_in_index(name: str, by_name: dict, by_path_suffix: dict | None = None) -> bool:
-    """Check if a name is resolvable in the database indexes."""
+def _name_in_index(name: str, by_name: dict, by_path_suffix: dict | None = None,
+                    data_names: set[str] | None = None) -> bool:
+    """Check if a name is resolvable in the database indexes or data directories."""
     if name in by_name:
         return True
     basename = name.rsplit("/", 1)[-1]
@@ -255,6 +256,11 @@ def _name_in_index(name: str, by_name: dict, by_path_suffix: dict | None = None)
         return True
     if by_path_suffix and name in by_path_suffix:
         return True
+    if data_names:
+        if name in data_names or name.lower() in data_names:
+            return True
+        if basename != name and (basename in data_names or basename.lower() in data_names):
+            return True
     return False
 
 
@@ -264,6 +270,7 @@ def find_undeclared_files(
     db: dict,
     emu_profiles: dict | None = None,
     target_cores: set[str] | None = None,
+    data_names: set[str] | None = None,
 ) -> list[dict]:
     """Find files needed by cores but not declared in platform config."""
     # Collect all filenames declared by this platform
@@ -333,7 +340,7 @@ def find_undeclared_files(
             # Archived files are grouped by archive
             if archive:
                 if archive not in archive_entries:
-                    in_repo = _name_in_index(archive, by_name, by_path_suffix)
+                    in_repo = _name_in_index(archive, by_name, by_path_suffix, data_names)
                     archive_entries[archive] = {
                         "emulator": profile.get("emulator", emu_name),
                         "name": archive,
@@ -364,10 +371,10 @@ def find_undeclared_files(
                 dest = f.get("path") or fname
 
             # Resolution: try name, then path basename, then path_suffix
-            in_repo = _name_in_index(fname, by_name, by_path_suffix)
+            in_repo = _name_in_index(fname, by_name, by_path_suffix, data_names)
             if not in_repo and dest != fname:
                 path_base = dest.rsplit("/", 1)[-1]
-                in_repo = _name_in_index(path_base, by_name, by_path_suffix)
+                in_repo = _name_in_index(path_base, by_name, by_path_suffix, data_names)
 
             checks = _parse_validation(f.get("validation"))
             undeclared.append({
@@ -587,7 +594,10 @@ def verify_platform(
         status_counts[s] = status_counts.get(s, 0) + 1
 
     # Cross-reference undeclared files
-    undeclared = find_undeclared_files(config, emulators_dir, db, emu_profiles, target_cores=target_cores)
+    from cross_reference import _build_data_dir_index
+    data_names = _build_data_dir_index()
+    undeclared = find_undeclared_files(config, emulators_dir, db, emu_profiles,
+                                       target_cores=target_cores, data_names=data_names)
     exclusions = find_exclusion_notes(config, emulators_dir, emu_profiles, target_cores=target_cores)
 
     # Ground truth coverage