fix: read the clone map beside the scanned tree

database.json

@@ -1,8 +1,9 @@
 {
   "schema_version": 1,
-  "generated_at": "2026-10-09T20:11:24Z",
+  "generated_at": "2026-10-09T20:41:14Z",
   "total_files": 19002,
   "total_size": 17497698358,
+  "bios_dir": "bios",
   "files": {
     "738c7f5fd0c8b57ee2e87ae7a97bf8e21a821d07": {
       "path": "bios/3D Realms/Duke Nukem 3D/DUKE.RTS",

schemas/database.schema.json

@@ -4,55 +4,136 @@
   "title": "RetroBIOS content database",
   "type": "object",
   "additionalProperties": false,
-  "required": ["schema_version", "generated_at", "total_files", "total_size", "files", "indexes"],
+  "required": [
+    "schema_version",
+    "generated_at",
+    "total_files",
+    "total_size",
+    "files",
+    "indexes"
+  ],
   "properties": {
-    "schema_version": { "const": 1 },
+    "schema_version": {
-    "generated_at": { "type": "string", "format": "date-time" },
+      "const": 1
-    "total_files": { "type": "integer", "minimum": 0 },
+    },
-    "total_size": { "type": "integer", "minimum": 0 },
+    "generated_at": {
+      "type": "string",
+      "format": "date-time"
+    },
+    "total_files": {
+      "type": "integer",
+      "minimum": 0
+    },
+    "total_size": {
+      "type": "integer",
+      "minimum": 0
+    },
+    "bios_dir": {
+      "type": "string",
+      "description": "Directory the files were scanned from; _mame_clones.json sits beside it."
+    },
     "files": {
       "type": "object",
-      "propertyNames": { "pattern": "^[a-f0-9]{40}$" },
+      "propertyNames": {
+        "pattern": "^[a-f0-9]{40}$"
+      },
       "additionalProperties": {
         "type": "object",
         "additionalProperties": false,
-        "required": ["path", "name", "size", "sha1", "md5", "sha256", "crc32", "adler32"],
+        "required": [
+          "path",
+          "name",
+          "size",
+          "sha1",
+          "md5",
+          "sha256",
+          "crc32",
+          "adler32"
+        ],
         "properties": {
-          "path": { "type": "string", "minLength": 1 },
+          "path": {
-          "name": { "type": "string", "minLength": 1 },
+            "type": "string",
-          "size": { "type": "integer", "minimum": 0 },
+            "minLength": 1
-          "sha1": { "type": "string", "pattern": "^[a-f0-9]{40}$" },
+          },
-          "md5": { "type": "string", "pattern": "^[a-f0-9]{32}$" },
+          "name": {
-          "sha256": { "type": "string", "pattern": "^[a-f0-9]{64}$" },
+            "type": "string",
-          "crc32": { "type": "string", "pattern": "^[a-f0-9]{8}$" },
+            "minLength": 1
-          "adler32": { "type": "string", "pattern": "^[a-f0-9]{8}$" },
+          },
-          "provenance": { "type": "object" }
+          "size": {
+            "type": "integer",
+            "minimum": 0
+          },
+          "sha1": {
+            "type": "string",
+            "pattern": "^[a-f0-9]{40}$"
+          },
+          "md5": {
+            "type": "string",
+            "pattern": "^[a-f0-9]{32}$"
+          },
+          "sha256": {
+            "type": "string",
+            "pattern": "^[a-f0-9]{64}$"
+          },
+          "crc32": {
+            "type": "string",
+            "pattern": "^[a-f0-9]{8}$"
+          },
+          "adler32": {
+            "type": "string",
+            "pattern": "^[a-f0-9]{8}$"
+          },
+          "provenance": {
+            "type": "object"
+          }
         }
       }
     },
     "indexes": {
       "type": "object",
       "additionalProperties": false,
-      "required": ["by_md5", "by_name", "by_crc32", "by_sha256", "by_path_suffix"],
+      "required": [
+        "by_md5",
+        "by_name",
+        "by_crc32",
+        "by_sha256",
+        "by_path_suffix"
+      ],
       "properties": {
-        "by_md5": { "$ref": "#/$defs/hashIndex" },
+        "by_md5": {
-        "by_crc32": { "$ref": "#/$defs/hashIndex" },
+          "$ref": "#/$defs/hashIndex"
-        "by_sha256": { "$ref": "#/$defs/hashIndex" },
+        },
-        "by_name": { "$ref": "#/$defs/listIndex" },
+        "by_crc32": {
-        "by_path_suffix": { "$ref": "#/$defs/listIndex" }
+          "$ref": "#/$defs/hashIndex"
+        },
+        "by_sha256": {
+          "$ref": "#/$defs/hashIndex"
+        },
+        "by_name": {
+          "$ref": "#/$defs/listIndex"
+        },
+        "by_path_suffix": {
+          "$ref": "#/$defs/listIndex"
+        }
       }
     }
   },
   "$defs": {
     "hashIndex": {
       "type": "object",
-      "additionalProperties": { "type": "string", "pattern": "^[a-f0-9]{40}$" }
+      "additionalProperties": {
+        "type": "string",
+        "pattern": "^[a-f0-9]{40}$"
+      }
     },
     "listIndex": {
       "type": "object",
       "additionalProperties": {
         "type": "array",
-        "items": { "type": "string", "pattern": "^[a-f0-9]{40}$" },
+        "items": {
+          "type": "string",
+          "pattern": "^[a-f0-9]{40}$"
+        },
         "minItems": 1,
         "uniqueItems": true
       }

scripts/common.py

@@ -912,7 +912,7 @@ def resolve_local_file(
 
     # MAME clone fallback: if a file was deduped, resolve via canonical
     if _depth < 3 and not has_strong_hash:
-        clone_map = get_mame_clone_map()
+        clone_map = get_mame_clone_map(db.get("bios_dir"))
         canonical = clone_map.get(name)
         if canonical and canonical != name:
             canonical_entry = {"name": canonical}
@@ -960,28 +960,34 @@ def resolve_local_file(
     return None, "not_found"
 
 
-_mame_clone_map_cache: dict[str, str] | None = None
+_mame_clone_map_cache: dict[str, dict[str, str]] = {}
 
 
-def get_mame_clone_map() -> dict[str, str]:
+def get_mame_clone_map(bios_dir: str | None = None) -> dict[str, str]:
-    """Load and cache the MAME clone map (clone_name -> canonical_name)."""
+    """The MAME clone map (clone_name -> canonical_name) of a scanned tree.
-    global _mame_clone_map_cache
+
-    if _mame_clone_map_cache is not None:
+    dedup.py writes _mame_clones.json beside the tree it scanned, and the
-        return _mame_clone_map_cache
+    map of one tree says nothing about another: read from the repository
-    clone_path = os.path.join(
+    root whatever ``--bios-dir`` said, an alternate tree's renamed clones
-        os.path.dirname(os.path.dirname(os.path.abspath(__file__))),
+    fell out of its packs. Without a tree the repository's own is read.
-        "_mame_clones.json",
+    """
-    )
+    if bios_dir is None:
+        bios_dir = os.path.join(
+            os.path.dirname(os.path.dirname(os.path.abspath(__file__))), "bios"
+        )
+    clone_path = os.path.join(os.path.dirname(os.path.abspath(bios_dir)), "_mame_clones.json")
+    cached = _mame_clone_map_cache.get(clone_path)
+    if cached is not None:
+        return cached
+    clone_map: dict[str, str] = {}
     if os.path.exists(clone_path):
         with open(clone_path) as f:
             data = json.load(f)
-        _mame_clone_map_cache = {}
         for canonical, info in data.items():
             for clone in info.get("clones", []):
-                _mame_clone_map_cache[clone] = canonical
+                clone_map[clone] = canonical
-    else:
+    _mame_clone_map_cache[clone_path] = clone_map
-        _mame_clone_map_cache = {}
+    return clone_map
-    return _mame_clone_map_cache
 
 
 

scripts/generate_db.py

@@ -386,6 +386,9 @@ def main():
         "generated_at": datetime.now(timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ"),
         "total_files": len(files),
         "total_size": total_size,
+        # The tree these paths belong to: dedup writes _mame_clones.json
+        # beside it, and the resolver reads the map there, not at the repo root.
+        "bios_dir": os.path.normpath(args.bios_dir),
         "files": files,
         "indexes": indexes,
     }

tests/test_e2e.py

@@ -6356,7 +6356,7 @@ struct BurnDriver BurnDrvneogeo = {
 
         db = self._gap_db("v4bios.zip", "a" * 40, "465c4e1c")
         original = cross_reference.get_mame_clone_map
-        cross_reference.get_mame_clone_map = lambda: {"bctvidbs.zip": "v4bios.zip"}
+        cross_reference.get_mame_clone_map = lambda *_a: {"bctvidbs.zip": "v4bios.zip"}
         try:
             entry = {"name": "bctvidbs.zip", "required": True}
             self.assertEqual(self._gap_source(entry, db), "bios")