feat: publish collection composition

scripts/common.py

@@ -808,6 +808,35 @@ def unique_emulator_profiles(profiles: dict[str, dict]) -> dict[str, dict]:
     }
 
 
+GAME_DATA_TOPS = ("RPG Maker", "ScummVM")
+
+
+def compute_composition(db: dict) -> dict:
+    """File and byte counts by tree area.
+
+    Three buckets, each re-derivable from paths alone: arcade ROM sets
+    (Arcade/), game and engine data (the RPG Maker/ and ScummVM/ trees),
+    and console or computer system files (everything else).
+    """
+    buckets = {
+        "systems": {"files": 0, "size_bytes": 0},
+        "arcade": {"files": 0, "size_bytes": 0},
+        "game_data": {"files": 0, "size_bytes": 0},
+    }
+    for entry in db.get("files", {}).values():
+        parts = entry.get("path", "").split("/")
+        top = parts[1] if len(parts) > 1 else ""
+        if top == "Arcade":
+            bucket = buckets["arcade"]
+        elif top in GAME_DATA_TOPS:
+            bucket = buckets["game_data"]
+        else:
+            bucket = buckets["systems"]
+        bucket["files"] += 1
+        bucket["size_bytes"] += entry.get("size", 0)
+    return buckets
+
+
 def group_identical_platforms(
     platforms: list[str],
     platforms_dir: str,

scripts/generate_readme.py

@@ -18,6 +18,7 @@ from datetime import datetime, timezone
 
 sys.path.insert(0, os.path.dirname(__file__))
 from common import (
+    compute_composition,
     list_registered_platforms,
     load_database,
     load_emulator_profiles,
@@ -143,6 +144,7 @@ def generate_readme(db: dict, platforms_dir: str) -> str:
         load_emulator_profiles("emulators", skip_aliases=False)
     )
     emulator_count = len(profiles)
+    comp = compute_composition(db)
 
     system_ids: set[str] = set()
     for p in profiles.values():
@@ -231,7 +233,10 @@ def generate_readme(db: dict, platforms_dir: str) -> str:
             f"- **{len(coverages)} platforms** supported with platform-specific verification",
             f"- **{emulator_count} emulators** profiled from source (RetroArch cores + standalone)",
             f"- **{len(system_ids)} systems** covered (NES, SNES, PlayStation, Saturn, Dreamcast, ...)",
-            f"- **{total_files:,} files** verified with MD5, SHA1, CRC32 checksums",
+            f"- **{total_files:,} files** verified with MD5, SHA1, CRC32 checksums:"
+            f" {comp['systems']['files']:,} system files,"
+            f" {comp['arcade']['files']:,} arcade ROM sets,"
+            f" {comp['game_data']['files']:,} game and engine data files",
             *_catalog_matched_line(db),
             f"- **{size_mb:.0f} MB** total collection size",
             "",

scripts/generate_site.py

@@ -21,6 +21,7 @@ from pathlib import Path
 
 sys.path.insert(0, os.path.dirname(__file__))
 from common import (
+    compute_composition,
     list_registered_platforms,
     load_database,
     load_emulator_profiles,
@@ -294,6 +295,8 @@ def generate_home(
         "",
         "</div>",
         "",
+        composition_sentence(db),
+        "",
     ]
 
     # Platforms FIRST (main action)
@@ -417,11 +420,24 @@ def compute_stats(db: dict, coverages: dict, profiles: dict) -> dict:
         "platforms": len(coverages),
         "emulators": len(unique),
         "systems": len(systems),
+        "catalog_matched": sum(
+            1 for f in db.get("files", {}).values() if f.get("provenance")
+        ),
         "source": REPO_URL,
         "downloads": RELEASE_URL,
     }
 
 
+def composition_sentence(db: dict) -> str:
+    comp = compute_composition(db)
+    return (
+        f"Of these files, {comp['systems']['files']:,} are console and "
+        f"computer system files, {comp['arcade']['files']:,} arcade ROM sets "
+        f"(`Arcade/`), and {comp['game_data']['files']:,} game and engine "
+        "data (the `RPG Maker/` and `ScummVM/` trees)."
+    )
+
+
 def generate_stats(stats: dict) -> str:
     return json.dumps(stats, indent=2) + "\n"
 
@@ -2568,6 +2584,7 @@ def main():
     )
 
     stats = compute_stats(db, coverages, profiles)
+    stats["composition"] = compute_composition(db)
     write_if_changed(str(docs / "stats.json"), generate_stats(stats))
 
     # Build system_id -> manufacturer page map (needed by all generators)