scripts/extract_characters.py

#!/usr/bin/env python3
"""
Dev-time script. Extracts buildXxx() functions from playground sandbox JS files.
Writes sandbox_cache/characters_registry.json and sandbox_cache/characters.js.

Usage:
    python3 scripts/extract_characters.py

Source: /Users/bolyos/Desktop/playground/<game>/sandbox/
"""
import json
import re
from pathlib import Path

PLAYGROUND_BASE = Path("/Users/bolyos/Desktop/playground")
OUT_DIR = Path(__file__).parent.parent / "sandbox_cache"
OUT_DIR.mkdir(exist_ok=True)

# Pure Three.js geometry primitives — not characters, exclude them
GEOMETRY_PRIMITIVES = {
    "buildBox", "buildCylinder", "buildSphere", "buildCone",
    "buildIcosahedron", "buildOctahedron", "buildDodecahedron",
    "buildTetrahedron", "buildLathe",
}

# Scan patterns in priority order
SCAN_PATTERNS = [
    "*/sandbox/src/characters.js",  # jungle modular structure — try first
    "*/sandbox/main.js",
    "*/sandbox/bundle.js",
]


def extract_functions(source: str) -> dict[str, str]:
    """Extract top-level function buildXxx() {...} blocks from JS source."""
    registry: dict[str, str] = {}
    pattern = re.compile(r'^(function (build\w+)\([^)]*\)\s*\{)', re.MULTILINE)
    for m in pattern.finditer(source):
        fn_name = m.group(2)
        if fn_name in GEOMETRY_PRIMITIVES:
            continue
        start = m.start()
        depth = 0
        for j in range(m.start(1), len(source)):
            if source[j] == "{":
                depth += 1
            elif source[j] == "}":
                depth -= 1
                if depth == 0:
                    registry[fn_name] = source[start : j + 1]
                    break
    return registry


def game_slug(js_path: Path) -> str:
    """Derive a clean slug from the game directory name."""
    # parent chain: .../playground/<game>/sandbox/[src/]file.js
    parts = js_path.parts
    try:
        pg_idx = next(i for i, p in enumerate(parts) if p == "playground")
        game_name = parts[pg_idx + 1]
    except (StopIteration, IndexError):
        game_name = js_path.parent.parent.name
    return re.sub(r"[^a-zA-Z0-9]+", "_", game_name).strip("_").lower()


# ── Collect all JS files to scan ─────────────────────────────────────────────
js_files: list[Path] = []
for pattern in SCAN_PATTERNS:
    for f in sorted(PLAYGROUND_BASE.glob(pattern)):
        if f not in js_files:
            js_files.append(f)

print(f"Files to scan: {len(js_files)}")
for f in js_files:
    print(f"  {f}")

# ── Extract, resolving buildHero collisions by namespacing ────────────────────
registry: dict[str, str] = {}
hero_sources: dict[str, str] = {}   # fn_name → game_slug for collision tracking

for js_file in js_files:
    slug = game_slug(js_file)
    try:
        text = js_file.read_text(errors="ignore")
    except OSError as e:
        print(f"  WARNING: could not read {js_file}: {e}")
        continue

    found = extract_functions(text)
    print(f"  {js_file.name} ({slug}): {len(found)} functions found")

    for fn_name, fn_body in found.items():
        if fn_name == "buildHero":
            namespaced = f"buildHero_{slug}"
            # Rename the function declaration to match the namespaced key
            fn_body_renamed = fn_body.replace(
                f"function buildHero(", f"function {namespaced}(", 1
            )
            registry[namespaced] = fn_body_renamed
            print(f"    ↳ buildHero → {namespaced}")
        else:
            if fn_name in registry:
                print(f"    ↳ {fn_name} already seen, skipping duplicate")
            else:
                registry[fn_name] = fn_body

print(f"\nTotal unique build functions: {len(registry)}")

# ── Character functions (names containing 'Character', 'Songoku', 'Hero', or Pokemon names) ──
character_fns = sorted([
    k for k in registry
    if any(tag in k for tag in ["Character", "Hero", "Songoku",
                                "Raticate", "Persian", "Tauros", "Snorlax",
                                "Graveler", "Onix", "Zubat", "Golbat",
                                "Pidgey", "Fearow", "Beedrill", "Butterfree",
                                "Voltorb", "Electrode", "Jigglypuff", "Abra",
                                "Alakazam", "Gengar", "Doduo", "Rapidash"])
])
print(f"\nCharacter functions ({len(character_fns)}): {character_fns}")

# ── Write outputs ─────────────────────────────────────────────────────────────
json_path = OUT_DIR / "characters_registry.json"
json_path.write_text(json.dumps(registry, indent=2))
print(f"\nWritten: {json_path} ({json_path.stat().st_size} bytes)")

js_path = OUT_DIR / "characters.js"
js_path.write_text("\n\n".join(registry.values()) + "\n")
print(f"Written: {js_path} ({js_path.stat().st_size} bytes)")