#!/usr/bin/env python3
import os
import re
import shlex

# Characters we'll allow directly in the *basename* (before .pk3)
ALLOWED_BASE = set("abcdefghijklmnopqrstuvwxyz0123456789_-")

def sanitize_base(name_base: str) -> str:
    # Lowercase and replace anything not in ALLOWED_BASE with underscore
    s = []
    for ch in name_base.lower():
        if ch in ALLOWED_BASE:
            s.append(ch)
        else:
            s.append("_")
    base = "".join(s)
    # Collapse multiple underscores
    base = re.sub(r"_+", "_", base)
    # Strip leading/trailing underscores
    base = base.strip("_")
    return base or "map"

def main():
    # Work in current directory; restrict to .pk3 files
    all_names = [f for f in os.listdir(".") if os.path.isfile(f)]
    pk3 = [n for n in all_names if n.lower().endswith(".pk3")]

    # 1) Find case-insensitive collisions
    groups = {}
    for name in pk3:
        key = name.lower()
        groups.setdefault(key, []).append(name)
    collision_groups = {k: v for k, v in groups.items() if len(v) > 1}

    # 2) Find names with "weird" characters
    weird = []
    allowed_full = set("abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789._-")
    for name in pk3:
        if any(ch not in allowed_full for ch in name):
            weird.append(name)

    to_fix = set(weird)
    for vs in collision_groups.values():
        to_fix.update(vs)

    if not to_fix:
        print("Nothing to fix: no weird characters or case-colliding names found.")
        return

    print(f"# Found {len(pk3)} .pk3 files")
    print(f"# {len(weird)} names with non-trivial characters")
    print(f"# {len(collision_groups)} case-insensitive collision groups")
    print(f"# Will propose renames for {len(to_fix)} files\n")

    # Names that are *not* being touched stay as-is, so we must not collide with them
    unaffected = set(pk3) - to_fix
    used = set(unaffected)

    rename_map = {}

    # Helper to allocate a unique name
    def allocate(base: str, ext: str) -> str:
        candidate = f"{base}{ext}"
        if candidate not in used:
            used.add(candidate)
            return candidate
        # Try with suffixes
        i = 2
        while True:
            candidate = f"{base}_{i}{ext}"
            if candidate not in used:
                used.add(candidate)
                return candidate
            i += 1

    # 3) Handle collision groups first, so they get clean, related names
    for key, vs in sorted(collision_groups.items()):
        sample = vs[0]
        if "." in sample:
            base_part, ext_part = sample.rsplit(".", 1)
            ext_part = "." + ext_part
        else:
            base_part, ext_part = sample, ""
        sani_base = sanitize_base(base_part)
        ext_lower = ext_part.lower()

        # Prefer already-lowercase names first
        vs_sorted = sorted(vs, key=lambda n: (not n.islower(), n))
        for idx, name in enumerate(vs_sorted):
            new_name = allocate(sani_base, ext_lower)
            rename_map[name] = new_name

    # 4) Handle "weird-only" names (that weren't in a collision group)
    weird_only = sorted(to_fix - set(rename_map.keys()))
    for name in weird_only:
        if "." in name:
            base_part, ext_part = name.rsplit(".", 1)
            ext_part = "." + ext_part
        else:
            base_part, ext_part = name, ""
        sani_base = sanitize_base(base_part)
        ext_lower = ext_part.lower()
        new_name = allocate(sani_base, ext_lower)
        rename_map[name] = new_name

    # 5) Print the mv commands (dry-run by default)
    print("# Review these commands carefully before running them.")
    print("# If they look good, you can save them to a shell script and execute.")
    for old, new in sorted(rename_map.items()):
        if old == new:
            continue
        print("mv -v -- " + shlex.quote(old) + " " + shlex.quote(new))

if __name__ == "__main__":
    main()