extract_glb_textures.py

"""Extract embedded images from a GLB file (stdlib only).

Usage:
    python3 extract_glb_textures.py <input.glb> <output_dir>
"""
import json
import struct
import sys
from pathlib import Path


def extract(glb_path: Path, out_dir: Path) -> None:
    out_dir.mkdir(parents=True, exist_ok=True)
    with open(glb_path, "rb") as f:
        magic, version, total = struct.unpack("<4sII", f.read(12))
        if magic != b"glTF":
            raise SystemExit(f"not a GLB: {glb_path}")
        json_len, json_type = struct.unpack("<II", f.read(8))
        if json_type != 0x4E4F534A:
            raise SystemExit("missing JSON chunk")
        gltf = json.loads(f.read(json_len))
        bin_len, bin_type = struct.unpack("<II", f.read(8))
        if bin_type != 0x004E4942:
            raise SystemExit("missing BIN chunk")
        bin_data = f.read(bin_len)

    images = gltf.get("images", [])
    buffer_views = gltf.get("bufferViews", [])
    materials = gltf.get("materials", [])
    textures = gltf.get("textures", [])

    print(f"GLB: {glb_path}")
    print(f"  JSON: {json_len} bytes, BIN: {bin_len} bytes")
    print(f"  images: {len(images)}, textures: {len(textures)}, materials: {len(materials)}")

    ext_for_mime = {"image/png": "png", "image/jpeg": "jpg", "image/jpg": "jpg"}

    image_paths = []
    for i, img in enumerate(images):
        mime = img.get("mimeType", "image/png")
        ext = ext_for_mime.get(mime, "bin")
        bv_idx = img.get("bufferView")
        if bv_idx is None:
            print(f"  image[{i}]: skipped (no bufferView, name={img.get('name')})")
            image_paths.append(None)
            continue
        bv = buffer_views[bv_idx]
        offset = bv.get("byteOffset", 0)
        length = bv["byteLength"]
        data = bin_data[offset:offset + length]
        name = img.get("name") or f"image_{i}"
        safe = "".join(c if c.isalnum() or c in "._-" else "_" for c in name)
        out = out_dir / f"{i:02d}_{safe}.{ext}"
        out.write_bytes(data)
        print(f"  image[{i}]: {out.name} ({length / 1024:.1f} KB, {mime})")
        image_paths.append(out)

    print("\n[textures]")
    for i, tex in enumerate(textures):
        src = tex.get("source")
        print(f"  texture[{i}] -> image[{src}]"
              + (f" = {image_paths[src].name}" if src is not None and image_paths[src] else ""))

    print("\n[materials]")
    for i, mat in enumerate(materials):
        name = mat.get("name", f"mat_{i}")
        pbr = mat.get("pbrMetallicRoughness", {})
        bct = pbr.get("baseColorTexture", {}).get("index")
        mrt = pbr.get("metallicRoughnessTexture", {}).get("index")
        nrm = mat.get("normalTexture", {}).get("index")
        emi = mat.get("emissiveTexture", {}).get("index")
        occ = mat.get("occlusionTexture", {}).get("index")
        print(f"  material[{i}] {name}:")
        print(f"    baseColor   -> texture[{bct}]" if bct is not None else "    baseColor   -> (none)")
        print(f"    metalRough  -> texture[{mrt}]" if mrt is not None else "    metalRough  -> (none)")
        print(f"    normal      -> texture[{nrm}]" if nrm is not None else "    normal      -> (none)")
        print(f"    emissive    -> texture[{emi}]" if emi is not None else "    emissive    -> (none)")
        print(f"    occlusion   -> texture[{occ}]" if occ is not None else "    occlusion   -> (none)")


if __name__ == "__main__":
    if len(sys.argv) != 3:
        print(__doc__)
        sys.exit(1)
    extract(Path(sys.argv[1]).expanduser(), Path(sys.argv[2]).expanduser())