WIP checkpoint: Creator Studio: Asset Import Pipeline — glTF model import backend

backend/app/routers/assets.py

@@ -1,5 +1,6 @@
 """Asset management router."""
 
+import json
 import os
 from pathlib import Path
 from typing import Optional
@@ -64,6 +65,177 @@ def _sanitize_filename(filename: str) -> str:
     return safe or "file"
 
 
+def _validate_gltf_json(content: bytes) -> dict:
+    """Validate that content is a valid .gltf file (JSON-based glTF).
+
+    Returns the parsed JSON dict on success.
+    Raises ValueError if the file is not valid glTF.
+    """
+    try:
+        data = json.loads(content)
+    except json.JSONDecodeError as e:
+        raise ValueError(f"Invalid JSON in .gltf file: {e}")
+
+    if not isinstance(data, dict):
+        raise ValueError("glTF file root must be a JSON object")
+
+    if "asset" not in data:
+        raise ValueError("Missing required 'asset' property in glTF file")
+
+    if "scenes" not in data and "meshes" not in data and "nodes" not in data:
+        # glTF 2.0 spec: at least one of scenes, meshes, nodes, accessors,
+        # bufferViews, buffers, materials, textures, samplers, skin,
+        # animations, or extensions must be present.
+        # For simplicity, we require at least one of scenes/meshes/nodes.
+        raise ValueError(
+            "glTF file must contain at least one of: scenes, meshes, or nodes"
+        )
+
+    return data
+
+
+def _validate_glb(content: bytes) -> dict:
+    """Validate that content is a valid .glb file (binary glTF).
+
+    Returns the parsed JSON chunk data on success.
+    Raises ValueError if the file is not valid glTF.
+    """
+    if len(content) < 12:
+        raise ValueError("glb file too short")
+
+    # Check magic bytes: must start with b'glTF'
+    magic = content[:4]
+    if magic != b'glTF':
+        raise ValueError(
+            f"Invalid glb magic bytes: expected b'glTF', got {magic!r}"
+        )
+
+    # Version check (we support version 2)
+    version = int.from_bytes(content[4:8], byteorder='little')
+    if version != 2:
+        raise ValueError(f"Unsupported glTF version: {version} (expected 2)")
+
+    # Total length check
+    total_length = int.from_bytes(content[8:12], byteorder='little')
+    if total_length != len(content):
+        raise ValueError(
+            f"glb total length mismatch: header says {total_length}, "
+            f"actual is {len(content)}"
+        )
+
+    # Parse JSON chunk
+    # Chunk format: length (4 bytes) + type (4 bytes) + data
+    offset = 12
+    json_chunk_data = None
+
+    while offset < len(content):
+        if offset + 8 > len(content):
+            raise ValueError("Incomplete chunk header in glb file")
+
+        chunk_length = int.from_bytes(content[offset:offset+4], byteorder='little')
+        chunk_type = content[offset+4:offset+8]
+
+        if chunk_length + 8 > len(content) - offset:
+            raise ValueError("Chunk extends beyond file end")
+
+        if chunk_type == b'JSON':
+            json_chunk_data = content[offset+8:offset+8+chunk_length]
+            # JSON chunk may have padding
+            break
+        elif chunk_type == b'BIN':
+            # Skip binary chunk
+            offset += 8 + chunk_length
+            continue
+        else:
+            # Unknown chunk type - skip it
+            offset += 8 + chunk_length
+            continue
+
+    if json_chunk_data is None:
+        raise ValueError("No JSON chunk found in glb file")
+
+    try:
+        data = json.loads(json_chunk_data)
+    except json.JSONDecodeError as e:
+        raise ValueError(f"Invalid JSON in glb chunk: {e}")
+
+    if not isinstance(data, dict):
+        raise ValueError("glTF file root must be a JSON object")
+
+    if "asset" not in data:
+        raise ValueError("Missing required 'asset' property in glTF file")
+
+    return data
+
+
+def _extract_metadata(data: dict) -> dict:
+    """Extract metadata from parsed glTF data.
+
+    Returns a dict with:
+    - bounding_box: {"min": [...], "max": [...]} or null
+    - vertex_count: int
+    - material_count: int
+    - has_animation: bool
+    """
+    metadata = {
+        "bounding_box": None,
+        "vertex_count": 0,
+        "material_count": 0,
+        "has_animation": False,
+    }
+
+    # Count materials
+    if "materials" in data and isinstance(data["materials"], list):
+        metadata["material_count"] = len(data["materials"])
+
+    # Check for animations
+    if "animations" in data and isinstance(data["animations"], list):
+        metadata["has_animation"] = len(data["animations"]) > 0
+
+    # Count vertices from meshes
+    vertex_count = 0
+    bounds_min = None
+    bounds_max = None
+
+    if "meshes" in data and isinstance(data["meshes"], list):
+        for mesh in data["meshes"]:
+            if "primitives" in mesh and isinstance(mesh["primitives"], list):
+                for primitive in mesh["primitives"]:
+                    # Accessor index for positions
+                    pos_accessor_idx = primitive.get("attributes", {}).get("POSITION")
+                    if pos_accessor_idx is not None:
+                        accessors = data.get("accessors", [])
+                        if isinstance(accessors, list) and 0 <= pos_accessor_idx < len(accessors):
+                            accessor = accessors[pos_accessor_idx]
+                            count = accessor.get("count", 0)
+                            vertex_count += count
+
+                            # Extract bounding box from min/max
+                            if "min" in accessor and "max" in accessor:
+                                acc_min = accessor["min"]
+                                acc_max = accessor["max"]
+                                if bounds_min is None:
+                                    bounds_min = acc_min
+                                    bounds_max = acc_max
+                                else:
+                                    bounds_min = [
+                                        min(a, b) for a, b in zip(bounds_min, acc_min)
+                                    ]
+                                    bounds_max = [
+                                        max(a, b) for a, b in zip(bounds_max, acc_max)
+                                    ]
+
+    metadata["vertex_count"] = vertex_count
+
+    if bounds_min is not None and bounds_max is not None:
+        metadata["bounding_box"] = {
+            "min": bounds_min,
+            "max": bounds_max,
+        }
+
+    return metadata
+
+
 # --- Endpoints ---
 
 @router.post("/", response_model=AssetResponse, status_code=201)
@@ -207,3 +379,86 @@ async def upload_asset_file(
     await db.commit()
     await db.refresh(asset)
     return asset
+
+
+@router.post("/import-gltf", response_model=AssetResponse, status_code=201)
+async def import_gltf(
+    current_user: CurrentUser,
+    file: UploadFile = File(...),
+    db: AsyncSession = Depends(get_db),
+):
+    """Import a glTF or glb file as an asset.
+
+    Accepts a .gltf or .glb file, validates it, extracts metadata
+    (bounding box, vertex count, material count, animation flag),
+    stores the file, and returns an AssetResponse with the new asset's ID.
+    """
+    filename = file.filename or ""
+    lower_filename = filename.lower()
+
+    # Validate file extension
+    if not (lower_filename.endswith(".gltf") or lower_filename.endswith(".glb")):
+        raise HTTPException(
+            status_code=400,
+            detail="File must be a .gltf or .glb file",
+        )
+
+    # Validate file size against configured maximum.
+    max_bytes = settings.MAX_UPLOAD_SIZE_MB * 1024 * 1024
+    content = await file.read()
+    file_size = len(content)
+    if file_size > max_bytes:
+        raise HTTPException(
+            status_code=413,
+            detail=f"File size {file_size} exceeds maximum allowed size {max_bytes} bytes",
+        )
+
+    # Validate file content based on format
+    try:
+        if lower_filename.endswith(".gltf"):
+            gltf_data = _validate_gltf_json(content)
+        else:
+            gltf_data = _validate_glb(content)
+    except ValueError as e:
+        raise HTTPException(
+            status_code=400,
+            detail=f"Invalid glTF file: {e}",
+        )
+
+    # Extract metadata
+    metadata = _extract_metadata(gltf_data)
+
+    # Create the asset record
+    asset = Asset(
+        name=Path(filename).stem,
+        description=None,
+        asset_type=AssetType.MODEL_GLTF,
+        file_path="",  # Will be set after file storage
+        owner_id=current_user.id,
+        world_id=None,
+        is_public=False,
+    )
+    db.add(asset)
+    await db.commit()
+    await db.refresh(asset)
+
+    # Sanitize filename for storage
+    safe_filename = _sanitize_filename(filename)
+    extension = ".gltf" if lower_filename.endswith(".gltf") else ".glb"
+    stored_name = f"{asset.id}_{safe_filename}{extension}"
+
+    # Write the file to the upload directory
+    upload_dir = Path(settings.UPLOAD_DIR)
+    upload_dir.mkdir(parents=True, exist_ok=True)
+    file_path = upload_dir / stored_name
+    file_path.write_bytes(content)
+
+    # Update asset with file info and metadata
+    asset.file_path = str(file_path)
+    asset.file_size = file_size
+    asset.mime_type = "model/gltf+json" if lower_filename.endswith(".gltf") else "model/gltf+binary"
+    asset.metadata_json = json.dumps(metadata)
+
+    await db.commit()
+    await db.refresh(asset)
+    return asset