gaussforge 0.5.0

High-performance Gaussian Splatting format conversion library

GaussForge Python Binding

High-performance Gaussian Splatting format conversion library for Python.

Installation

pip install gaussforge

Quick Start

import gaussforge

# Create instance
gf = gaussforge.GaussForge()

# Read a PLY file
with open("model.ply", "rb") as f:
    data = f.read()

result = gf.read(data, "ply")
if "error" not in result:
    print(f"Loaded {result['data']['numPoints']} points")

# Convert to another format
converted = gf.convert(data, "ply", "splat")
if "error" not in converted:
    with open("output.splat", "wb") as f:
        f.write(converted["data"])

Supported Formats

• ply - Standard PLY format
• compressed.ply - Compressed PLY format
• splat - Splat format
• ksplat - K-Splat format
• spz - SPZ compressed format
• sog - SOG format

API Reference

GaussForge()

Create a new GaussForge instance.

read(data: bytes, format: str, strict: bool = False) -> dict

Read Gaussian data from bytes.

• data: Raw file data as bytes
• format: Input format name
• strict: Enable strict validation (default: False)

Returns a dict with data key containing the parsed Gaussian data, or error key on failure.

write(ir: dict, format: str, strict: bool = False) -> dict

Write Gaussian IR to bytes.

• ir: Gaussian intermediate representation dict
• format: Output format name
• strict: Enable strict validation (default: False)

Returns a dict with data key containing the encoded bytes, or error key on failure.

convert(data: bytes, in_format: str, out_format: str, strict: bool = False) -> dict

Convert between formats directly.

• data: Input file data as bytes
• in_format: Input format name
• out_format: Output format name
• strict: Enable strict validation (default: False)

Returns a dict with data key containing the converted bytes, or error key on failure.

get_model_info(data: bytes, format: str, file_size: int = 0) -> dict

Get detailed model information.

• data: Raw file data as bytes
• format: Input format name
• file_size: Optional file size for reporting

Returns a dict with data key containing model info, or error key on failure.

get_supported_formats() -> list[str]

Get list of supported format names.

get_version() -> str

Get library version string.

Building from Source

cd python
pip install -e .

License

Apache-2.0