Collada-dae-parser NPM

collada-dae-parser

Parse collada .dae file vertex positions, textures, normals and animations

What does it do?

collada-dae-parser parses a collada file and outputs JSON. This is useful for displaying skeletal animations in the browser.

collada-dae-parser is only concerned with giving you JSON. An animation system is outside of this modules scope, but skeletal-animation-system could be a useful start.

To Install

# API
$ npm install --save collada-dae-parser
# CLI
$ npm install -g collada-dae-parser

Making use of the parser

If you're unfamiliar with skeletal animation, the demo is a good starting point. Here's where we're parsing our collada file and buffering our graphics data. We're parsing at runtime in the demo, but in a real application you'd want to parse your collada files during a build step.

# Run the demo locally. Changes to the `src` and `demo` directories will live reload in your browser
# PRs and issues are welcome!
git clone https://github.com/chinedufn/collada-dae-parser
cd collada-dae-parser
npm install
npm run demo

Collada Support

collada-dae-parser tries to be useful for WebGL games and interactive demos, but does not try to support the entire collada spec. If you're trying to parse a model that is not supported, collada-dae-parser will try to let you know how to tweak it.

CLI

Output stringified JSON to stdout

# parse from stdin
cat my-3d-model.dae | dae2json > parsed-model.json

# parse from file
dae2json my-3d-modal.dae > parsed-model.json

API

`parseDae(xmlFile)` -> `object`

This function returns the parsed collada object.

xmlFile

Required

Type: string or Buffer

Your collada file data. Not the filename, the file contents.

Returned Object

TODO: Document this

var parseDae = require('collada-dae-parser')
var parsedCollada = parseDae(fs.readFileSync(fileName))
console.log(parsedCollada)
/*
  {
    jointNamePositionIndex: {...},
    jointInverseBindPoses: {...},
    keyframes: {...},
    vertexNormalIndices: [...],
    vertexNormals: [...],
    vertexPositionIndices: [...],
    vertexPositions: [...],
    vertexUVIndices: [...],
    vertexUVs: [...]
  }
*/

jointNamePositionIndex

Type: Object

lorem ipsum

jointInverseBindPoses

Type: Object

lorem ipsum

keyframes

Type: Object

lorem ipsum

vertexNormalIndices

Type: Array

lorem ipsum

vertexNormals

Type: Array

lorem ipsum

vertexPositionIndices

Type: Array

lorem ipsum

vertexPositions

Type: Array

lorem ipsum

vertexUVIndices

Type: Array

lorem ipsum

vertexUVs

Type: Array

lorem ipsum

TODO:

src: basic cli (potentially pull into own repo, but start here)
src: Allow file buffer to be passed in
src: Factor in bind shape matrix
src: Stop exporting bind shape matrix
demo: fix normals in demo lighting
demo: toggle between 2 animations
src: Remove callback from API
warning: Throw descriptive error message if user attempts to export multiple geometries
A separate package that uses collada-dae-parser to implement a stateless skeletal animation system
src / demo: add a textured demo model
src: rounding values. Currently lots of .999999 and 1.000001
src: All of the TODO: statements in code
demo: full screen demo with controls overlay
demo: Support mobile touch events
demo: allow zoom in, zoom out in demo
doc: Add a GIF of every test fixture animation and demo
doc: Documentation
warning: Somehow let the user know if their joints use non rigid transforms - until we actually support this
warning: You didn't export a geometry. Link them to the documentation
warning: If the base model or armature's location, translation and scale aren't 1.0 let user know that we don't support that. They should apply location, translation and scale before exporting
research: Look into supporting library materials

References

waZim. (2010) "Step by Step Skeletal Animation in C++ and OpenGL, Using COLLADA Part 1 & Part 2"
- Heavily informed the initial parser
Markus Ruh. (2012) Vertex Skinning
- Heavily informed the vertex skinning in the initial demo vertex shader
Jerimiah van Oosten. (2011) "GPU Skinning of MD5 Models in OpenGL and Cg"
- Heavily informed the vertex skinning in the initial demo vertex shader