rob-webaverse / preview-backend

[System] Preview Service

Preview service is an entirely seperate system residing at Github Repo used by application to generate the preview of any asset.

Usage

https://app.webaverse.com/screenshot.html?url=https://webaverse.github.io/assets/male.vrm&ext=vrm&type=png

Inputs (as Query Params)

• url: {URL of the asset that can be downloadable by the screenshot system} [Required]
• ext: {vrm|glb|vox|png|jpg|jpeg|gif} [Required]
• type: {png|jpeg} [Required]

Output

• Output screenshot will be posted back to the calling service.

Architecture

Flow Diagram

Location

Webaverse App
└───src
   └───screenshot.js

Functions

GLTF/GLB/VRM Loader

Uses: Totum

	let  object;
	try {
		object = await  metaversefileApi.load(url);
	} catch (err) {
		console.warn(err);
	}
	return  object;

• The above code is a common code that applies to GLTF, GLB and load the asset into context.
• Loader uses MetaverFile Moudle that auto detects the extension type and returns us the object as scene app.

VRM Loader

Uses: Totum

	let  object;
	try {
		object = await  metaversefileApi.load(url);
	} catch (err) {
		console.warn(err);
	}
	return  object;

• The above code is a common code that applies to VRM and load the asset into context.
• Loader uses MetaverFile Moudle that auto detects the extension type and returns us the object as scene app.

       const _getTailBones = skeleton => {
         const result = [];
         const _recurse = bones => {
           for (let i = 0; i < bones.length; i++) {
             const bone = bones[i];
             if (bone.children.length === 0) {
               if (!result.includes(bone)) {
                 result.push(bone);
               }
             } else {
               _recurse(bone.children);
             }
           }
         };
         _recurse(skeleton.bones);
         return result;
       };
       const _findFurthestParentBone = (bone, pred) => {
         let result = null;
         for (; bone; bone = bone.parent) {
           if (pred(bone)) {
             result = bone;
           }
         }
         return result;
       };
       const _countCharacters = (name, regex) => {
         let result = 0;
         for (let i = 0; i < name.length; i++) {
           if (regex.test(name[i])) {
             result++;
           }
         }
         return result;
       };
       const _findEye = (tailBones, left) => {
         const regexp = left ? /l/i : /r/i;
         const eyeBones = tailBones.map(tailBone => {
           const eyeBone = _findFurthestParentBone(tailBone, bone => /eye/i.test(bone.name) && regexp.test(bone.name.replace(/eye/gi, '')));
           if (eyeBone) {
             return eyeBone;
           } else {
             return null;
           }
         }).filter(spec => spec).sort((a, b) => {
           const aName = a.name.replace(/shoulder/gi, '');
           const aLeftBalance = _countCharacters(aName, /l/i) - _countCharacters(aName, /r/i);
           const bName = b.name.replace(/shoulder/gi, '');
           const bLeftBalance = _countCharacters(bName, /l/i) - _countCharacters(bName, /r/i);
           if (!left) {
             return aLeftBalance - bLeftBalance;
           } else {
             return bLeftBalance - aLeftBalance;
           }
         });
         const eyeBone = eyeBones.length > 0 ? eyeBones[0] : null;
         if (eyeBone) {
           return eyeBone;
         } else {
           return null;
         }
       };

• The VRM loader is further equipped with the logic to find the eyes of the avatar and center them. It makes use of two important functions _getTailBones & _findEye
• _getTailBones is called on the skeleton that return the tail bones.
• _findEye is called on the tail bones returned from the above step.