oursky / arcms Goto Github PK

• unzip the asset file and preview the model
• example: https://manu.ninja/webgl-3d-model-viewer-using-three-js

Current VODownloader flow are hidden in the chain of completion callback. It makes the code difficult to follow. i.e. you need to go through 5 function in details starting from the public function downloadVirtualObject to know how is the download works.

A better way is to write flow control logic in one function, and keep the actual download logic spread in those function. In your case, there is actually 3 async request. We don't need to separate those codes into 5 async function.

In additional, the download is sequential, which is not necessary.

So please take a look at https://developer.apple.com/documentation/dispatch/dispatchgroup

In result, you will have following code:

self.downloadJSON(url: url, completion: { urls in
    if !urls["textures"] {
        print("Fail to download \(texture)")
        completion(false)
    }
    allDownload =  DispatchGroup()
    self.downloadTextures(url: urls["textures"], dispatchGroup: allDownload, completion: { success in
        if !success {
            print("Fail to download \(texture)")
            completion(false)
        }
    })
    self.downloadSCN(url: urls["scn"], completion: { success in
        if !success {
            print("Fail to download \(texture)")
            completion(false)
        }
    })
    allDownload.notify(queue: .main){
      completion(true)
    }
})

And I found there is various non-symmetry API design. For example, there is removeFromLoading not no addToLoading. addToLoading are do in-line.

• Smoothing the movement of the object to retain the feel of illusion
• Try to use rectangle recognition to see whether tracking can be optimized

Please add the following to .gitignore:

*.pem
*.cert
*.pfx
*.key
.env
docker-compose.override.yml

• on afterSave, extract the zip and generate a preview image from the .scn file
  • the generated preview image shall be attach to the object record
  • then we can display the preview image in CMS

We should have a configurable cache size and persist the files into disk.

• Create lambda function
• The app should call this lambda function to get the data

• Put https://github.com/oursky/arserver_skygear under /cloudcode folder
• and the current project inside /ios folder

• Init project setup
• Buddybuild
• Auto upload to hockeyapp
• Add lint
• Add complexity check

• Read and decode QR code in this format

{url: 'xxxx'}

• After reading the QR code, then to download the 3D object, during download, show HUD loading dialog
• If QR format not correct, show HUD dialog with message "Sorry, not able to recognize this code"
• After 3D object downloaded, close the loading dialog and place the 3D object
• The 3D object should be cached locally after download