If you had ever run into 500 or 504 errors when downloading large derivatives (usually in GBs) using the GET :urn/manifest/:derivativeurn endpoint, that was most likely due to a limitation on processing long running GET requests in our backend.
As a workaround we can download the derivatives in chunks by specifying a “Range” option in the request header:
Implementing such a mechanism may sound like a tedious task, but we can take this as a opportunity to download large derivatives in chunks with multiple requests in parallel. And by dong so there’s every chance that we’d get also a significant speed boost.
To do this we can simply fire multiple requests in parallel with the range header option to fetch data chunks - take Node.js and Axios for an example we can put our code together like below:
function fetchChunkAsync(tmpPath, rangeString, tokenString) {
return new Promise( (resolve, reject)=>{
axios({
method: 'get',
headers: {
'Authorization': 'Bearer ' + tokenString,
'Range': 'bytes=' + rangeString
}
url: 'https://developer.api.autodesk.com/modelderivative/v2/designdata/:urn/manifest/:derivativeUrn',
responseType: 'stream'
})
.then( response =>{
response.data.pipe(fs.createWriteStream(tmpPath));
response.data.on('end', ()=> resolve())
}).catch(error => reject(error))
})
}You can calculate the ideal chunk sizes and set up config objects with paths for temporary storages, put them into a array in the right sequence and pass to the above function to fetch the data and return a “Promise” for callback handling when done.
Finally, merge all chunks together in the final file soon as all requests and their temporary persistence wraps up:
Promise.all(chunkConfigs.map(config=>fetchChunkAsync(config.tmpPath, config.rangeString, tokenString))).then(()=>{
chunkConfigs.forEach( config=> {
// concat chunks together and clean up
})
})
We will consider stating this limitation explicitly in the documentation for the relevant endpoints and we may have more to update later. Thanks and until next time!