1st feature request : big models performance fix

[alperdincer]

Thanks for this repo. Congrats on 1st issue :) We are also working on the similar web app with different technologies, but we can switch to this app.

As we tested our app with 35K+ features, we have tested it with this app and it is not working as performant as your demo. For big files, you can switch from GeoJSON to Vector Tiles with https://github.com/mapbox/geojson-vt library before adding to MapBoxGL map.

Your opinions are important for us, may be we can join our repos together in the future to add more features in a single app.

Thanks again for this great app.

A.

[lbutler]

Hi Alper,

Thanks for checking out my app.

I haven't tried it on any larger models, and I'm honestly not surprised that it would be slow.

Agreed that getting this working on larger models is a worthwhile endeavour. I avoided any optimisations in this first release, so there is plenty of low hanging fruit to fix.

From what I understand, the geojson-vt library is already under the hood in mapbox gl js, but I'll double-check. I'll also load some larger models and do some performance testing to see where the slowdowns are occurring.

Below are some of the refactoring ideas I had for the app which may also increase performance

• Move EPANET engine to a separate repo and run calculations and transformation from inp to geojson on a web worker
• Rethink how we store and access time-series data, at the moment I dump it all on the geojson properties, but it may make more sense to store it in memory and access when needed
• Use Deck.GL to do the data visualisation of time series data on the map

Is your app and/or repo public? I want to check it out if possible, even if its just photos or a video.

[alperdincer]

Hi Luke,

Sorry for late response.

We are preparing our web app to show at a forum about water systems and we are quite busy, but after that event, I'll share you detailed information about application and discuss how we can develop together.

Best regards.

[lbutler]

The engine that runs this app is now a separate repo, it can be found at epanet-js.

I also rewrote how the time series data is stored and accessed, the whole project is a mess and probably needs a rewrite but the speed was drastically improved.

I'm going to try it with a larger model this week and then close off this issue if the problem is solved. If not I will try to find some more performance gains.

[lbutler]

I've tried the a couple of larger models, both with around 20k features and things are now working correctly.

There are still a few areas where we could make improvements, such as pushing the simulation to a web worker but for now it's working. I need to do a serious refactor of this app to utilise epanet-js better.

@alperdincer, let me know how it goes and if you see an improvement