Avoiding Rate Limit (Error 429)

[vihdutta]

On larger projects, its difficult to download the entire assembly without hitting a rate limit. I was hoping there could be an option in the config file to pace the API requests to ensure its within the API limits of OnShape.

I got the error on a project with ~1000 parts below:

! ERROR (429) while using Onshape API
! {
  "message" : "Too many requests.",
  "moreInfoUrl" : "",
  "status" : 429,
  "code" : 0
}

[Gregwar]

Hello,

I am not sure of the values that should be used, I am checking with Onshape team

They give numbers expressed in requests/year (?) here:
https://onshape-public.github.io/docs/auth/limits/

In your experience, how does 429 behave, is it just blocking for a short time and then letting you send requests again soon?

[Gregwar]

I am still waiting for more feedbacks from OnShape team about this

[T-K-233]

It seems that Onshape is enforcing this API limit starting from Feburary.

Perhaps we can get around this by integrating this tool as an Onshape Application?

[Gregwar]

Hello,

I indeed saw this
https://www.reddit.com/r/Onshape/s/UgYsmFPPaE

I will talk to Onshape team soon and try to see what options are. The announced rates are really low and would make export tools likely not viable.

For what I understand, creating an integrated app would mean hosting it, which imply having a server running non stop and serving as an intermediary for requests, or maybe doing the whole export for you. I feel like this is bloating very much the workflow.

[Gregwar]

Keeping updates after a first exchange with Onshape.

They suggest an integrated app would be a good solution, although I feel like this is a lot of work, a need to invest in dedicated server, and a loss of privacy for all the users (if I understand correctly, users would have to rely on the app to issue the requests).

Also, they say native URDF export was being considered, although no confirmed ETA. I don't feel really motivated in investing efforts in an integrated app that would become obsolete without having a clear timeline in mind.

I should have more information this week

[leeygang]

Keeping updates after a first exchange with Onshape.

They suggest an integrated app would be a good solution, although I feel like this is a lot of work, a need to invest in dedicated server, and a loss of privacy for all the users (if I understand correctly, users would have to rely on the app to issue the requests).

Also, they say native URDF export was being considered, although no confirmed ETA. I don't feel really motivated in investing efforts in an integrated app that would become obsolete without having a clear timeline in mind.

I should have more information this week

@Gregwar is there any update?

[Gregwar]

Integrated app is on the way and waiting for approval

[T-K-233]

In the latest release Onshape provides a built-in URDF exporter: https://forum.onshape.com/discussion/30410

[T-K-233]

^ Tried their built-in exporter a bit. The exported file is quite unusable, and there does not seem to have any option for custom configurations...

[marcin-nrobotics]

thanks for engaging with them @Gregwar! will the integrated app have the same config functionality? will it allow for custom postprocessors?

[Gregwar]

Actually, you can apply post processors yourself from the intermediate representation, this is detailed here:
https://onshape-to-robot.readthedocs.io/en/latest/processors.html#retrieve-and-convert-modes

• If you use --retrieve, the export will stop after everything is retrieved from onshape, creating a robot.pkl
• You can force the robot.pkl to be produced anyway using --save-pickle
• If you use --convert, the robot.pkl will be loaded and only the post processing will be applied

Thus, you should be able to use the integrated app to retrieve only and apply the processing at home

I encourage you to use this feature as well since it can save API calls on your side, when iterating on the post processing part alone

[Gregwar]

^ Tried their built-in exporter a bit. The exported file is quite unusable, and there does not seem to have any option for custom configurations...

I am still waiting for news, the integrated app is technically ready since more than a month but currently waiting for approval

[Gregwar]

Ok, the application seems live now 🙂
Don't hesitate to tell me if it works for you

[leeygang2]

when I subscribe the Onshape-to-Robot application and then try to grant the access to my model, I hit the error:

Oops! An Error Occurred
The server returned a "404 Not Found".

Something is broken. Please let us know what you were doing when this error occurred. We will fix it as soon as possible. Sorry for any inconvenience caused.

[T-K-233]

Thank you so much @Gregwar. The Onshape-to-Robot Application works very well on my end.

Sharing for other's reference, here's my workflow using this app:

1. After subscribing to the application, there will be a onshape-to-robot icon show up when viewing an assembly.

2. I copy-paste my local config.json content to the "ADVANCED" tab, and do "retrieve only" run

3. The export take some time, but then it will provide a zip file containing the config.json, robot.pkl, and the STL mesh files.

4. Then I can simply download and unpack the zip file to my local project, and use the previous workflow to postprocess the stuff.

Hope this helps. Here's my entire code workflow also for reference.