Perceptilabs command line options - port number?

There is a verbosity parameter -v for debugging etc.

Is there a parameter to allow the localhost port number to be specified? I would like to to run >1 instance and be able to avoid other localhost conflicts (even if jupyter lab prefers 8888, I’m sure I’ve seen 8080 for something else)

And… are there any other interesting options to explore?

We unfortunately don’t have a port number option yet. It’s something we have had in the pipeline for a looong time, but other priorities have pushed it down.
Hopefully we can get one out in not too long, but no ETA on this yet. I’ll add some more priority to it thanks to your request though :slight_smile:

As for other options, there is one to run with the old kernel, but at this point it’s just unstable and has less features. We will have some interesting feature flags coming out in pl-nightly soon, but they will be obscured with the real release to promote using those features over old ones.

Are there any specific options outside of port number you would be interested in?

Apart from port number, nothing major springs to mind right now (though it could be a bit useful to be able so specify which browser to open by executable path; Chrome is currently my default, but I also have FF, Opera and Edge… sometimes reserving a browser for particular uses is convenient)

Assuming verbosity -v=1 is the default, what does -v=2 show?

I tend to expect 1 = Errors, 2 = 1 + Warnings, 3 = 2 + Info…

I just took a look in the code, currently it is:
always: stdin
-v=1: stderr
-v=2: stdout

In hindsight I’m not sure where I got -v=3 from, it should be the same as -v=2.

1 Like