search

convexengineering / gplibrary

Useful subsystem models
MIT License
10 stars 11 forks source link

use parsing notation #137

Closed mjburton11 closed 6 years ago

bqpd commented 6 years ago

test models please

mjburton11 commented 6 years ago

This is definitely going to fail tests on gas/solar. Is there a way to make this backwards compatible with all the repos so they don't break everything? or should I just merge it when ready and then go and fix everything...

bqpd commented 6 years ago

hmm good question. I feel like merging to library first is probably cleaner?

mjburton11 commented 6 years ago

so everything will just be broken for a while haha. That will motivate me to change everything quickly

bqpd commented 6 years ago

yeah...the good alternative is probably to have parallel development and bring back the old gpkit-models branch flag for individual repo PR tests only :p

mjburton11 commented 6 years ago

Yeah... that's not a bad idea. Let's talk some more about this tmr

nillst commented 6 years ago

do we want to talk about this after team meeting today?

mjburton11 commented 6 years ago

I have a meeting at 1 now so that won't work for me...

bqpd commented 6 years ago

heh, let's talk about it after your meeting then!

bqpd commented 6 years ago

I think we should add S bounds (for example, W_W <= 0.5*C_L*fs.V**2*S)

mjburton11 commented 6 years ago

@bqpd do you know why this is failing on jenkins, it runs for me locally

bqpd commented 6 years ago

test this please

bqpd commented 6 years ago

test models please

bqpd commented 6 years ago

@mjburton11, yup it's a bug...fixing...

bqpd commented 6 years ago

test this please

bqpd commented 6 years ago

fixed, still failing on cvxopt

mjburton11 commented 6 years ago

darn cvxopt...

bqpd commented 6 years ago

@mjburton11, it seems to be because agust[0] in GustL goes quite low...any chance you could adjust the fit to raise it above 1e-10? (it was going to 1e-11)

mjburton11 commented 6 years ago

Do you still want me to try and go higher than 1e-10?

bqpd commented 6 years ago

...yeah, it'd be good to have gplibrary working on cvxopt...

bqpd commented 6 years ago

That conversion you made wasn't the relevant 1e-10 to 1e-5 conversion to make, I think :p It's from the fit! Do you have cvxopt installed? It solves with cvxopt when it's in a Bounded

mjburton11 commented 6 years ago

I changed it to 1e-3 and it solves with cvxopt locally on my machine

bqpd commented 6 years ago

Looks like only wing test is failing now?

mjburton11 commented 6 years ago

it might be a problem with the tip and root variables in beam. I tried lowering them but the substitutions aren't working, take a look at the last commit

bqpd commented 6 years ago

I think I got it working, hold back a sec

bqpd commented 6 years ago

it's working!

bqpd commented 6 years ago

alright, I'm back to prepping for Aurora. All yours

mjburton11 commented 6 years ago

merged!!!

bqpd commented 6 years ago

yeehaw!

On Wed, Nov 15, 2017 at 5:32 PM, Michael Burton notifications@github.com wrote:

merged!!!

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/convexengineering/gplibrary/pull/137#issuecomment-344751311, or mute the thread https://github.com/notifications/unsubscribe-auth/ABagGGFcetJhHfcI24KMJLzP43v0FnGsks5s22aEgaJpZM4QbDWF .