Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

GUI #29

Merged
merged 7 commits into from
Oct 9, 2015
Merged

GUI #29

merged 7 commits into from
Oct 9, 2015

Conversation

casperdcl
Copy link
Sponsor Member

Addresses #28.
~350it/s overhead when enabled with gui=True.
Looks promising but still needs testing and tidying.

@casperdcl
Copy link
Sponsor Member Author

Any idea why py32 coverage/test build is failing on travis but all the rest passes?

@casperdcl
Copy link
Sponsor Member Author

erm... @hadim , @kmike any ideas on the weird Travis fail?

@hadim
Copy link
Contributor

hadim commented Oct 9, 2015

It's weird cuz only happenning on py32... can you try to relaunch Travis to see if it reproduces again ?

@lrq3000
Copy link
Member

lrq3000 commented Oct 9, 2015

I think I had the exact same bug on two of my other projects. This is due to Travis which has recently updated to coverage 4, and this weirdly produces a bug, but only with Python 3.2 (not with other below or greater Python versions).

To fix this, you just need to add a statement to load coverage < 4 in your travis config.

@lrq3000
Copy link
Member

lrq3000 commented Oct 9, 2015

Here's the Travis issue ticket:

travis-ci/travis-ci#4866

Proposed fix by ZuluPro (which worked for me), in .travis.yml, add this:

install:
[other instructions here...]
- pip install 'coverage<4'

@lrq3000
Copy link
Member

lrq3000 commented Oct 9, 2015

And BTW nice work here, I didn't try it for myself, but if the overhead is so small when using the GUI option, it's really amazing!

@casperdcl
Copy link
Sponsor Member Author

thx @lrq3000 ... if you have matplotlib you can easily checkout this branch and test it with the scripts in the examples folder.

@lrq3000
Copy link
Member

lrq3000 commented Oct 9, 2015

Ah nice to know that there's a simple example to use @casperdcl :) I'll try that when I have some time (I'm a little bit overwhelmed by projects currently, and that's an understatement :( ).

@hadim
Copy link
Contributor

hadim commented Oct 9, 2015

@casperdcl I guess you want to merge this before we made the transition.

@lrq3000
Copy link
Member

lrq3000 commented Oct 9, 2015

The workaround didn't work, coverage 4 is still used:

py32 installed: coverage==4.0,coveralls==1.0,docopt==0.6.2,nose==1.3.7,requests==2.8.0,tqdm==1.0,wheel==0.24.0

@lrq3000
Copy link
Member

lrq3000 commented Oct 9, 2015

I think that you need to put it before installing coveralls, at least that's what I did in my other project, because I guess else installing coveralls will automatically install the dependencies such as coverage.

Please try this:

install:

  • pip install 'coverage<4'
  • pip install tox coveralls

@kmike
Copy link
Contributor

kmike commented Oct 9, 2015

It should be changed in https://github.com/tqdm/tqdm/blob/master/tox.ini. Btw, I think we can remove coveralls from tox.ini

@lrq3000
Copy link
Member

lrq3000 commented Oct 9, 2015

@kmike ah maybe you're right, I didn't make a specific tox config in my other project, so the fix in .travis.yml was enough for me.

@casperdcl
Copy link
Sponsor Member Author

Ok. Third forced update later... should work now

@hadim
Copy link
Contributor

hadim commented Oct 9, 2015

Looks like it will pass :-)

@lrq3000
Copy link
Member

lrq3000 commented Oct 9, 2015

Seems all good :)

@hadim hadim closed this Oct 9, 2015
@hadim hadim reopened this Oct 9, 2015
@casperdcl
Copy link
Sponsor Member Author

wth that took a while -.- I was so freaked out by the bug but couldn't believe the latest version of coverage would break... On my scientific install (py27) coverage4 throws up a unicode-related error. not. cool.

casperdcl added a commit that referenced this pull request Oct 9, 2015
@casperdcl casperdcl merged commit 0341a41 into master Oct 9, 2015
@casperdcl casperdcl deleted the gui branch October 9, 2015 21:40
@hadim
Copy link
Contributor

hadim commented Oct 9, 2015

Nice work dude !

@casperdcl
Copy link
Sponsor Member Author

i also recently found a bug in matlab's cuda-related linux libraries. that was even more not cool. ;) hope you guys can work out a way of using this in things like gooey (haven't looked at its source)

@casperdcl casperdcl mentioned this pull request Oct 9, 2015
Closed
@lrq3000
Copy link
Member

lrq3000 commented Oct 9, 2015

Yes it's very not cool when the libraries are broken :(

@lrq3000
Copy link
Member

lrq3000 commented Nov 30, 2015

About Gooey integration, here's a starting point:

chriskiehl/Gooey#120

The problem is that we'd need to directly modify Gooey in order to support integration of tqdm. Another way is to make Gooey able to callback to get progress updates, but this seems problematic according to the main developper:

chriskiehl/Gooey#60

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

4 participants