Closed pcm32 closed 4 years ago
Also, I think that I'm using some dplyr constructs, is there a way to specify a library call for this particular function (I could call it inside the function, but I wonder if you are using some other way). dplyr is part of the conda package for monocle3-cli.
I have removed as much as possible redundant travis.org and travis.com (to leave only the last one for PRs), to reduce our carbon footprint :-).
Many thanks for the PR @pcm32! I can do the rest. Just to confirm, you want it to always write a full marker table, an optional top marker table and an optional marker-cluster plot, is that right?
Yes, sounds right, unless that you think that top markers output should be the relevant one and the full one the optional. Both ways work for me. Thanks @nh3! I can do release and so on later if you want.
Hehehe, apparently I was a bit off on how to implement this.
The inner workings aren't straightforward and I didn't write documentation, it took me sometime too. In the end, I chose to save top markers as a mandatory output.
I think this is now ready for review.
Being the author of the PR, I cannot review it. @pinin4fjords or @a-solovyev12 could you please take a look? Thanks @nh3 for the changes, it looks good to me at least.
Thank you @pinin4fjords and @pcm32. Is there a way to make travis CI run the other tests? Otherwise it seems it wouldn't allow us to merge without force.
Try pushing a dummy spacing change or something @nh3 ? Has worked in previous situations where Travis got confused.
Thanks @pcm32! Actually I wonder how did you do this. Was always a bit tired to see the same tests run 2-3 times...
I have removed as much as possible redundant travis.org and travis.com (to leave only the last one for PRs), to reduce our carbon footprint :-).
I went to the travis-ci.org and .com dashboard, once I was sure that this was active in the travis-ci.com side, I turned off the travis-ci.org ones. Then I only activated the PR one on the travis-ci.com (turned off the branch one). But we can use branch if you want instead of PR, I just wouldn't use both.
This PR adds the markers test and markers filtering functionality to produce top markers in monocle. There are a few things that I'm not sure how they work in the current layout of the package, for instance whether I need to do something to inhering the output_plot and output_plot_format, or how to deal with the case where the method can output more than a single table (looking to inherit the output_table option).
@nh3 could you advise on what I'm missing to complete this? Thanks!