[WIP] CSV datatype#191
Conversation
There was a problem hiding this comment.
Overall this seems pretty solid to me - can you create a tabular file on master (which won't have the delimiter metadata defined), then checkout this branch and verify these column providers still work with the old datasets. I think there are other places we are defensive for older datasets and newer metadata - I just wanted to make sure you have confirmed that isn't needed here.
There was a problem hiding this comment.
Chatted with @blankenberg about this - it seems that this is what no_value= does in the MetadataElement definition about. So I guess the no_value for delimiter should be \t.
There was a problem hiding this comment.
Hey, I just checked - to me it looks okay (file: csv-tab-commentheader-csv)
|
Yeah - I just check it and somehow it is getting '\t' as the delimiter even though the So does the scatterplot work for you? I don't get any visualizations enabled except IPython if I upload a |
|
Haven't heard any news on this for a couple weeks - any chance for status update? I am going to merge #204 soon unless there is a strong objection which will introduce more conflicts to this. |
|
Yeah, the notification slipped my mind. The scatterplots works for me, maybe it's not in your configuration? The thing is, that everything in |
Conflicts: lib/galaxy/datatypes/tabular.py
For me it seems like the viz plugins only detect the datatype if the plugin datatype (e.g. ``tabular.CSV``) is explicitly instantiated by some extension directly. In your sample config you have ``subclass="True"`` so ``tabular.CSV`` wasn't working. This is probably a bug with the datatype code - but nonetheless I think dropping subclass from the sample is an okay workaround - because it does seem to be a direct mapping.
|
Figured out the viz problem - I think a small tweak to the sample config fixes it. I have opened a PR against your branch that resolves this and merges in the latest dev so the PR reapplies cleanly. |
Updates for PR galaxyproject#191 (checked the 508194f)
|
Thank you for the contribution @vavrusa and @jmchilton for the review. |
|
🎉 thanks! |
* update outputs * add default notebook * add an inputs folder
* update outputs * add default notebook * add an inputs folder
* update outputs * add default notebook * add an inputs folder
* update outputs * add default notebook * add an inputs folder
5 participants
This is another attempt at the CSV type support, I'd like to know your thoughts before I move forward with it.
Fixes
stats/gsummaryworks with RPy2 classic mode, as the original RPy is pretty much deadHow to test
I have attached some example files at https://gist.github.com/vavrusa/f8bb8d059c6c96dc7585
to test it, the one with the first line as a header should be sniffed correctly when uploading, the rest should be correctly displayed when saved as
tabular. Visualization for all should be correct.