-
Notifications
You must be signed in to change notification settings - Fork 1k
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
Various Improvements for the Rule Builder Widget #5823
Conversation
3bf2182
to
90b74b3
Compare
Previously these were only available in preview mode.
- Replace "Okay" with "Apply" on button labels. - Create CSS style for buttons in rule editor part of the component. - More consistency - right-align all the buttons, more consistent spacing. - Make primary action that button blue to make the buttons look a little bland.
Leverages the [pyre-to-regexp](https://github.com/jmchilton/pyre-to-regexp) that I recently published for this purpose.
Hi @jmchilton I had a couple of other small thoughts for the Rule Builder that I wondered if I should add here (not sure if you've already done something on these)...
|
"Collection Name" is definitely not mandatory, that only needs to be set if multiple collections are being created at the same time. I'd expect typically one collection would be built at a time - though I understand control/condition situations are currently generally handled as separate collections so I may be wrong about this. There is some discussion of this here. If a list identifier is not set the build button is not supposed to be clickable, this a bug and it will fix it. I will make it bold for sure though. In some scenarios this is pre-populated. Like if you select multiple datasets in the history panel and then go to create a collection from them - the list identifier is pre-populated as the name. For general tabular data I'm more skeptical about pre-picking a column but I should find a way to put it in the list in a way where it is more clear it is a decision the user MUST make in addition to just making it bold. Thanks for the suggestion!
I'm hesitant to combine URLs and file paths into one column definition. I understand the Galaxy uploader works that way and most of the time it is fine - but the backend for that is terrible hacks all the way down and prevents you from doing things like uploading tables that start with URLs for instance. I'd like to at least start with this trying to be more structured. I think a more conservative approach I'd rather start with is searching the supplied data and pre-populating a URL column definition if it looks like there is a single column with a single URL in it. Maybe I could customize that message in that case also - so it says "We have found a URL we are assuming contains the target data, click here or remove the column definition to undo this if this is incorrect." The widget also pre-assigns FTP directory if you select that option in the upload page - so we could do something similar in that situation. So you'd only ever see that message if you are uploading a tabular data file without an obvious target URL - maybe it has multiple URLs or maybe it is a sample sheet with FTP paths embedded somehow. These would be more niche, more advanced situations.
I see, that is confusing. Hmm... the problem is that the paired indicator also gets translated underneath to an identifier - so this is a particular kind of identifier. Let me think about this though because you are definitely correct this is confusing. There is some help text on mouse over of the column definition on the outer page that says "This should be a short description of the replicate, sample name, condition, etc... that describes each level of the list structure." But that is a very subtle thing. You've clearly put a lot thought into this and I really appreciate all the comments. This is great feedback that I think is going to make the widget much more usable. Keep it coming! |
Thanks @jmchilton. |
Review individuals commits for additional details - I've tried to keep them atomic and focused.