BUG: stream of feedback from @mih and comments by @jshuenis

[jsheunis]

1

"trying to click on something (e.g. keywords), and nothing happens".
I do not have a good intuition of what to make of that. Best thing I can come up with is that the visual presentation of a keyword does somewhat resemble that of a button, and this might have triggered the expectation of an action when clicking on it.

So: keywords are shown for a dataset, and on the summary cards of a dataset's subdatasets. The keywords on subdatasets are clickable. The keywords on top for the dataset itself are not clickable. This is intentional, because the dataset keywords is a property of the dataset itself, and not of its subdatasets. The keyword search below takes all subdataset keywords into account.
But I agree, it could be expected that the dataset keywords are clickable due to their shape, and then it's a UX issue if they aren't. I could probably make them appear different visually.

2

IIRC javascript should have access to the page URL that the current catalog page was reached from (in terms of browser navigation) -- i.e. the standard "back button" -- I often find myself looking for the in page "back" navigation, and there is none (only to-top, and down)

Investigate whether to implement an in-page back button that serves the same purpose as browser back-button but could make for a more expected/intuitive UX.

3

Some of the keywords are funky: https://psychoinformatics-de.github.io/sfb1451-projects-catalog/#/dataset/a808f9e1-d638-4771-ae4c-ff21da29151a/272e2844cd7c58dda8dfe288e7831ad2fa9c34b6 has "suffix" and "subject"

this comes from the bids_dataset extractor+translator, and they can be handled by those to either be included or excluded. I think specifically the keywords pointed out here can be excluded from the bids_dataset output, since they do not add any useful functionality

4

the preview for a catalog dataset landing page generated by matrix is just "CRC1451 Catalog" -- this could/should have essential info on the actual dataset

Good point. For the general case, we will need to somehow get info from the superdataset into the index.html <meta> tag. Related issue here: #285.

For the SFB we can enter it manually: sfb1451/metadata-catalog#16

5

when I search at the catalog entrypoint for a keyword (like that "suffix"), the catalog knows about it (autocompletes)

But if I hit return, I get no hits -- because none of the immediate subdatasets matches. This is nevertheless a UX issue. It could be found, it knows about it.

This is unexpected behaviour. The keyword search space is populated by all keywords of immediate subdatasets, therefore if a keyword is known about (from the keyword search field), then there should be an immediate subdataset that gets matched. If not, this would be some bug. To be investigated.

6

When I click on the copy button here:

I get tons of whitespace

(three lines, plus indentation)

We need to strip whitespace after copying the string and before pasting it to the clipboard.

7

Citation contains HTML markup

Need to filter out markup received from crossref, or perhaps the API has functionality to exclude that?

8

This is now the most frequent property in the catalog. If it is easy to turn off if == zero, I think we should consider it.

Keep the remaining zero in the corresponding tab: Subdatasets (0).

9

Binder functionality:

• Update repo/notebook name: parameter-test is confusing
• Use bash kernel notebook, not python
• Check if useful progress reporting can be shown

10

datalad clone https://gin.g-node.org/Aswendt_Lab/2023_Goetz_RecoveryGenes.git
should be:
datalad clone https://gin.g-node.org/Aswendt_Lab/2023_Goetz_RecoveryGenes

Should add another step to GIN url mangling in the catalog code to remove the trailing .git

[jsheunis]

This issue was divided into separate issues which were subsequently addresses by PRs