Unending spinning of dataset images #807
Conversation
Failed request and incorrect name structure will cause forever loading
If no file suffix, the name will be undefined
The thumbnail response can be empty string. When this happens, the default thumbnail will not be used.
When duplicate images exist, except the first image request, all others will have no responses.
1 failed test on run #60 ↗︎
Details:
|
|||||||||||||||||||||
| Test | Artifacts | |
|---|---|---|
| Homepage > Footer |
Screenshots
Video
|
|
Review all test suite changes for PR #807 ↗︎
There was a problem hiding this comment.
I do not think we want to remove the check for an image title. When applying these changes, clicking on the first image in the gallery of dataset 205 takes me to: /datasets/file/205/1?path=files/derivative/sub-7/sub-7-sam-1/Dog7%2020X%20Caspase-3%20(1).jp2 but does not show an associated biolucida viewer. It looks like the name of the image should be: Dog 20X Caspase-3 (1).jp2
| @@ -104,7 +104,7 @@ export default { | |||
| }, | |||
| computed: { | |||
| isReady() { | |||
| return this.data.title && (this.thumbnail || this.useDefaultImg) && (this.data.link || this.data.userData) | |||
There was a problem hiding this comment.
why are we removing the check for the title? Images should have a title associated with them
There was a problem hiding this comment.
Due to the failed biolucida requests, the file name will be undefined. The main reason for removing the title check is to avoid unending spinning caused by this. This may cause some inconvenience for people who are doing the portal QC. I updated the existing biolucida testing which will run through all the datasets and help to generate a report of related issues. You will be able to find an example report in this ticket.
| @@ -808,6 +811,9 @@ export default { | |||
| const name = response.name | |||
| if (name) { | |||
| this.$set(item, 'title', name.substring(0, name.lastIndexOf('.'))) | |||
| if (name.lastIndexOf('.') === -1) { | |||
There was a problem hiding this comment.
this is happening because we are removing the check for the title. I believe every image should have a title. If an image does not have one, we need to figure out why biolucida is not returning one
There was a problem hiding this comment.
The change here is because of the file name structure. The original code's default file name format should end with a suffix, e.g. .jp2 in Dog 20X Caspase-3 (1).jp2. However, in some datasets, the file name will not have the suffix (some other reason caused this). name.substring(0, name.lastIndexOf('.')) will be an empty string. For example, some image files in dataset 178 (the issue is caught in the above example report).
I understand that removing the check for an image title is not an ideal solution. The main reason why the name does not show in the gallery card and biolucida viewer doesn't show the image for dataset 205 here is because |
Do we know why we are getting those errors from Biolucida? My main concern is that the image being in the gallery implies that it has a viewer associated with it, but when they click on it to view it, it will not show up |
|
@egauzens I remember Maci mentioning something about some older datasets using a "local" path for some of the images instead of the correct dataset path (if I remember correctly). |
Description
All the changes are focused on fixing the unending spinning issue mentioned in ticket Investigate unending spinning of dataset images for all datasets
Type of change
How Has This Been Tested?
I have done manual testing. Accessed all the dataset galleries and checked whether all spinning icons were gone.
Checklist: