User Feedback On Importer in Insight

[saloynton]

To arrange and confirm a timeframe with Johan that would be suitable for the trial of Insight
with the functionally of the importer.

• Meeting arranged for user feedback on new workflow for import for 10th Jan. Meeting arranged with Keria as Johan not back from holiday.

• Import Feedback 10th January 2011

Note the feedback has been arrange around the functionality that is currently implemented (this is set in italics).

What is currently implemented?
Selections:

Select a dataset
select images and import them in the dataset
select folders and import images (candidates) contained in the folders. No need to select images individually.
select a Project
select a folder and turn it into a dataset automatically. This dataset will be linked to the project. Images contained in the folder will be imported into the new datasets.

The feedback from this step was that there was not sufficient information to clearly confirm where the data is being imported (The ability of selecting where the data is stored from the outset might be slow to catch on). The comparison was made to previous import page shown in Figure 1. This was highlighted to provide the confirmation of where the data is/will be stored. And so might what be required to confirm to the user where the data is will be stored. The additional comment was made of having the link in the import activities window to allow a user go straight to the data.

• Figure 1

No container selected or not valid container selected (e.g. project selected and try to import and image)
If images are selected, a default date-based dataset is created e.g. 07-01-11 and the images will be imported in that dataset. Note that the dataset will NOT be linked to the project even if a project is selected.

If the images are going to be imported into a default folder then the default date-based dataset needs to be confirmed for the user. This was again suggested that the default date based folder is shown in the import box.

"If a folder is selected (and ticked to be turned into a dataset) the dataset will be created and not added to any project.

This can create a potential breakdown in the workflow of Project > Dataset. Keira highlighted how all her existing work is set out in the format of Project > Dataset and so she would not want Datasets without an association to a Project and so break the existing established structure.

Note: JM It is currently possible in the importer to import images into dataset NOT linked to a project. Offers more flexibility, I don't really see the breakdown here. Existing users can continue to do what they were doing.

Options:

Ability to turn the archived flag per image/folder.
Ability to set default value for archived or turn option enabled from config file (currently client side)

Currently, only option to tag the images. All images in the import queue will be tagged. If a folder is selected, all the images in the folder will be tagged. Possibility to add existing tags or to create new tags.

The first comment about tagging by Keria was still that she would find it more valuable to tag at the microscope. This is where she would find it more useful to add that level of information.
The workflow of adding tags and the retention of the last tag last used when adding a new data. Keria felt there was the potential to make a mistake of adding tags to data she had not intended. Because of this she would prefer the default to be rest to no tags.
The main piece of annotation information used by Keria for her work is filled out in the description. This was identified as the key aspect to have in the import. Keria identified that she may not use the additional fields of rating, comment, and attachment but she would not see any reason to have them excluded from the options.

Note JM: As discussed before demo, tags not kept. Description field per image or one field for all.
Understood preferences for tagging but won't see that happening any time soon.

When a file is imported:

For single image, a thumbnail is displayed in the row corresponding to the file. Mouse over to view a bigger thumbnail. Double-click on thumbnail or name to launch the viewer.
For a multi-images file, up to 3 thumbnails are displayed + number of images w/o displayed thumbnails . Mouse over to view a bigger thumbnail. Double-click on thumbnail to launch the viewer.
Icon in the data manager will change to indicating that images have been imported in the folder (dataset/project for now). No automatic refresh, might be too intrusive. Very easy to do refresh if it is better.

The refresh option was done automatically by Keria, an alternative of having a smaller refresh icon on the folder to indicate that the folder needs to refreshed. But this was already automatically part of Keria workflow so it was more difficult to judge.

When file failed to import:
name of the files with error message. Mouse over to view part of the stack trace. Option to submit the error with or w/o the file

Extra notes

The existing functionality of having to view and exclude the specific file types in a folder was key and so wanted to ensure that this would remain in the import functionality. This is shown in figure 2 in the bottom where can view R3D.dv files only.

Note JM: File filters not yet implemented. On the TODOs list.

• Figure 2

Additional positive comments

• The import button works more efficiently than before, Keria prefers the reduction in the step to import, where previously it took two steps to import ok > import. The move to ok and the import begins is a positive improvement.

• Positive feedback over the thumbnail viewing once images imported.

[jburel]

I have re-activated the importer functionality
see ​http://trac.openmicroscopy.org.uk/omero/changeset/7869/insight.

A first round of tests might be good.

[jburel]

For improvement see ​Insight:r7875

[jburel]

Referencing ticket #3415 has changed sprint.

[jburel]

tickets created following this observations: #3828, #3808, #3807, #3806, #3799