Feature Request

[geopast]

Hi folks, Brian Foo suggested posting a feature request for ScribeAPI here. What I would like is an option for ScribeAPI to handle single documents that have information on both sides (e.g. duplex) so that the UI displays both sides simultaneously. So basically juxtapose two images side-by-side for a single document at a time. Happy to discuss further if needed - neil@geopast.com. Many thanks for creating some really awesome software. Best wishes, Neil

[nonword]

Interesting. Is it essential that the two sides be shown simultaneously? Although Scribe supports multi-page documents in the Mark workflow, one has to use the thumbnail viewer to navigate between pages. We didn't consider a view that presented multiple images side by side simultaneously - in part because our materials so far have been pretty high detail and need to be presented at full width. But the request is noted. Do you have sample images that demonstrate the need for a side-by-side viewer?

[geopast]

Thanks for the reply. In my use case, the document structure (postcards) changes over time. In most examples some transcription data is located on one side and some on the other but in another example all the data is on the first side and not the other or vice versa. Also both sides often need to be viewed together to be able to make overall decisions about the data to be transcribed e.g marking. Please see below examples as requested and happy to discuss further.

Example 1 (data on both sides)

Example 2 (different data on both sides)

Example 3 (data on one side only - but first side provides context)

[nonword]

Thanks for posting sample images. Looks like an interesting project and I see how presenting the images separately wouldn't be ideal; Neither image should be transcribed in isolation. This would be a bit of work to build as it would mean modifying the SubjectSetViewer component to (optionally) present the two images of your subject-sets simultaneously (one on top of the other, I assume). Additionally, although Scribe allows you to view all of the subjects in a subject-set, the initial subject (image) presented is the one that Scribe obliges you to transcribe. It would take some work to modify Scribe to allow one to transcribe any of the images (subjects) in a presented subject set. This is a side-effect of some early assumptions we made that would take some explanation.. but suffice it to say the system really fundamentally targets single-image tasks. The three primary Scribe devs are buried in other things right now, so although there's interest in revising some of these things, it's not going to come from us soon, unfortunately.

Although it's a hack, you could build composite images from the two sides of the postcards and present them as single images (subjects).. If you saved the original image dimensions in the subject "meta_data" hash (or really, anywhere), it should be straightforward to transpose annotation region coordinates back to the original derivatives after the project has collected data..

Another thought is you could consider https://www.zooniverse.org/lab . I'm actually not sure if it handles presenting multiple images simultaneously, but it's more mature than Scribe in a couple other ways, so a project to consider if you haven't already.

[gr-]

If multi-page temporarily meets your needs, you can add multi-page support this way : put an extra field "set_key" in your group csv file, and provide each subject with a value that matches the set it belongs to (for instance "set_1"). Set values have no meaning. No requirement for specific development. Hope it can help.