Possibility to tag documents and document parts
This is a story specific to the LECTAUREP project.
-
It would be a good thing to be able to create tags associated to documents and/or documentparts.
Ideally, the creation of tags would follow the same model as Gitlab or Github, where you can freely choose the name of the tag (and a color?). If you have in mind how the tagging works in Transkribus, it is very rigid: you can't tag a collection or a document, you can only tag a documentpart (page) with only 4 pre-defined tags: 'New', 'In Progress', 'Final' and 'Ground Truth'. It is not enough to cover the needs of a project because there are many stages within the "in progress" step (for example: automatically segmented but not reviewed, or segmentation corrected or manually generated but transcription automatically made and not reviewed, etc). If a team/project can create their own tags in stead of using pre-defined tags, they can define their own workflow. Tagging documents would serve a different purpose than tagging at documentparts level, but both would be very useful to navigate within the transcriptions. -
It would then be very useful to have a tool to select only documentparts tagged a certain way (or all the dp without a specific tag). Think of a situation where you have 200 images in your document: if you can choose, for example, "only the images tagged with 'ready'" than you can train a model without having to unselect 100 images...
-
It would be very helpful to have the same system at document level: it would help navigating within the list of documents, especially when you start to have dozens of pages of documents.
Voir :