Open spyysalo opened 12 years ago
Note that there's some preliminary comments on possible format in static/standoff.html
(commented out for now)
The above use case is an example for a document-level field. We would like to suggest to make this more generic and allow a user to define document level fields. A document level field would be defined by a name and its type. Additionally it might be necessary to allow to assign a default value to it (e.g. annotator name) and to define it as a read-only field.
These fields can be used for multiple purposes, e.g. tracking the annotator name on a document level, showing the detected document language, showing the document type, tracking the annotation status (or progress, as mentioned above), etc.
In combination with the search feature the document-level fields could be an effective tool to easily get a sub-set of documents which a user needs to work on.
For example:
The above use cases can be done without loosing anytime to locate a document.
Is there a chance to get this feature into the next release?
To train a statistical model on a partial annotated brat corpus it is necessary to distinguish between already fully annotated documents and non or only partial annotated documents.
Suggestion from user: