-
Notifications
You must be signed in to change notification settings - Fork 54
No, at this time we've focused on building a secure and lightweight centralized annotation system rather than a distributed model. In addition, Anafora requires a Linux environment to run. That said, although it will require some technical skill, there's no reason that you can't install Anafora on your linux machine (or in a virtual machine) and then work directly with that instance.
No. Anafora requires communication with the central server to save and open annotations, so without any internet connection, you will be unable to begin or save annotation.
However, Anafora does cache annotations locally, allowing you to continue working in the same document/window during an internet outage. If you've made annotations while no internet was available, when you reconnect and reload, you'll be prompted to save your locally cached annotations to the server. This means that Anafora can work well with a slow, intermittent, or tethered internet connection, well beyond what is feasible with X11 tunneling or completely synchronous web-based tools, but it still does require an internet connection to perform annotations.
Anafora is designed to read AnaforaXML only, however, the format is similar to that used by many other tools, and we have built scripts to convert files from other annotation software (e.g. Protégé/Knowtator) into AnaforaXML in the past.
Yes! Anafora has been designed to show overlapping spans with a double-bordered gray box. Clicking this box will bring up a menu allowing you to choose your desired annotation.
Anafora reads from plaintext files, and does no parsing or processing of the data within. So, although the plaintext tree files can be read into the program, they will be displayed as they are, and annotators will not be able to interact with nodes or labels beyond selecting spans.
Not at the present time. That said, if this is a requirement for your project, the best approach is to add a Text input box as a property of the relation/entity, and then simply copy and paste the relation's unique ID into that box. This way, the link is available in the raw data, although the tool does not check for removed relations, etc.