Embedded brat visualization

[DDDurrant]

I am working with brat visualization as demonstrated here http://brat.nlplab.org/embed.html. If this isn't the right place to post this, then please point me in the right direction.

The issue is when a user inputs something that resembles an html self closing tag <hr/> it will convert it to <hr></hr> It doesn't even need to actually match an existing tag. It only needs to have "<" followed by any series of caricatures, then end with "/>". Patterns I have tried "" "<a:a/>" "<a:a a/>". they each come out as such "", "<a:a></a:a>", "<a:a a></a:a>". I have tested this behavior in chrome, firefox, and safari.

This wouldn't be a big deal, but it shifts all the annotations off by however many characters are added to the text. I didn't read anywhere that documents this behavior. I have noticed that when I try it on https://corenlp.run/ that also uses brat visualization, it doesn't follow this same behavior. Is there some setting that I need to add, or do I need to manipulate the text so that it doesn't see it as html?