When Evita applies, it will not create an event if there already was one on that same location. But it would be nice to allow a location to be tagged as an event by two separate components.
This is now not possible because when adding an event we check whether there is one already there. That check is needed because some chunks are considered more than once because of peculiarities in the Evita code. Either make sure that Evita does not consider the same locations twice or make the check dependent on the source of the event, which means adding a source attribute which should perhaps be added to each tag. Note that this is hard with the lex tag since the tag itself comes from the tokenizer but the pos attribute from the tagger, we may just credit the preprocessor for it, or maybe we just add the source attribute to TimeML tags.
When Evita applies, it will not create an event if there already was one on that same location. But it would be nice to allow a location to be tagged as an event by two separate components.
This is now not possible because when adding an event we check whether there is one already there. That check is needed because some chunks are considered more than once because of peculiarities in the Evita code. Either make sure that Evita does not consider the same locations twice or make the check dependent on the source of the event, which means adding a source attribute which should perhaps be added to each tag. Note that this is hard with the lex tag since the tag itself comes from the tokenizer but the pos attribute from the tagger, we may just credit the preprocessor for it, or maybe we just add the source attribute to TimeML tags.