BLEMMA attribute

The BLEMMA attribute is very similar to LEMMA, but it is applied at the atom level of the disambiguation output and not the word level. This allows the user to match sub-tokens (atoms) that, at the word level, would be part of compounds, collocations or idiomatic expressions.

For example, this rule (using LEMMA):

SCOPE SENTENCE
{
    DOMAIN(dom1)
    {
        LEMMA("safety")
    }
}

when run against this text:

The manufacturing plant is in compliance with the fire safety standard.

doesn't find a match, because the term safety is part of the collocation lemma fire safety and is not recognized as a lemma by itself at the word level.

On the other hand, this rule (using BLEMMA):

SCOPE SENTENCE
{
    DOMAIN(dom1)
    {
        BLEMMA("safety")
    }
}

matches safety at the atom level, where it is not aggregated with lemma fire.