MediumNLPnlp-114

Triple Extraction Pattern

Problem

Extract simple subject-verb-object triples from a token list. For every position whose token is in verbs, emit [previous token, verb, next token] when both neighbors exist.

Verbs at the first or last position are skipped, and a set of verbs with no matches yields an empty list.

Hint

Iterate the interior positions from 1 to len(tokens) - 2.

Test cases

case 1

in: [["cats","eat","fish"],["eat"]]

expected: [["cats","eat","fish"]]

case 2

in: [["eat","fish"],["eat"]]

expected: []

case 3

in: [["cats","eat","fish","run"],["eat","run"]]

expected: [["cats","eat","fish"]]

case 4

in: [["a","b","c"],[]]

expected: []

case 5

in: [["sleep"],["sleep"]]

expected: []

Starter codePython

def triple_extraction_pattern(tokens, verbs):
    # Your code here
    pass

5 test cases · run your code against them in the editor