input data format

[MikeWangWZHL]

could you elaborate the format of corpus_index_*.txt, I load it with pickle and see a struture of lists, [ "gives", "give", "obj", "surplus", "city", null ], what are the meanings of these fields? and what is the interger mean? like this '4' here [ [..], [...], 4 ],

[eecrazy]

Like in the above training example, it is consisted of 8 context events, and 5 candidate events. The interger 4 means the index of the correct candidate event (starting from 0, so 4 means the last candidate event among the 5 is the correct answer).

Each event is a list structure with 6 elements, like ('gives', 'give', 'obj', 'surplus', 'city', None), they are: verb, lemmatized verb, dependency relation between the verb and the shared entity 'city', subject, object, indirect object to the verb.