def _extract(self, stim):
words = [w.text for w in stim]
pos = nltk.pos_tag(words)
if len(words) != len(pos):
raise PliersError(
"The number of words in the ComplexTextStim does not match "
"the number of tagged words returned by nltk's part-of-speech"
" tagger.")
data = {}
onsets = []
durations = []
for i, w in enumerate(stim):
p = pos[i][1]
if p not in data:
data[p] = [0] * len(words)
data[p][i] += 1
onsets.append(w.onset)
durations.append(w.duration)
return ExtractorResult(np.array(list(data.values())).transpose(),
stim,
self,
features=list(data.keys()),
onsets=onsets,
durations=durations)
def _extract(self, stims):
words = [w.text for w in stims]
pos = nltk.pos_tag(words)
if len(words) != len(pos):
raise PliersError(
"The number of words does not match the number of tagged words"
"returned by nltk's part-of-speech tagger.")
results = []
tagset = nltk.data.load('help/tagsets/upenn_tagset.pickle').keys()
for i, s in enumerate(stims):
pos_vector = dict.fromkeys(tagset, 0)
pos_vector[pos[i][1]] = 1
values = [list(pos_vector.values())]
results.append(
ExtractorResult(values,
s,
self,
features=list(pos_vector.keys())))
return results