from ldp.gesture import LRB, GestureType
from ldp.data import Utterances, Subjects
from util.count import FeatureCounter
lrb = LRB()
gtype = GestureType()
subjects = Subjects()
utterances = Utterances()
count = FeatureCounter('Subject', 'Session', 'Project', 'Gesture')
P2 = set(subjects.project(2))
columns = 'subject, session, c_lrb, c_g_type'
filter = 'session in (1,2,3,4,5,8) and c_lrb != ""'
def pprint(args): print "\t".join(args)
for subj, sess, h, g in utterances(columns, filter, limit=''):
proj = 2 if subj in P2 else 3
H = lrb.valid_values(h.upper())
G = gtype.valid_values(g, subcodes=False)
for (h, g) in zip(H, G):
code = "{0}+{1}".format(h, g)
count(subj, sess, proj, code)
count.print_report('Gesture')
import re
from nlp import Tokenizer
from ldp.data import Utterances
from util.count import FeatureCounter
count = FeatureCounter('Subject', 'Session', 'Speaker', 'Word')
utterances = Utterances()
parse = Tokenizer()
words = [word.rstrip() for word in open('words.txt')]
rgx = re.compile(r'\b(?:' + '|'.join(words) + r')\b')
columns = 'subject, session, row, p_utts, c_utts'
where = 'session < 8'
# pretty-print with tab delims
def pprint(*args):
print '\t'.join(str(x) for x in args)
pprint(*'SUBJ SESS SPKR ROW UTT MATCH'.split(' ')) # header
for subj, sess, row, p, c in utterances(columns, where, project=2):
for spkr, utt in [('P', p), ('C', c)]:
matches = rgx.findall(utt)
for word in matches:
count(subj, sess, spkr, word)
if matches:
pprint(subj, sess, spkr, row, utt, ', '.join(matches))
print
count.print_report('Word')
import re
from ldp.data import Utterances
from nlp.lex import Tokenizer
from util.count import FeatureCounter
utts = Utterances()
tokenize = Tokenizer()
count = FeatureCounter('Subject', 'Session', 'Speaker')
columns = 'subject, session, key, c_utts, p_utts'
wordchar = re.compile(r'\w')
grandmother = re.compile(r'G')
father = re.compile(r'F|@')
for subj, sess, key, c, p in utts(columns,
where='session in ("11", "12")',
limit='',
project=2):
for spkr, utt in [('CHILD', c), ('MOTHER', p)]:
if spkr == 'MOTHER':
if father.search(key):
spkr = 'FATHER'
elif grandmother.search(key):
spkr = 'GRANDMOTHER'
for t in tokenize(utt):
if wordchar.search(t): count(subj, sess, spkr)
count.print_report('Speaker')
from ldp.gesture import LRB
from ldp.data import Utterances, Subjects
from util.count import FeatureCounter
lrb = LRB()
subjects = Subjects()
utterances = Utterances()
count = FeatureCounter('Subject', 'Session', 'Project', 'LRB')
P2 = set(subjects.project(2))
columns = 'subject, session, c_lrb, c_g_type'
filter = 'session in (1,2,3,4,5,8) and c_lrb != ""'
def pprint(args): print "\t".join(args)
for subj, sess, h, g in utterances(columns, filter, limit=''):
proj = 2 if subj in P2 else 3
for h in lrb.valid_values(h.upper()):
count(subj, sess, proj, h)
count.print_report('LRB')