def features(inputDir):
features = []
labels = []
#getting data from given directory as list
csv_data = list(tools.get_data(inputDir))
for dialogue in csv_data:
for i in range(len(dialogue)):
features.append(sentFeatures(dialogue, i))
labels.append(sentLabels(dialogue[i]))
return features, labels, csv_data
# -*- coding: utf-8 -*- """ Created on Sat Apr 11 12:45:35 2020 @author: Likhita Suresh """ import pycrfsuite import hw2_corpus_tools as tools import os path = os.getcwd() + "\\trainSmall" print(path) csv_data = list(tools.get_data(path)) print(csv_data)
import pycrfsuite
import hw2_corpus_tools as tools
import os, sys
#getting path to train date - appending train data file to current path
#path = os.getcwd()+"\\trainSmall"
inputDir = sys.argv[1]
testDir = sys.argv[2]
outputFile = sys.argv[3]
#print (path)
#print (csv_data[0][0][2][0].token)
csv_data = list(tools.get_data(inputDir))
#feature for every token
def tokenFeatures(wordList):
tokens = []
letter_pairs = ["th", "he", "in"]
for word in wordList:
tokens.append('TOKEN_' + word.token)
for pair in letter_pairs:
if pair in word:
tokens.append('TOKEN_' + pair)
for i in range(len(wordList) - 1):
tokens.append('TOKEN_' + wordList[i].token + wordList[i + 1].token)
return tokens