Last active
April 8, 2018 07:40
-
-
Save yanshengjia/234c08088de3fcb0151e25734d46458a to your computer and use it in GitHub Desktop.
Convert ARPA format language model to JSON
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
# !/usr/bin/python | |
# -*- coding:utf-8 -*- | |
# @author: Shengjia Yan | |
# @date: 2017-11-29 Wednesday | |
# @email: i@yanshengjia.com | |
import json | |
import codecs | |
class ARPAParser: | |
def __init__(self, parent=None): | |
self.parent = parent | |
self.n_max = 3 | |
self.lm_path = '../data/lm/corpus.lm' | |
self.arpa_file = open(self.lm_path, 'r') | |
self.info = '' | |
self.lm = {} | |
self.output_path = '../data/lm/lm.json' | |
# the n in ngrams up to 3 | |
def extract(self): | |
n = 1 | |
current = 1 | |
counter = 0 | |
for line in self.arpa_file.readlines(): | |
counter += 1 # line number starts from 1 | |
line = line.strip() # line: probability\tngrams\tbackoff_weight, ngrams:[word1 word2 word3] | |
if line == '' or line=='\\data\\' or line=='\\end\\': | |
continue | |
if counter >= 3 and counter <= (self.n_max + 2): | |
self.info += line + '\n' | |
continue | |
if line == "\\" + str(n) + "-grams:": | |
current = n | |
n += 1 | |
continue | |
list = line.split('\t') | |
dict = {} | |
dict['log_p'] = list[0] | |
if len(list) == 3: # backoff weight exist | |
dict['log_bw'] = list[2] | |
else: # backoff weight doesn't exist (equals to 1, log10(bw)==0) | |
dict['log_bw'] = 0.0 | |
self.lm[list[1]] = dict | |
def saveLM(self): | |
with open(self.output_path, 'w') as output_file: | |
lm = json.dumps(self.lm, ensure_ascii=False) | |
output_file.write(lm) | |
def main(): | |
parser = ARPAParser() | |
parser.extract() | |
parser.saveLM() | |
if __name__ == "__main__": | |
main() |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment