Skip to content

Instantly share code, notes, and snippets.

@putheakhem
Created July 16, 2019 15:10
Show Gist options
  • Save putheakhem/8b84ca603f97d6a1a55bd14f560590ca to your computer and use it in GitHub Desktop.
Save putheakhem/8b84ca603f97d6a1a55bd14f560590ca to your computer and use it in GitHub Desktop.
Use for validation accuracy Neural Machine Translation
import sys
import codecs
import os
import math
import operator
import json
from functools import reduce
def fetch_data(cand, ref):
""" Store each reference and candidate sentences as a list """
references = []
if '.txt' in ref:
reference_file = codecs.open(ref, 'r', 'utf-8')
references.append(reference_file.readlines())
else:
for root, dirs, files in os.walk(ref):
for f in files:
reference_file = codecs.open(os.path.join(root, f), 'r', 'utf-8')
references.append(reference_file.readlines())
candidate_file = codecs.open(cand, 'r', 'utf-8')
candidate = candidate_file.readlines()
return candidate, references
def count_ngram(candidate, references, n):
clipped_count = 0
count = 0
r = 0
c = 0
for si in range(len(candidate)):
# Calculate precision for each sentence
ref_counts = []
ref_lengths = []
# Build dictionary of ngram counts
for reference in references:
ref_sentence = reference[si]
ngram_d = {}
words = ref_sentence.strip().split()
ref_lengths.append(len(words))
limits = len(words) - n + 1
# loop through the sentance consider the ngram length
for i in range(limits):
ngram = ' '.join(words[i:i+n]).lower()
if ngram in list(ngram_d.keys()):
ngram_d[ngram] += 1
else:
ngram_d[ngram] = 1
ref_counts.append(ngram_d)
# candidate
cand_sentence = candidate[si]
cand_dict = {}
words = cand_sentence.strip().split()
limits = len(words) - n + 1
for i in range(0, limits):
ngram = ' '.join(words[i:i + n]).lower()
if ngram in cand_dict:
cand_dict[ngram] += 1
else:
cand_dict[ngram] = 1
clipped_count += clip_count(cand_dict, ref_counts)
count += limits
r += best_length_match(ref_lengths, len(words))
c += len(words)
if clipped_count == 0:
pr = 0
else:
pr = float(clipped_count) / count
bp = brevity_penalty(c, r)
return pr, bp
def clip_count(cand_d, ref_ds):
"""Count the clip count for each ngram considering all references"""
count = 0
for m in list(cand_d.keys()):
m_w = cand_d[m]
m_max = 0
for ref in ref_ds:
if m in ref:
m_max = max(m_max, ref[m])
m_w = min(m_w, m_max)
count += m_w
return count
def best_length_match(ref_l, cand_l):
"""Find the closest length of reference to that of candidate"""
least_diff = abs(cand_l-ref_l[0])
best = ref_l[0]
for ref in ref_l:
if abs(cand_l-ref) < least_diff:
least_diff = abs(cand_l-ref)
best = ref
return best
def brevity_penalty(c, r):
if c > r:
bp = 1
else:
bp = math.exp(1-(float(r)/c))
return bp
def geometric_mean(precisions):
return (reduce(operator.mul, precisions)) ** (1.0 / len(precisions))
def BLEU(candidate, references):
precisions = []
for i in range(4):
pr, bp = count_ngram(candidate, references, i+1)
precisions.append(pr)
bleu = geometric_mean(precisions) * bp
return bleu
if __name__ == "__main__":
candidate, references = fetch_data(sys.argv[1], sys.argv[2])
bleu = BLEU(candidate, references)
print(bleu)
out = open('bleu_out.txt', 'w')
out.write(str(bleu))
out.close()
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment