s416094-mlworkshops/extract.py
Kuba Kowalczyk 2d28b5271b fix
2019-05-14 07:31:05 +02:00

55 lines
1.5 KiB
Python

import re
import fileinput
import csv
wars = []
corrects=0
with open("results.txt", "r", encoding='utf-8') as ins:
for line in ins:
if line.startswith('Scores'):
correct = int(re.findall('\d+', line)[0])
substitutions = int(re.findall('\d+', line)[1])
deletes = int(re.findall('\d+', line)[2])
inserts = int(re.findall('\d+', line)[3])
war = (substitutions+deletes+inserts)/(substitutions+deletes+correct)
wars.append(war)
if substitutions==0 and deletes==0 and inserts==0:
corrects = corrects + 1
srr = corrects / len(wars)
print(srr)
war_avg = sum(wars) / len(wars)
print(war_avg)
with open("wer_avg.txt", "w") as wer_avg_file:
wer_avg_file.write("WER avg: ")
wer_avg_file.write(str(war_avg))
with open("srr.txt", "w") as srr_file:
srr_file.write("SRR: ")
srr_file.write(str(srr))
with open("wer_srr.txt", "w") as wer_srr:
wer_srr.write("SRR: ")
wer_srr.write(str(srr))
wer_srr.write("\n")
wer_srr.write("WER avg: ")
wer_srr.write(str(war_avg))
counter = 0;
with open("wikiniews_results.tsv",'r', encoding="utf-8") as wik_in:
with open("result.tsv",'w', encoding="utf-8") as f_out:
writer = csv.writer(f_out, delimiter='\t', lineterminator='\n')
reader = csv.reader(wik_in, delimiter='\t')
result = []
for row in reader:
# add new column values
row.append(wars[counter])
result.append(row)
counter=counter+1
writer.writerows(result)