Introducing utils.helpers for miscellaneous helper functions

bin/benchmark_nc.py

@@ -9,7 +9,7 @@ import sys
 # To use util.tc
 sys.path.append(os.path.abspath(os.path.dirname(os.path.dirname(sys.argv[0]))))
 import util.taskcluster as tcu
-from util.benchmark import keep_only_digits
+from util.helpers import keep_only_digits
 
 import paramiko
 import argparse
@@ -171,8 +171,8 @@ def all_files(models=[]):
         assert len(fa) == len(fb)
         assert len(fa) == 1
 
-        fa = keep_only_digits(fa[0])
-        fb = keep_only_digits(fb[0])
+        fa = int(keep_only_digits(fa[0]))
+        fb = int(keep_only_digits(fb[0]))
 
         if fa < fb:
             return -1

bin/benchmark_plotter.py

@@ -8,8 +8,7 @@ import sys
 
 # To use util.tc
 sys.path.append(os.path.abspath(os.path.dirname(os.path.dirname(sys.argv[0]))))
-import util.taskcluster as tcu
-from util.benchmark import keep_only_digits
+from util.helpers import keep_only_digits
 
 import argparse
 import numpy
@@ -35,7 +34,7 @@ def reduce_filename(f):
     '''
 
     f = os.path.basename(f).split('.')
-    return keep_only_digits(f[-3])
+    return int(keep_only_digits(f[-3]))
 
 def ingest_csv(datasets=None, range=None):
     existing_files = filter(lambda x: os.path.isfile(x[1]), datasets)

bin/import_cv2.py

@@ -27,7 +27,7 @@ from multiprocessing.dummy import Pool
 from multiprocessing import cpu_count
 from util.downloader import SIMPLE_BAR
 from util.text import Alphabet, validate_label
-from util.feeding import secs_to_hours
+from util.helpers import secs_to_hours
 
 
 FIELDNAMES = ['wav_filename', 'wav_filesize', 'transcript']

bin/import_lingua_libre.py

@@ -28,7 +28,7 @@ from glob import glob
 
 from util.downloader import maybe_download
 from util.text import Alphabet, validate_label
-from util.feeding import secs_to_hours
+from util.helpers import secs_to_hours
 
 FIELDNAMES = ['wav_filename', 'wav_filesize', 'transcript']
 SAMPLE_RATE = 16000

bin/import_m-ailabs.py

@@ -26,7 +26,7 @@ from glob import glob
 
 from util.downloader import maybe_download
 from util.text import Alphabet, validate_label
-from util.feeding import secs_to_hours
+from util.helpers import secs_to_hours
 
 FIELDNAMES = ['wav_filename', 'wav_filesize', 'transcript']
 SAMPLE_RATE = 16000

bin/import_slr57.py

@@ -29,7 +29,7 @@ from glob import glob
 
 from util.downloader import maybe_download
 from util.text import Alphabet, validate_label
-from util.feeding import secs_to_hours
+from util.helpers import secs_to_hours
 
 FIELDNAMES = ['wav_filename', 'wav_filesize', 'transcript']
 SAMPLE_RATE = 16000

bin/import_ts.py

@@ -27,7 +27,7 @@ from os import path
 
 from util.downloader import maybe_download
 from util.text import validate_label
-from util.feeding import secs_to_hours
+from util.helpers import secs_to_hours
 
 FIELDNAMES = ['wav_filename', 'wav_filesize', 'transcript']
 SAMPLE_RATE = 16000

stats.py

@@ -3,7 +3,8 @@
 import argparse
 import os
 
-from util.feeding import read_csvs, secs_to_hours
+from util.helpers import secs_to_hours
+from util.feeding import read_csvs
 
 def main():
     parser = argparse.ArgumentParser()

util/benchmark.py

@@ -1,15 +0,0 @@
-#!/usr/bin/env python
-# -*- coding: utf-8 -*-
-
-from __future__ import absolute_import, division, print_function
-
-def keep_only_digits(s):
-    r'''
-    local helper to just keep digits
-    '''
-    fs = ''
-    for c in s:
-        if c.isdigit():
-            fs += c
-
-    return int(fs)

util/helpers.py

@@ -0,0 +1,9 @@
+
+def keep_only_digits(txt):
+    return ''.join(filter(lambda c: c.isdigit(), txt))
+
+
+def secs_to_hours(secs):
+    hours, remainder = divmod(secs, 3600)
+    minutes, seconds = divmod(remainder, 60)
+    return '%d:%02d:%02d' % (hours, minutes, seconds)