def test_put_file_contents_str(self):
s, fpath = "123", os.path.join(self.resources, "foo.bar")
self.assertTrue(not os.path.isfile(fpath))
self.assertTrue(Doraemon.put_file_contents(fpath, s))
self.assertTrue(os.path.isfile(fpath))
self.assertEqual(1, len(Doraemon.get_file_contents(fpath)))
self.assertEqual(s, Doraemon.get_file_contents(fpath)[0])
if os.path.isfile(fpath):
os.remove(fpath)
self.assertTrue(not os.path.isfile(fpath))
def test_put_file_contents_list(self):
_l, fpath = range(0, 10), os.path.join(self.resources, "foo.bar")
self.assertTrue(not os.path.isfile(fpath))
self.assertTrue(Doraemon.put_file_contents(fpath, _l))
self.assertTrue(os.path.isfile(fpath))
self.assertEqual(10, len(Doraemon.get_file_contents(fpath)))
self.assertEqual(map(lambda num: str(num), _l),
Doraemon.get_file_contents(fpath))
if os.path.isfile(fpath):
os.remove(fpath)
self.assertTrue(not os.path.isfile(fpath))
def test_get_file_filter(self):
self.assertEqual(
0,
len(
Doraemon.get_files(self.resources,
file_filter=lambda fpath: ".foobar" == fpath
.split(".")[-1])))
def save_transferred(self, raw_fname, transferred, output_dir=None):
output_fpath = "transferred_{0}.txt".format(
os.path.splitext(os.path.basename(raw_fname))[0])
if output_dir is not None:
if not os.path.isdir(output_dir):
os.makedirs(output_dir)
output_fpath = os.path.join(output_dir, output_fpath)
return (Doraemon.put_file_contents(output_fpath,
transferred), output_fpath)
def test_save_transferred(self):
tvaf = TransferVoiceAnnotationFiles()
tag_file = os.path.join(self.resources, "test.TAG")
transferred = tvaf.transfer_tagfile(tag_file)
is_saved, output_fpath = tvaf.save_transferred(tag_file, transferred)
self.assertTrue(is_saved)
self.assertTrue(os.path.isfile(output_fpath))
self.assertEqual(transferred, Doraemon.get_file_contents(output_fpath))
if os.path.isfile(output_fpath):
os.remove(output_fpath)
self.assertTrue(not os.path.isfile(output_fpath))
def transfer_tagfile(self,
tagfile_path,
in_charset=None,
out_charset=None):
transferred = []
lines = Doraemon.get_file_contents(tagfile_path,
in_charset=in_charset,
out_charset=out_charset)
for line in lines:
line = line.split(" ")
try:
if len(line) < 4 or line[1].startswith("0"):
continue
except Exception:
continue
transferred.append(line[-1])
return transferred
def test_get_files_by_suffix_lower(self):
self.assertTrue(
len(Doraemon.get_files_by_suffix(self.resources, ".TXT")) > 0)
def test_get_files_by_suffix(self):
self.assertEqual(
0, len(Doraemon.get_files_by_suffix(self.resources, ".foobar")))
def test_get_file_by_get_files(self):
filepath = os.path.join(self.resources, "none_contents_file.txt")
files = Doraemon.get_files(filepath)
self.assertEqual(files, [filepath])
def test_get_files(self):
files = Doraemon.get_files(self.resources)
self.assertTrue(
os.path.join(self.resources, "d", "e", "2.txt") in files)
def test_get_none_contents_file_contents(self):
fpath = os.path.join(self.resources, "none_contents_file.txt")
self.assertEqual(0, len(Doraemon.get_file_contents(fpath)))
# coding=utf-8
import os
import sys
sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
if __name__ == "__main__":
from doraemon import Doraemon
fpath = sys.argv[1]
contents = Doraemon.get_file_contents(fpath,
in_charset="gbk",
out_charset="utf-8")
Doraemon.put_file_contents("utf_8_{0}".format(os.path.basename(fpath)),
contents)
def get_tagfiles(self, input_dir):
return Doraemon.get_files_by_suffix(input_dir, ".tag")