# -*- coding: utf-8 -*-
""" Store all usefull functions here """
import os
import sys
import time
import re
from os.path import join
from string import maketrans
from htmlentitydefs import name2codepoint
def chmod(*args):
try:
os.chmod(*args)
except:
pass
def decode(string):
""" decode string with utf if possible """
try:
return string.decode("utf8", "replace")
except:
return string
def remove_chars(string, repl):
""" removes all chars in repl from string"""
if type(string) == str:
return string.translate(maketrans("", ""), repl)
elif type(string) == unicode:
return string.translate(dict([(ord(s), None) for s in repl]))
def save_path(name):
#remove some chars
if os.name == 'nt':
return remove_chars(name, '/\\?%*:|"<>')
else:
return remove_chars(name, '/\\"')
def save_join(*args):
""" joins a path, encoding aware """
return fs_encode(join(*[x if type(x) == unicode else decode(x) for x in args]))
# File System Encoding functions:
# Use fs_encode before accesing files on disk, it will encode the string properly
if sys.getfilesystemencoding().startswith('ANSI'):
def fs_encode(string):
try:
string = string.encode('utf-8')
finally:
return string
fs_decode = decode #decode utf8
else:
fs_encode = fs_decode = lambda x: x # do nothing
def get_console_encoding(enc):
if os.name == "nt":
if enc == "cp65001": # aka UTF-8
print "WARNING: Windows codepage 65001 is not supported."
enc = "cp850"
else:
enc = "utf8"
return enc
def compare_time(start, end):
start = map(int, start)
end = map(int, end)
if start == end: return True
now = list(time.localtime()[3:5])
if start < now < end: return True
elif start > end and (now > start or now < end): return True
elif start < now > end < start: return True
else: return False
def formatSize(size):
"""formats size of bytes"""
size = int(size)
steps = 0
sizes = ["B", "KiB", "MiB", "GiB", "TiB"]
while size > 1000:
size /= 1024.0
steps += 1
return "%.2f %s" % (size, sizes[steps])
def formatSpeed(speed):
return formatSize(speed) + "/s"
def freeSpace(folder):
if os.name == "nt":
import ctypes
free_bytes = ctypes.c_ulonglong(0)
ctypes.windll.kernel32.GetDiskFreeSpaceExW(ctypes.c_wchar_p(folder), None, None, ctypes.pointer(free_bytes))
return free_bytes.value
else:
from os import statvfs
s = statvfs(folder)
return s.f_bsize * s.f_bavail
def uniqify(seq, idfun=None):
# order preserving
if idfun is None:
def idfun(x): return x
seen = {}
result = []
for item in seq:
marker = idfun(item)
# in old Python versions:
# if seen.has_key(marker)
# but in new ones:
if marker in seen: continue
seen[marker] = 1
result.append(item)
return result
def parseFileSize(string, unit=None): #returns bytes
if not unit:
m = re.match(r"(\d*[\.,]?\d+)(.*)", string.strip().lower())
if m:
traffic = float(m.group(1).replace(",", "."))
unit = m.group(2)
else:
return 0
else:
if isinstance(string, basestring):
traffic = float(string.replace(",", "."))
else:
traffic = string
#ignore case
unit = unit.lower().strip()
if unit in ("gb", "gig", "gbyte", "gigabyte", "gib", "g"):
traffic *= 1 << 30
elif unit in ("mb", "mbyte", "megabyte", "mib", "m"):
traffic *= 1 << 20
elif unit in ("kb", "kib", "kilobyte", "kbyte", "k"):
traffic *= 1 << 10
return traffic
def lock(func):
def new(*args):
#print "Handler: %s args: %s" % (func,args[1:])
args[0].lock.acquire()
try:
return func(*args)
finally:
args[0].lock.release()
return new
def fixup(m):
text = m.group(0)
if text[:2] == "&#":
# character reference
try:
if text[:3] == "&#x":
return unichr(int(text[3:-1], 16))
else:
return unichr(int(text[2:-1]))
except ValueError:
pass
else:
# named entity
try:
name = text[1:-1]
text = unichr(name2codepoint[name])
except KeyError:
pass
return text # leave as is
def html_unescape(text):
"""Removes HTML or XML character references and entities from a text string"""
return re.sub("&#?\w+;", fixup, text)
if __name__ == "__main__":
print freeSpace(".")
print remove_chars("ab'cdgdsf''ds'", "'ghd")