def _fuzz(word, fuzziness=0.2):
"""Fuzz a word with noise.
Parameters
----------
word : str
A word to fuzz
fuzziness : float
How fuzzy to make the word
Returns
-------
str
A fuzzed word
"""
while True:
new_word = []
for ch in word:
if random() > fuzziness: # noqa: S311
new_word.append(ch)
else:
if random() > 0.5: # noqa: S311
new_word.append(choice(printable)) # noqa: S311
elif random() > 0.8: # noqa: S311
new_word.append(unichr(randint(0, 0x10FFFF))) # noqa: S311
else:
new_word.append(unichr(randint(0, 0xFFFF))) # noqa: S311
if random() > 0.5: # noqa: S311
new_word.append(ch)
new_word = ''.join(new_word)
if new_word != word:
return new_word
def test___init__(self):
expected = b'\xc3\xa9\xe0\xaf\xb2\xe0\xbe\x84'
if six.PY3:
expected = expected.decode('utf-8')
message = six.unichr(233) + six.unichr(0x0bf2) + six.unichr(3972)
exc = exception.IronicException(message)
self.assertEqual(expected, exc.__str__())
def _build_illegal_xml_regex():
"""Constructs a regex to match all illegal xml characters.
Expects to be used against a unicode string."""
# Construct the range pairs of invalid unicode characters.
illegal_chars_u = [
(0x00, 0x08), (0x0B, 0x0C), (0x0E, 0x1F), (0x7F, 0x84),
(0x86, 0x9F), (0xFDD0, 0xFDDF), (0xFFFE, 0xFFFF)]
# For wide builds, we have more.
if sys.maxunicode >= 0x10000:
illegal_chars_u.extend(
[(0x1FFFE, 0x1FFFF), (0x2FFFE, 0x2FFFF), (0x3FFFE, 0x3FFFF),
(0x4FFFE, 0x4FFFF), (0x5FFFE, 0x5FFFF), (0x6FFFE, 0x6FFFF),
(0x7FFFE, 0x7FFFF), (0x8FFFE, 0x8FFFF), (0x9FFFE, 0x9FFFF),
(0xAFFFE, 0xAFFFF), (0xBFFFE, 0xBFFFF), (0xCFFFE, 0xCFFFF),
(0xDFFFE, 0xDFFFF), (0xEFFFE, 0xEFFFF), (0xFFFFE, 0xFFFFF),
(0x10FFFE, 0x10FFFF)])
# Build up an array of range expressions.
illegal_ranges = [
"%s-%s" % (six.unichr(low), six.unichr(high))
for (low, high) in illegal_chars_u]
# Compile the regex
return re.compile(six.u('[%s]') % six.u('').join(illegal_ranges))
def test_dash_issue(self):
html = '<strong>—</strong>'
self.server.response['get.data'] = html
grab = build_grab()
grab.go(self.server.get_url())
# By default &#[128-160]; are fixed
self.assertFalse(grab.doc.select('//strong/text()').text()
== six.unichr(151))
self.assertTrue(grab.doc.select('//strong/text()').text()
== six.unichr(8212))
# disable fix-behaviour
grab.setup(fix_special_entities=False)
grab.go(self.server.get_url())
# By default &#[128-160]; are fixed
self.assertTrue(grab.doc.select('//strong/text()').text()
== six.unichr(151))
self.assertFalse(grab.doc.select('//strong/text()').text()
== six.unichr(8212))
# Explicitly use unicode_body func
grab = build_grab()
grab.go(self.server.get_url())
#print(':::', grab.doc.unicode_body())
self.assertTrue('—' in grab.doc.unicode_body())
def decompress(compressed):
"""Decompress a list of output ks to a string."""
# Build the dictionary.
dict_size = 0x10000
dictionary = dict((six.unichr(i), six.unichr(i)) for i in range(dict_size))
result = io.StringIO()
w = compressed.pop(0)
result.write(w)
for k in compressed:
if k in dictionary:
entry = dictionary[k]
elif k == dict_size:
entry = w + w[0]
else:
raise ValueError('Bad compressed k: %s' % k)
result.write(entry)
# Add w+entry[0] to the dictionary.
dictionary[dict_size] = w + entry[0]
dict_size += 1
w = entry
return result.getvalue()
def test_render_response_utf8(self):
req = apirequest.APIRequest("FakeAction", "FakeVersion", {})
resp = {
'utf8': six.unichr(40960) + u'abcd' + six.unichr(1972)
}
data = req._render_response(resp, 'uuid').decode()
self.assertIn('<utf8>ꀀabcd޴</utf8>', data)
def _init_cache():
'''Creates a mapping of wide->narrow and narrow->wide characters'''
global _wide_to_narrow
global _narrow_to_wide
_wide_to_narrow = {}
_narrow_to_wide = {}
char_names = {six.unichr(i): unicodedata.name(six.unichr(i), None)
for i in range(0, 65536)
}
for wide_ch, name in char_names.items():
if name is None:
continue
if name.upper().startswith('FULLWIDTH '):
half_name = name[len('FULLWIDTH '):]
else:
half_name = 'HALFWIDTH {}'.format(name)
try:
half_ch = unicodedata.lookup(half_name)
except KeyError:
pass
else:
_wide_to_narrow[wide_ch] = half_ch
_narrow_to_wide[half_ch] = wide_ch
logger.debug('Mapped %d characters from wide<->narrow',
len(_wide_to_narrow))
def setUp(self):
# These tests verify the UTF-8 decoder/validator on the various test cases from
# http://www.cl.cam.ac.uk/~mgk25/ucs/examples/UTF-8-test.txt
vs = []
for k in _create_utf8_test_sequences():
vs.extend(k[1])
# All Unicode code points
for i in range(
0, 0xffff): # should by 0x10ffff, but non-wide Python build is limited to 16-bits
if i < 0xD800 or i > 0xDFFF: # filter surrogate code points, which are disallowed to encode in UTF-8
vs.append((True, six.unichr(i).encode("utf-8")))
# FIXME: UnicodeEncodeError: 'utf-8' codec can't encode character '\ud800'
# in position 0: surrogates not allowed
if False:
# 5.1 Single UTF-16 surrogates
for i in range(0xD800, 0xDBFF): # high-surrogate
ss = six.unichr(i).encode("utf-8")
vs.append((False, ss))
for i in range(0xDC00, 0xDFFF): # low-surrogate
ss = six.unichr(i).encode("utf-8")
vs.append((False, ss))
# 5.2 Paired UTF-16 surrogates
for i in range(0xD800, 0xDBFF): # high-surrogate
for j in range(0xDC00, 0xDFFF): # low-surrogate
ss1 = six.unichr(i).encode("utf-8")
ss2 = six.unichr(j).encode("utf-8")
vs.append((False, ss1 + ss2))
vs.append((False, ss2 + ss1))
self._TEST_SEQUENCES = vs
def setUp(self):
path = tests.get_data_path(('PP', 'simple_pp', 'global.pp'))
self.cube_2d = iris.load_cube(path)
# Generate the unicode cube up here now it's used in two tests.
unicode_str = six.unichr(40960) + u'abcd' + six.unichr(1972)
self.unicode_cube = iris.tests.stock.simple_1d()
self.unicode_cube.attributes['source'] = unicode_str
def gen_utf8 ( file_bytes ):
global args
file_dict = {}
rows_dict = {}
# convert to utf-8
errors = 'strict'
if args.errors:
errors = 'replace'
try:
file_utf8 = file_bytes.decode('utf-8', errors)
pass
except UnicodeDecodeError:
print ( "file {0} is not valid utf-8, try analysing file as bytes using flag -b or enable error replacement with flag -e\n".format(source))
sys.exit ( 2 )
except:
traceback.print_exc()
sys.exit ( 1 )
for c in file_utf8:
if not c in file_dict:
file_dict[ c ] = 0
file_dict[ c ] += 1
if c > six.unichr(0x7f):
cp = ord( c )
cp = ( int( cp / 16 ) ) * 16
rows_dict[ six.unichr( cp ) ] = True
return ( file_dict, rows_dict )
def _mouse_handler(self, cli, mouse_event):
"""
Handle mouse events in a pane. A click in a non-active pane will select
it, one in an active pane, will send the mouse event to the application
running inside it.
"""
process = self.process
x = mouse_event.position.x
y = mouse_event.position.y
# The containing Window translates coordinates to the absolute position
# of the whole screen, but in this case, we need the relative
# coordinates of the visible area.
y -= self.process.screen.line_offset
if not self.has_focus(cli):
# Focus this process when the mouse has been clicked.
if mouse_event.event_type == MouseEventTypes.MOUSE_UP:
self.set_focus(cli)
else:
# Already focussed, send event to application when it requested
# mouse support.
if process.screen.sgr_mouse_support_enabled:
# Xterm SGR mode.
ev, m = {
MouseEventTypes.MOUSE_DOWN: ('0', 'M'),
MouseEventTypes.MOUSE_UP: ('0', 'm'),
MouseEventTypes.SCROLL_UP: ('64', 'M'),
MouseEventTypes.SCROLL_DOWN: ('65', 'M'),
}.get(mouse_event.event_type)
self.process.write_input(
'\x1b[<%s;%s;%s%s' % (ev, x + 1, y + 1, m))
elif process.screen.urxvt_mouse_support_enabled:
# Urxvt mode.
ev = {
MouseEventTypes.MOUSE_DOWN: 32,
MouseEventTypes.MOUSE_UP: 35,
MouseEventTypes.SCROLL_UP: 96,
MouseEventTypes.SCROLL_DOWN: 97,
}.get(mouse_event.event_type)
self.process.write_input(
'\x1b[%s;%s;%sM' % (ev, x + 1, y + 1))
elif process.screen.mouse_support_enabled:
# Fall back to old mode.
if x < 96 and y < 96:
ev = {
MouseEventTypes.MOUSE_DOWN: 32,
MouseEventTypes.MOUSE_UP: 35,
MouseEventTypes.SCROLL_UP: 96,
MouseEventTypes.SCROLL_DOWN: 97,
}.get(mouse_event.event_type)
self.process.write_input('\x1b[M%s%s%s' % (
six.unichr(ev),
six.unichr(x + 33),
six.unichr(y + 33)))
def set_unichr(x):
if isinstance(x, string_types):
return unichr(int(x, 16))
elif isinstance(x, integer_types):
return unichr(x)
else:
return unichr(int(x))
def test_reading_utf8_without_flag(self):
with open('test/utf8_without_leader_flag.dat', 'rb') as fh:
reader = MARCReader(fh, to_unicode=False)
record = next(reader)
self.assertEqual(type(record), Record)
utitle = record['240']['a']
self.assertEqual(type(utitle), binary_type)
self.assertEqual(utitle,
b'De la solitude a\xcc\x80 la communaute\xcc\x81.')
with open('test/utf8_without_leader_flag.dat', 'rb') as fh:
reader = MARCReader(fh, to_unicode=True, hide_utf8_warnings=True)
record = next(reader)
self.assertEqual(type(record), Record)
utitle = record['240']['a']
self.assertEqual(type(utitle), text_type)
# unless you force utf-8 characters will get lost and
# warnings will appear in the terminal
self.assertEqual(utitle, 'De la solitude a la communaute .')
# force reading as utf-8
with open('test/utf8_without_leader_flag.dat', 'rb') as fh:
reader = MARCReader(fh, to_unicode=True, force_utf8=True,
hide_utf8_warnings=True)
record = next(reader)
self.assertEqual(type(record), Record)
utitle = record['240']['a']
self.assertEqual(type(utitle), text_type)
self.assertEqual(utitle, u'De la solitude a' + unichr(0x0300) +
' la communaute' + unichr(0x0301) + '.')
def _next_code_point(val, val_iter, yield_char=False, to_int=lambda x: x):
"""Provides the next *code point* in the given Unicode sequence.
This generator function yields complete character code points, never incomplete surrogates. When a low surrogate is
found without following a high surrogate, this function raises ``ValueError`` for having encountered an unpaired
low surrogate. When the provided iterator ends on a high surrogate, this function yields ``None``. This is the
**only** case in which this function yields ``None``. When this occurs, the user may append additional data to the
input unicode sequence and resume iterating through another ``next`` on this generator. When this function receives
``next`` after yielding ``None``, it *reinitializes the unicode iterator*. This means that this feature can only
be used for values that contain an ``__iter__`` implementation that remains at the current position in the data
when called (e.g. :class:`BufferQueue`). At this point, there are only two possible outcomes:
* If next code point is a valid low surrogate, this function yields the combined code point represented by the
surrogate pair.
* Otherwise, this function raises ``ValueError`` for having encountered an unpaired high surrogate.
Args:
val (unicode|BufferQueue): A unicode sequence or unicode BufferQueue over which to iterate.
val_iter (Iterator[unicode|BufferQueue]): The unicode sequence iterator over ``val`` from which to generate the
next integer code point in the range ``0x0`` to ``0x10FFFF``.
yield_char (Optional[bool]): If True **and** the character code point resulted from a surrogate pair, this
function will yield a :class:`CodePoint` representing the character code point and containing the original
unicode character. This is useful when the original unicode character will be needed again because UCS2
Python builds will error when trying to convert code points greater than 0xFFFF back into their
unicode character representations. This avoids requiring the user to mathematically re-derive the
surrogate pair in order to successfully convert the code point back to a unicode character.
to_int (Optional[callable]): A function to call on each element of val_iter to convert that element to an int.
"""
high = next(val_iter)
low = None
code_point = to_int(high)
if _LOW_SURROGATE_START <= code_point <= _LOW_SURROGATE_END:
raise ValueError('Unpaired low surrogate in Unicode sequence: %d' % code_point)
elif _HIGH_SURROGATE_START <= code_point <= _HIGH_SURROGATE_END:
def combine_surrogates():
low_surrogate = next(val_iter)
low_code_point = to_int(low_surrogate)
if low_code_point < _LOW_SURROGATE_START or low_code_point > _LOW_SURROGATE_END:
raise ValueError('Unpaired high surrogate: %d' % code_point)
# Decode the surrogates
real_code_point = _NON_BMP_OFFSET
real_code_point += (code_point - _HIGH_SURROGATE_START) << 10
real_code_point += (low_code_point - _LOW_SURROGATE_START)
return real_code_point, low_surrogate
try:
code_point, low = combine_surrogates()
except StopIteration:
yield None
val_iter = iter(val) # More data has appeared in val.
code_point, low = combine_surrogates()
if yield_char and low is not None:
out = CodePoint(code_point)
if isinstance(val, six.text_type):
# Iterating over a text type returns text types.
out.char = high + low
else:
out.char = six.unichr(high) + six.unichr(low)
else:
out = code_point
yield out
def format_bar(cnt):
scaled = cnt*width/max_count
full = int(floor(scaled))
eighths = int(ceil((scaled-full)*8))
if eighths:
return full*six.unichr(0x2588) + six.unichr(0x2588+(8-eighths))
else:
return full*six.unichr(0x2588)
def load(self, fileName, hdrf):
f = open(fileName, 'rb')
strf = ''.join(struct.unpack('10s', f.read(10)))
if strf != '#?RADIANCE':
f.close()
return False
f.seek(1, 1)
cmd = []
c = 0
oldc = 0
while True:
oldc = c
c = ord(f.read(1))
if c is 0xa and oldc is 0xa:
break
cmd.append(str(sm.unichr(c)))
reso = []
while True:
c = ord(f.read(1))
cstr = str(sm.unichr(c))
reso.append(cstr)
if c is 0xa:
break
resoStr = "".join(reso)
resoStrUnformat = re.match('\-Y (?P<_0>\d+) \+X (?P<_1>\d+)', resoStr)
(ws, hs) = map(itemgetter(1), sorted(resoStrUnformat.groupdict().items()))
w = int(ws)
h = int(hs)
hdrf.width = w
hdrf.height = h
self.scanline = [[-1 for i in range(4)] for i in range(w)]
if not self.scanline:
print("File closed because scanline not found.")
f.close()
return False
scnidx = 0
# Convert image
for [] in range(h - 1, -1, -1):
# If self.scanline doesn't update is because of this
if (self.deCrunch(scnidx, w, f) is False):
break
# This should update the cols array in hdrf which is the HDR Class.
# If hdrf.cols doesn't update is because of this
self.workOnRGBE(hdrf, w)
scnidx = scnidx + 1
f.close()
def test_decode_entities(self):
html = u'→'
self.assertEquals(six.unichr(8594), decode_entities(html))
html = u'→'
self.assertEquals(six.unichr(8594), decode_entities(html))
html = u'→'
self.assertEquals(six.unichr(8594), decode_entities(html))
def toggle_advanced_panel(self):
"""Hides/Shows the advanced options panel"""
if bool(self.show_advanced.get()):
self.subFrame.grid(column=0,row=7,columnspan=4,rowspan=6, sticky=Tk.E+Tk.W, ipady=5)
self.subFrame.grid_columnconfigure(2, weight=1)
self.toggle_button.configure(text='Advanced ' + six.unichr(9660))
else:
self.subFrame.grid_forget()
self.toggle_button.configure(text='Advanced '+ six.unichr(9654))
def handle_charref(self, name):
if self.savedata is None:
return
if name[:1].lower() in ('x', 'u'):
c = unichr(int(name[1:], 16))
else:
c = unichr(int(name))
log.debug("Num ent: &%s; = %r", name, c)
self.handle_data(c)
def testUnicodeInRoute(self):
url = "/update/3/GMP/53.0/20170421105455/Linux_x86_64-gcc3/null/{}/Linux%204.4.0-53-generic%20(GTK%203.18.9,libpulse%208.0.0)/mint/1.0/update.xml"
# Test an arbitrary number of unicode chars
cdec = step = 42
while cdec < 4200:
channel = ''.join(unichr(c) for c in range(cdec, cdec + step) if unichr(c).isalpha())
cdec += step
url = url.format(channel)
ret = self.client.get(url)
self.assertEqual(ret.status_code, 200)
def _write(self, output, value):
coord = int(round(value * 1e5, 0))
coord <<= 1
coord = coord if coord >= 0 else ~coord
while coord >= 0x20:
output.write(six.unichr((0x20 | (coord & 0x1f)) + 63))
coord >>= 5
output.write(six.unichr(coord + 63))
def unichr2(i):
"""Returns a Unicode string of one character with ordinal 32 <= i,
otherwise an escaped control character.
"""
if 32 <= i:
return six.unichr(i)
elif six.unichr(i) in controlchars:
# we just return the character, unescaped
# if people want to escape them they can use escapecontrols
return six.unichr(i)
return "\\u%04x" % i
def repl(m):
if m.group(2) is not None:
high = int(m.group(1), 16)
low = int(m.group(2), 16)
if 0xD800 <= high <= 0xDBFF and 0xDC00 <= low <= 0xDFFF:
cp = ((high - 0xD800) << 10) + (low - 0xDC00) + 0x10000
return unichr(cp)
else:
return unichr(high) + unichr(low)
else:
return unichr(int(m.group(1), 16))
def show_status(self, changed):
status = ''
# set 'geotagged' status
if self.metadata.latlong:
status += six.unichr(0x2690)
# set 'unsaved' status
if changed:
status += six.unichr(0x26A1)
self.status.setText(status)
self._elide_name()
if changed:
self.image_list.new_metadata.emit(True)
def __init__(self,
start=0,
end=10,
width=12,
fill=six.unichr(0x25C9).encode("utf-8"),
blank=six.unichr(0x25CC).encode("utf-8"),
marker=six.unichr(0x25CE).encode("utf-8"),
format='[%(fill)s%(marker)s%(blank)s] %(progress)s%%',
incremental=True,
stdout=sys.stdout):
super(AnimatedProgressBar, self).__init__(start,end,width,fill,blank,marker,format,incremental)
self.stdout = stdout
def test_unaccepted_control_chars(self):
"""Tests we cannot assign the unaccepted control chars without escaping.
Source: https://en.wikipedia.org/wiki/Valid_characters_in_XML#XML_1.0
"""
exc_msg = ("All strings must be XML compatible: Unicode or ASCII, no "
"NULL bytes or control characters")
for code in xliff.ASCII_CONTROL_CODES:
self.unit.target = u'Een&#x%s;' % code.lstrip('0') or '0'
assert self.unit.target == u'Een%s' % six.unichr(int(code, 16))
self.unit.target = u'Een%s' % six.unichr(int(code, 16))
assert self.unit.target == u'Een%s' % six.unichr(int(code, 16))
def _write(self, output, curr_value, prev_value, factor):
curr_value = self._py2_round(curr_value * factor)
prev_value = self._py2_round(prev_value * factor)
coord = curr_value - prev_value
coord <<= 1
coord = coord if coord >= 0 else ~coord
while coord >= 0x20:
output.write(six.unichr((0x20 | (coord & 0x1f)) + 63))
coord >>= 5
output.write(six.unichr(coord + 63))
def handle_charref(self, ref):
# called for each character reference, e.g. for ' ', ref will be '160'
# Reconstruct the original character reference.
if ref.startswith('x'):
value = unichr(int(ref[1:], 16))
else:
value = unichr(int(ref))
if value in _cp1252.keys():
self.pieces.append('&#%s;' % hex(ord(_cp1252[value]))[1:])
else:
self.pieces.append('&#%(ref)s;' % locals())
def _send_cmd(self, command, server):
"""Internal method to send a single command to the pexpect object.
Args::
command: the command to send
server: the pexpect object to send to
Returns:
The formatted output of the command as a string, or -1 on timeout
"""
try:
if self.options["unix_line_endings"]:
server.send("{0}{1}".format(
command,
six.unichr(0x000A),
))
elif self.options["windows_line_endings"]:
server.send("{0}{1}{2}".format(
command,
six.unichr(0x000D),
six.unichr(0x000A),
))
else:
server.sendline(command)
cmd_response = server.expect(
self.options["shell_prompts"] +
self.options["extra_prompts"] +
self.options["passwd_prompts"],
self.options["cmd_timeout"],
)
if cmd_response >= (
len(self.options["shell_prompts"]) +
len(self.options["extra_prompts"])
) and len(self.options["second_password"] or "") > 0:
server.sendline(self.options["second_password"])
server.expect(
self.options["shell_prompts"] +
self.options["extra_prompts"],
self.options["cmd_timeout"],
)
except (pexpect.TIMEOUT, pexpect.EOF):
return self._try_for_unmatched_prompt(
server,
server.before,
command,
)
return format_output(server.before, command, self.options)
def test_unicode_scalar(self):
""" Storage of variable-length unicode strings (auto-creation) """
self.f.attrs['x'] = u"Hello" + six.unichr(0x2340) + u"!!"
out = self.f.attrs['x']
self.assertEqual(out, u"Hello" + six.unichr(0x2340) + u"!!")
self.assertEqual(type(out), six.text_type)
aid = h5py.h5a.open(self.f.id, b"x")
tid = aid.get_type()
self.assertEqual(type(tid), h5py.h5t.TypeStringID)
self.assertEqual(tid.get_cset(), h5py.h5t.CSET_UTF8)
self.assertTrue(tid.is_variable_str())
def handle_entityref(self, name):
c = unichr(name2codepoint[name])
print("Named ent:", c)
return default
elif l == 1:
return ord(s)
elif l == 2:
return struct.unpack('>H', s)[0]
elif l == 3:
return struct.unpack('>L', b'\x00' + s)[0]
elif l == 4:
return struct.unpack('>L', s)[0]
else:
raise TypeError('invalid length: %d' % l)
# decode_text
PDFDocEncoding = ''.join(
six.unichr(x) for x in (
0x0000,
0x0001,
0x0002,
0x0003,
0x0004,
0x0005,
0x0006,
0x0007,
0x0008,
0x0009,
0x000a,
0x000b,
0x000c,
0x000d,
0x000e,
def handle_charref(self, name):
if name.startswith('x'):
c = unichr(int(name[1:], 16))
else:
c = unichr(int(name))
print("Num ent :", c)
def replchar(m):
num = m.group(1)
return six.unichr(int(num))
def repl_oct_char(match):
return unichr(int(match.group(0)[1:], 8))
def handle_charref(self, name):
if name.startswith('x'):
codepoint = int(name[1:], 16)
else:
codepoint = int(name)
self.handle_ref(six.unichr(codepoint))
def _unicodeExpand(s):
return r_unicodeEscape.sub(lambda m: six.unichr(int(m.group(0)[2:], 16)),
s)
def render_char(self, code_point, is_first):
char = six.unichr(code_point)
if self.needs_escape(char, is_first):
return re.escape(char)
return char
from __future__ import print_function
import collections
import sys
import unicodedata
import six
from six.moves import range # pylint: disable=redefined-builtin
import tensorflow as tf
# Conversion between Unicode and UTF-8, if required (on Python2)
_native_to_unicode = (lambda s: s.decode("utf-8")) if six.PY2 else (lambda s: s)
# This set contains all letter and number characters.
_ALPHANUMERIC_CHAR_SET = set(
six.unichr(i) for i in range(sys.maxunicode)
if (unicodedata.category(six.unichr(i)).startswith("L") or
unicodedata.category(six.unichr(i)).startswith("N")))
def encode(text):
"""Encode a unicode string as a list of tokens.
Args:
text: a unicode string
Returns:
a list of tokens as Unicode strings
"""
if not text:
return []
ret = []
def property_chars(self, prefix):
return "".join(
six.unichr(x) for x in range(sys.maxunicode)
if unicodedata.category(six.unichr(x)).startswith(prefix))
def generate_ids(prefix=None,
postfix=None,
count=1,
length=None,
use_unicode=False,
_stepler_prefix=None):
"""Generate unique identificators, based on UUID.
Arguments:
prefix (str|None): prefix of unique ids
postfix (str|None): postfix of unique ids
count (int): count of unique ids
length (int|None): length of unique ids
use_unicode (boolean|False): generate str with unicode or not
_stepler_prefix (str, optional): Resources prefix is used to call
``generate_ids`` inside ``stepler.config`` and avoid cross imports
problem. By default it has value ``stepler.config.STEPLER_PREFIX``.
Returns:
generator: unique ids
"""
# TODO(schipiga): thirdparty module should know nothing about stepler
# configured values. We hack it for usability.
# ``If``-statement is used to allow ``generate_ids`` inside
# ``stepler.config`` and prevent cross imports problem.
if not _stepler_prefix:
from stepler import config
_stepler_prefix = config.STEPLER_PREFIX
if prefix:
prefix = prefix.strip('-')
if postfix:
postfix = postfix.strip('-')
# hash usually should have >= 7 symbols
uid_length = min_uid_length = 7
# calculate if it's possible to generate UID with requested length,
# postfix and prefix
if length:
const_length = len(_stepler_prefix + '-' +
(prefix + '-' if prefix else '') +
# uid will be here
('-' + postfix if postfix else ''))
uid_length = length - const_length
if uid_length < min_uid_length:
raise ValueError(
"According to passed prefix and postfix minimal length to "
"generate unique id must be equal or greater "
"than {0}.".format(const_length + min_uid_length))
for _ in range(count):
# mix constant stepler prefix to separate tested objects
uid = _stepler_prefix
if prefix:
uid += '-' + prefix
if use_unicode:
uid += '-' + u"".join(
six.unichr(random.choice(range(0x0400, 0x045F)))
for _ in range(uid_length))
else:
uid_val = str(uuid.uuid4())
uid_val = (uid_val * (uid_length // len(uid_val) + 1))[:uid_length]
uid += '-' + uid_val
if postfix:
uid += '-' + postfix
yield uid
"""various classes for establishing ground truth"""
from .classification import classes_to_numpy, classes_from_numpy, BLANK_CLASS
from .opencv_utils import show_image_and_wait_for_key, draw_segments, draw_classes
import numpy
import string
from six import text_type, unichr, moves
NOT_A_SEGMENT = unichr(10)
class Grounder(object):
def ground(self, imagefile, segments, external_data):
"""given an ImageFile, grounds it, through arbitrary data (better defined in subclasses)"""
raise NotImplementedError()
class TerminalGrounder(Grounder):
"""
Labels by using raw_input() to capture a character each line
"""
def ground(self, imagefile, segments, _=None):
classes = []
character = ""
print("Found %s segments to ground." % len(segments))
print("Type 'exit' to stop grounding the file.")
print("Type ' ' for anything that is not a character.")
print("Grounding will exit automatically after all segments.")
print("Going back to a previous segment is not possible at this time.")
for num in range(len(segments)):
while len(character) != 1:
def testUnicodeStr(self):
html = self.doc_class('<dtml-var a><dtml-var b>')
expected = u'\u07d0\xc8'
res = html(a=force_str(six.unichr(2000)), b=self._recode(b'\xc3\x88'))
self.assertEqual(res, expected)
def hook(obj):
result = {}
for key in obj:
result[key] = unichr(int(obj[key], 16))
return result
def _escape_char(c):
if isinstance(c, int):
c = six.unichr(c)
return c if c in LEGAL_CHARS else ESCAPE_FMT.format(ord(c))
def handle_charref(self, name):
if name.lower().startswith('x'):
text = six.unichr(int(name[1:], 16))
else:
text = six.unichr(int(name))
self._enqueue(TEXT, text)
def handle_charref(self, name):
"""Handle entries in the form &#NNNN; e.g. ⃡"""
self.handle_data(six.unichr(int(name)))
if sys.version_info >= (2, 7) or sys.version_info >= (3, 2):
import unittest as ut
else:
try:
import unittest2 as ut
except ImportError:
raise ImportError(
'unittest2 is required to run the test suite with python-%d.%d' %
(sys.version_info[:2]))
# Check if non-ascii filenames are supported
# Evidently this is the most reliable way to check
# See also h5py issue #263 and ipython #466
# To test for this, run the testsuite with LC_ALL=C
try:
testfile, fname = tempfile.mkstemp(unichr(0x03b7))
except UnicodeError:
UNICODE_FILENAMES = False
else:
UNICODE_FILENAMES = True
os.close(testfile)
os.unlink(fname)
del fname
del testfile
class TestCase(ut.TestCase):
"""
Base class for unit tests.
"""
@classmethod
from feature_extraction import FEATURE_DATATYPE
import numpy
import cv2
from opencv_utils import get_opencv_version
from six import unichr
CLASS_DATATYPE = numpy.uint16
CLASS_SIZE = 1
CLASSES_DIRECTION = 0 # vertical - a classes COLUMN
BLANK_CLASS = unichr(35) # marks unclassified elements
def classes_to_numpy(classes):
"""given a list of unicode chars, transforms it into a numpy array"""
import array
# utf-32 starts with constant ''\xff\xfe\x00\x00', then has little endian 32 bits chars
# this assumes little endian architecture!
assert unichr(15).encode('utf-32') == b'\xff\xfe\x00\x00\x0f\x00\x00\x00'
assert array.array("I").itemsize == 4
int_classes = array.array("I", "".join(classes).encode('utf-32')[4:])
assert len(int_classes) == len(classes)
classes = numpy.array(int_classes, dtype=CLASS_DATATYPE, ndmin=2) # each class in a column. numpy is strange :(
classes = classes if CLASSES_DIRECTION == 1 else numpy.transpose(classes)
return classes
def classes_from_numpy(classes):
"""reverses classes_to_numpy"""
classes = classes if CLASSES_DIRECTION == 0 else classes.tranpose()
classes = list(map(unichr, classes))
def property_chars(self, prefix):
#pylint:disable=no-self-use
return "".join(six.unichr(x) for x in range(sys.maxunicode) \
if unicodedata.category(six.unichr(x)).startswith(prefix))
def sign(n):
return cmp(n, 0)
r_unicodeEscape = re.compile(r'(\\u[0-9A-Fa-f]{4}|\\U[0-9A-Fa-f]{8})')
def _unicodeExpand(s):
return r_unicodeEscape.sub(lambda m: six.unichr(int(m.group(0)[2:], 16)),
s)
narrow_build = False
try:
six.unichr(0x10FFFF)
except ValueError:
narrow_build = True
if narrow_build:
def _unicodeExpand(s):
try:
return r_unicodeEscape.sub(
lambda m: six.unichr(int(m.group(0)[2:], 16)), s)
except ValueError:
warnings.warn(
'Encountered a unicode char > 0xFFFF in a narrow python build. '
'Trying to degrade gracefully, but this can cause problems '
'later when working with the string:\n%s' % s)
return r_unicodeEscape.sub(
class Junit(py.xml.Namespace):
pass
# We need to get the subset of the invalid unicode ranges according to
# XML 1.0 which are valid in this python build. Hence we calculate
# this dynamically instead of hardcoding it. The spec range of valid
# chars is: Char ::= #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD]
# | [#x10000-#x10FFFF]
_legal_chars = (0x09, 0x0A, 0x0D)
_legal_ranges = ((0x20, 0x7E), (0x80, 0xD7FF), (0xE000, 0xFFFD), (0x10000,
0x10FFFF))
_legal_xml_re = [
u"%s-%s" % (six.unichr(low), six.unichr(high))
for (low, high) in _legal_ranges if low < sys.maxunicode
]
_legal_xml_re = [six.unichr(x) for x in _legal_chars] + _legal_xml_re
illegal_xml_re = re.compile(u"[^%s]" % u"".join(_legal_xml_re))
del _legal_chars
del _legal_ranges
del _legal_xml_re
_py_ext_re = re.compile(r"\.py$")
def bin_xml_escape(arg):
def repl(matchobj):
i = ord(matchobj.group())
if i <= 0xFF:
def handle_entityref(self, name):
try:
text = six.unichr(entities.name2codepoint[name])
except KeyError:
text = '&%s;' % name
self._enqueue(TEXT, text)
' ': '_',
',': '_',
'.': '_',
'-': '_',
'/': '_PER_',
'%': 'PERCENT',
'[': '',
']': '',
'(': '',
')': '',
"'": '',
'8': 'EIGHT',
'15': 'FIFTEEN',
'30': 'THIRTY',
'\\': '_',
six.unichr(160): '_',
six.unichr(176): 'DEG_',
six.unichr(186): 'DEG_',
six.unichr(8211): '_',
}
def main():
"""Main entry point for UNECE code .xls parsing."""
parser = argparse.ArgumentParser(
description='Reads in a .xls file and generates a units module for '
'OpenHTF.',
prog='python units_from_xls.py')
parser.add_argument('xlsfile', type=str, help='the .xls file to parse')
parser.add_argument('--outfile',
type=str,
def test_render_response_utf8(self):
req = apirequest.APIRequest("FakeAction", "FakeVersion", {})
resp = {'utf8': six.unichr(40960) + u'abcd' + six.unichr(1972)}
data = req._render_response(resp, 'uuid').decode()
self.assertIn('<utf8>ꀀabcd޴</utf8>', data)
from __future__ import print_function
import collections
import sys
import unicodedata
import six
from six.moves import range # pylint: disable=redefined-builtin
import tensorflow.compat.v1 as tf
# Conversion between Unicode and UTF-8, if required (on Python2)
_native_to_unicode = (lambda s: s.decode('utf-8')) if six.PY2 else (lambda s: s)
# This set contains all letter and number characters.
_ALPHANUMERIC_CHAR_SET = set(
six.unichr(i)
for i in range(sys.maxunicode)
if (
unicodedata.category(six.unichr(i)).startswith('L')
or unicodedata.category(six.unichr(i)).startswith('N')
)
)
def encode(text):
"""Encode a unicode string as a list of tokens.
Args:
text: a unicode string
Returns:
a list of tokens as Unicode strings
def _get_all_chars():
for i in range(0xFFFF):
yield six.unichr(i)
class TestSshClient(base.TestCase):
SELECT_POLLIN = 1
@mock.patch('paramiko.RSAKey.from_private_key')
@mock.patch('six.StringIO')
def test_pkey_calls_paramiko_RSAKey(self, cs_mock, rsa_mock):
cs_mock.return_value = mock.sentinel.csio
pkey = 'mykey'
ssh.Client('localhost', 'root', pkey=pkey)
rsa_mock.assert_called_once_with(mock.sentinel.csio)
cs_mock.assert_called_once_with('mykey')
rsa_mock.reset_mock()
cs_mock.reset_mock()
pkey = mock.sentinel.pkey
# Shouldn't call out to load a file from RSAKey, since
# a sentinel isn't a basestring...
ssh.Client('localhost', 'root', pkey=pkey)
self.assertEqual(0, rsa_mock.call_count)
self.assertEqual(0, cs_mock.call_count)
def _set_ssh_connection_mocks(self):
client_mock = mock.MagicMock()
client_mock.connect.return_value = True
return (self.patch('paramiko.SSHClient'),
self.patch('paramiko.AutoAddPolicy'),
client_mock)
def test_get_ssh_connection(self):
c_mock, aa_mock, client_mock = self._set_ssh_connection_mocks()
s_mock = self.patch('time.sleep')
c_mock.return_value = client_mock
aa_mock.return_value = mock.sentinel.aa
# Test normal case for successful connection on first try
client = ssh.Client('localhost', 'root', timeout=2)
client._get_ssh_connection(sleep=1)
aa_mock.assert_called_once_with()
client_mock.set_missing_host_key_policy.assert_called_once_with(
mock.sentinel.aa)
expected_connect = [mock.call(
'localhost',
port=22,
username='******',
pkey=None,
key_filename=None,
look_for_keys=False,
timeout=10.0,
password=None,
sock=None
)]
self.assertEqual(expected_connect, client_mock.connect.mock_calls)
self.assertEqual(0, s_mock.call_count)
def test_get_ssh_connection_over_ssh(self):
c_mock, aa_mock, client_mock = self._set_ssh_connection_mocks()
proxy_client_mock = mock.MagicMock()
proxy_client_mock.connect.return_value = True
s_mock = self.patch('time.sleep')
c_mock.side_effect = [client_mock, proxy_client_mock]
aa_mock.return_value = mock.sentinel.aa
proxy_client = ssh.Client('proxy-host', 'proxy-user', timeout=2)
client = ssh.Client('localhost', 'root', timeout=2,
proxy_client=proxy_client)
client._get_ssh_connection(sleep=1)
aa_mock.assert_has_calls([mock.call(), mock.call()])
proxy_client_mock.set_missing_host_key_policy.assert_called_once_with(
mock.sentinel.aa)
proxy_expected_connect = [mock.call(
'proxy-host',
port=22,
username='******',
pkey=None,
key_filename=None,
look_for_keys=False,
timeout=10.0,
password=None,
sock=None
)]
self.assertEqual(proxy_expected_connect,
proxy_client_mock.connect.mock_calls)
client_mock.set_missing_host_key_policy.assert_called_once_with(
mock.sentinel.aa)
expected_connect = [mock.call(
'localhost',
port=22,
username='******',
pkey=None,
key_filename=None,
look_for_keys=False,
timeout=10.0,
password=None,
sock=proxy_client_mock.get_transport().open_session()
)]
self.assertEqual(expected_connect, client_mock.connect.mock_calls)
self.assertEqual(0, s_mock.call_count)
@mock.patch('time.sleep')
def test_get_ssh_connection_two_attemps(self, sleep_mock):
c_mock, aa_mock, client_mock = self._set_ssh_connection_mocks()
c_mock.return_value = client_mock
client_mock.connect.side_effect = [
socket.error,
mock.MagicMock()
]
client = ssh.Client('localhost', 'root', timeout=1)
client._get_ssh_connection(sleep=1)
# We slept 2 seconds: because sleep is "1" and backoff is "1" too
sleep_mock.assert_called_once_with(2)
self.assertEqual(2, client_mock.connect.call_count)
def test_get_ssh_connection_timeout(self):
c_mock, aa_mock, client_mock = self._set_ssh_connection_mocks()
timeout = 2
time_mock = self.patch('time.time')
time_mock.side_effect = utils.generate_timeout_series(timeout + 1)
c_mock.return_value = client_mock
client_mock.connect.side_effect = [
socket.error,
socket.error,
socket.error,
]
client = ssh.Client('localhost', 'root', timeout=timeout)
# We need to mock LOG here because LOG.info() calls time.time()
# in order to preprend a timestamp.
with mock.patch.object(ssh, 'LOG'):
self.assertRaises(exceptions.SSHTimeout,
client._get_ssh_connection)
# time.time() should be called twice, first to start the timer
# and then to compute the timedelta
self.assertEqual(2, time_mock.call_count)
@mock.patch('select.POLLIN', SELECT_POLLIN, create=True)
def test_timeout_in_exec_command(self):
chan_mock, poll_mock, _, _ = (
self._set_mocks_for_select([0, 0, 0], True))
# Test for a timeout condition immediately raised
client = ssh.Client('localhost', 'root', timeout=2)
with testtools.ExpectedException(exceptions.TimeoutException):
client.exec_command("test")
chan_mock.fileno.assert_called_once_with()
chan_mock.exec_command.assert_called_once_with("test")
chan_mock.shutdown_write.assert_called_once_with()
poll_mock.register.assert_called_once_with(
chan_mock, self.SELECT_POLLIN)
poll_mock.poll.assert_called_once_with(10)
@mock.patch('select.POLLIN', SELECT_POLLIN, create=True)
def test_exec_command(self):
chan_mock, poll_mock, select_mock, client_mock = (
self._set_mocks_for_select([[1, 0, 0]], True))
chan_mock.recv_exit_status.return_value = 0
chan_mock.recv.return_value = b''
chan_mock.recv_stderr.return_value = b''
client = ssh.Client('localhost', 'root', timeout=2)
client.exec_command("test")
chan_mock.fileno.assert_called_once_with()
chan_mock.exec_command.assert_called_once_with("test")
chan_mock.shutdown_write.assert_called_once_with()
select_mock.assert_called_once_with()
poll_mock.register.assert_called_once_with(
chan_mock, self.SELECT_POLLIN)
poll_mock.poll.assert_called_once_with(10)
chan_mock.recv_ready.assert_called_once_with()
chan_mock.recv.assert_called_once_with(1024)
chan_mock.recv_stderr_ready.assert_called_once_with()
chan_mock.recv_stderr.assert_called_once_with(1024)
chan_mock.recv_exit_status.assert_called_once_with()
client_mock.close.assert_called_once_with()
def _set_mocks_for_select(self, poll_data, ito_value=False):
gsc_mock = self.patch('tempest.lib.common.ssh.Client.'
'_get_ssh_connection')
ito_mock = self.patch('tempest.lib.common.ssh.Client._is_timed_out')
csp_mock = self.patch(
'tempest.lib.common.ssh.Client._can_system_poll')
csp_mock.return_value = True
select_mock = self.patch('select.poll', create=True)
client_mock = mock.MagicMock()
tran_mock = mock.MagicMock()
chan_mock = mock.MagicMock()
poll_mock = mock.MagicMock()
select_mock.return_value = poll_mock
gsc_mock.return_value = client_mock
ito_mock.return_value = ito_value
client_mock.get_transport.return_value = tran_mock
tran_mock.open_session().__enter__.return_value = chan_mock
if isinstance(poll_data[0], list):
poll_mock.poll.side_effect = poll_data
else:
poll_mock.poll.return_value = poll_data
return chan_mock, poll_mock, select_mock, client_mock
_utf8_string = six.unichr(1071)
_utf8_bytes = _utf8_string.encode("utf-8")
@mock.patch('select.POLLIN', SELECT_POLLIN, create=True)
def test_exec_good_command_output(self):
chan_mock, poll_mock, _, _ = (
self._set_mocks_for_select([1, 0, 0]))
closed_prop = mock.PropertyMock(return_value=True)
type(chan_mock).closed = closed_prop
chan_mock.recv_exit_status.return_value = 0
chan_mock.recv.side_effect = [self._utf8_bytes[0:1],
self._utf8_bytes[1:], b'R', b'']
chan_mock.recv_stderr.return_value = b''
client = ssh.Client('localhost', 'root', timeout=2)
out_data = client.exec_command("test")
self.assertEqual(self._utf8_string + 'R', out_data)
@mock.patch('select.POLLIN', SELECT_POLLIN, create=True)
def test_exec_bad_command_output(self):
chan_mock, poll_mock, _, _ = (
self._set_mocks_for_select([1, 0, 0]))
closed_prop = mock.PropertyMock(return_value=True)
type(chan_mock).closed = closed_prop
chan_mock.recv_exit_status.return_value = 1
chan_mock.recv.return_value = b''
chan_mock.recv_stderr.side_effect = [b'R', self._utf8_bytes[0:1],
self._utf8_bytes[1:], b'']
client = ssh.Client('localhost', 'root', timeout=2)
exc = self.assertRaises(exceptions.SSHExecCommandFailed,
client.exec_command, "test")
self.assertIn('R' + self._utf8_string, six.text_type(exc))
def test_exec_command_no_select(self):
gsc_mock = self.patch('tempest.lib.common.ssh.Client.'
'_get_ssh_connection')
csp_mock = self.patch(
'tempest.lib.common.ssh.Client._can_system_poll')
csp_mock.return_value = False
select_mock = self.patch('select.poll', create=True)
client_mock = mock.MagicMock()
tran_mock = mock.MagicMock()
chan_mock = mock.MagicMock()
# Test for proper reading of STDOUT and STDERROR
gsc_mock.return_value = client_mock
client_mock.get_transport.return_value = tran_mock
tran_mock.open_session().__enter__.return_value = chan_mock
chan_mock.recv_exit_status.return_value = 0
std_out_mock = mock.MagicMock(StringIO)
std_err_mock = mock.MagicMock(StringIO)
chan_mock.makefile.return_value = std_out_mock
chan_mock.makefile_stderr.return_value = std_err_mock
client = ssh.Client('localhost', 'root', timeout=2)
client.exec_command("test")
chan_mock.makefile.assert_called_once_with('rb', 1024)
chan_mock.makefile_stderr.assert_called_once_with('rb', 1024)
std_out_mock.read.assert_called_once_with()
std_err_mock.read.assert_called_once_with()
self.assertFalse(select_mock.called)
def repl_hex_char(match):
return unichr(int(match.group(0)[2:], 16))
def _draw_text_as_text(self, gc, x, y, s, prop, angle, ismath, mtext=None):
writer = self.writer
color = rgb2hex(gc.get_rgb())
style = {}
if color != '#000000':
style['fill'] = color
if gc.get_alpha() != 1.0:
style['opacity'] = six.text_type(gc.get_alpha())
if not ismath:
font = self._get_font(prop)
font.set_text(s, 0.0, flags=LOAD_NO_HINTING)
fontsize = prop.get_size_in_points()
fontfamily = font.family_name
fontstyle = prop.get_style()
attrib = {}
# Must add "px" to workaround a Firefox bug
style['font-size'] = six.text_type(fontsize) + 'px'
style['font-family'] = six.text_type(fontfamily)
style['font-style'] = prop.get_style().lower()
style['font-weight'] = prop.get_weight().lower()
attrib['style'] = generate_css(style)
if mtext and (angle == 0 or mtext.get_rotation_mode() == "anchor"):
# If text anchoring can be supported, get the original
# coordinates and add alignment information.
# Get anchor coordinates.
transform = mtext.get_transform()
ax, ay = transform.transform_point(mtext.get_position())
ay = self.height - ay
# Don't do vertical anchor alignment. Most applications do not
# support 'alignment-baseline' yet. Apply the vertical layout
# to the anchor point manually for now.
angle_rad = angle * np.pi / 180.
dir_vert = np.array([np.sin(angle_rad), np.cos(angle_rad)])
v_offset = np.dot(dir_vert, [(x - ax), (y - ay)])
ax = ax + v_offset * dir_vert[0]
ay = ay + v_offset * dir_vert[1]
ha_mpl_to_svg = {'left': 'start', 'right': 'end',
'center': 'middle'}
style['text-anchor'] = ha_mpl_to_svg[mtext.get_ha()]
attrib['x'] = str(ax)
attrib['y'] = str(ay)
attrib['style'] = generate_css(style)
attrib['transform'] = "rotate(%f, %f, %f)" % (-angle, ax, ay)
writer.element('text', s, attrib=attrib)
else:
attrib['transform'] = generate_transform([
('translate', (x, y)),
('rotate', (-angle,))])
writer.element('text', s, attrib=attrib)
if rcParams['svg.fonttype'] == 'svgfont':
fontset = self._fonts.setdefault(font.fname, set())
for c in s:
fontset.add(ord(c))
else:
writer.comment(s)
width, height, descent, svg_elements, used_characters = \
self.mathtext_parser.parse(s, 72, prop)
svg_glyphs = svg_elements.svg_glyphs
svg_rects = svg_elements.svg_rects
attrib = {}
attrib['style'] = generate_css(style)
attrib['transform'] = generate_transform([
('translate', (x, y)),
('rotate', (-angle,))])
# Apply attributes to 'g', not 'text', because we likely
# have some rectangles as well with the same style and
# transformation
writer.start('g', attrib=attrib)
writer.start('text')
# Sort the characters by font, and output one tspan for
# each
spans = {}
for font, fontsize, thetext, new_x, new_y, metrics in svg_glyphs:
style = generate_css({
'font-size': six.text_type(fontsize) + 'px',
'font-family': font.family_name,
'font-style': font.style_name.lower(),
'font-weight': font.style_name.lower()})
if thetext == 32:
thetext = 0xa0 # non-breaking space
spans.setdefault(style, []).append((new_x, -new_y, thetext))
if rcParams['svg.fonttype'] == 'svgfont':
for font, fontsize, thetext, new_x, new_y, metrics in svg_glyphs:
fontset = self._fonts.setdefault(font.fname, set())
fontset.add(thetext)
for style, chars in list(six.iteritems(spans)):
chars.sort()
same_y = True
if len(chars) > 1:
last_y = chars[0][1]
for i in xrange(1, len(chars)):
if chars[i][1] != last_y:
same_y = False
break
if same_y:
ys = six.text_type(chars[0][1])
else:
ys = ' '.join(six.text_type(c[1]) for c in chars)
attrib = {
'style': style,
'x': ' '.join(six.text_type(c[0]) for c in chars),
'y': ys
}
writer.element(
'tspan',
''.join(unichr(c[2]) for c in chars),
attrib=attrib)
writer.end('text')
if len(svg_rects):
for x, y, width, height in svg_rects:
writer.element(
'rect',
x=six.text_type(x), y=six.text_type(-y + height),
width=six.text_type(width), height=six.text_type(height)
)
writer.end('g')
