def parse(cls,
bstr=None,
filename=None,
bytes_input=None,
file_input=None,
offset_bytes=0,
headers_only=False,
recursive=True):
"""
Parse an MP4 file or bytes into boxes
:param bstr: The bitstring to parse
:type bstr: bitstring.ConstBitStream
:param filename: Filename of an mp4 file
:type filename: str
:param bytes_input: Bytes of an mp4 file
:type bytes_input: bytes
:param file_input: Filename or file object
:type file_input: str, file
:param offset_bytes: Start parsing at offset.
:type offset_bytes: int
:param headers_only: Ignore data and return just headers. Useful when data is cut short
:type: headers_only: boolean
:param recursive: Recursively load sub-boxes
:type: recursive: boolean
:return: BMFF Boxes or Headers
"""
if filename:
bstr = bs.ConstBitStream(filename=filename,
offset=offset_bytes * 8)
elif bytes_input:
bstr = bs.ConstBitStream(bytes=bytes_input,
offset=offset_bytes * 8)
elif file_input:
bstr = bs.ConstBitStream(auto=file_input, offset=offset_bytes * 8)
log.debug("Starting parse")
log.debug("Size is %d bits", bstr.len)
while bstr.pos < bstr.len:
log.debug("Byte pos before header: %d relative to (%d)",
bstr.bytepos, offset_bytes)
log.debug("Reading header")
header = cls.parse_header(bstr)
log.debug("Header type: %s", header.box_type)
log.debug("Byte pos after header: %d relative to (%d)",
bstr.bytepos, offset_bytes)
if headers_only:
yield header
# move pointer to next header if possible
try:
bstr.bytepos = header.start_pos + header.box_size
except ValueError:
log.warning("Premature end of data")
raise
else:
yield cls.parse_box(bstr, header, recursive=recursive)
def _message_to_bitarray(message: bytes) -> bitstring.ConstBitStream:
"""Encodes a message as a bitarray with length multiple of 5."""
barr = bitstring.ConstBitStream(message)
padding_len = 5 - (len(barr) % 5)
if padding_len < 5:
# The bitarray length has to be multiple of 5. If not, it is right-padded with zeros.
barr = bitstring.ConstBitStream(
bin="{}{}".format(barr.bin, '0' * padding_len))
return barr
def testCBS(self):
a = bitstring.ConstBitStream('0b11000')
b = bitstring.ConstBitStream('0b11000')
self.assertFalse(a is b)
# self.assertTrue(a._datastore is b._datastore)
def consume_subsignature(self, region):
"""
Consume subsignature packets
If a string is passed in as region, it is converted to a bitstream for you
"""
if isinstance(region, string_types):
if PY3:
region = region.encode()
region = bitstring.ConstBitStream(bytes=region)
elif isinstance(region, binary_type):
region = bitstring.ConstBitStream(bytes=region)
self.subsignature_consumer.consume(self, region)
def __init__(self, path):
self._header_raw = None
self.header = None
if isinstance(path, bytes):
self._replay = bitstring.ConstBitStream(path)
elif isinstance(path, str):
with open(path, 'rb') as f:
self._replay = bitstring.ConstBitStream(f.read())
assert hasattr(self, '_replay')
self._parse_meta()
self._parse_header()
def unpack_bit_packed_values(data_bytes, value_nbits, dtype):
total_bit_count = len(data_bytes) * 8
number_of_samples = total_bit_count / value_nbits
number_of_unused_bits = total_bit_count % value_nbits
# Note: We need to first reverse the order of the bytes. The BitStream library will read from the MSB of the first byte.
# Due to the packing format we will therefore need to read it in reverse. First removing the remaining bits that do not contain any sample values.
buffer = array.array("B", data_bytes[::-1]).tostring()
bit_string = bitstring.ConstBitStream(bytes=buffer)
unused_bits = bit_string.read(number_of_unused_bits)
unpacked_values = []
for sample_index in range(number_of_samples):
unpacked_values.append(bit_string.read(value_nbits))
if dtype == 'uint':
unpacked_values = [x.uint for x in unpacked_values]
if dtype == 'intbe':
# Note that because we have revered the order of the bytes this is BE now rather than LE, as in the original format.
unpacked_values = [x.intbe for x in unpacked_values]
# Now we need to reverse the samples back to their original order.
unpacked_values = unpacked_values[::-1]
return unpacked_values
def getbits_rom(f):
# random rom file they gave me
buff = bytearray()
for b in f.read():
# Reverse bits, swap nibbles
buff += chr(munge(ord(b)))
return bitstring.ConstBitStream(bytes=buff)
def encode():
st = bitstring.ConstBitStream(filename=sys.argv[2])
f2 = open(sys.argv[3], 'w')
first_code = True
def write_b(b):
nonlocal first_code
nonlocal f2
if first_code:
f2.write(b)
first_code = False
else:
f2.write(' ' + b)
while True:
try:
a = st.read(18).uint
b = num2word[a]
write_b(b)
except bitstring.ReadError:
tail = st[st.pos:] + '0b1'
if tail.len < 18:
tail += bitstring.Bits(18 - tail.len)
b = num2word[tail.uint]
write_b(b)
break
f2.close()
def bitarray_to_u5(barr):
assert barr.len % 5 == 0
ret = []
s = bitstring.ConstBitStream(barr)
while s.pos != s.len:
ret.append(s.read(5).uint)
return ret
def __init__(self, user, record):
md5 = hashlib.md5()
md5.update(user.encode())
md5.update(record)
_hash = md5.hexdigest()
try:
_datetime = dt.datetime.strptime(record[0:10].decode(),
'%y%m%d%H%M')
except ValueError:
_datetime = dt.datetime(2015, 1, 1)
pressure_data = binascii.unhexlify(record[16:24])
# blood pressure data is in 30 bits
bs = bitstring.ConstBitStream(bytes=pressure_data)
bs.pos += 2
_pulse, _dia, _sys = bs.readlist(['uint:10, uint:10, uint:10'])
_pp = _sys - _dia
_map = _dia + _pp / 3
super().__init__(hash=_hash,
datetime=_datetime,
sys=_sys,
dia=_dia,
pulse=_pulse,
pp=_pp,
map=_map)
def __init__(self, data, offset, remaining=None):
# okay, we need to calculate the size first
self.size_in_bits = 0
bs = bitstring.ConstBitStream(bytes=data)
bs.bytepos = offset
for field in self._struct:
fmt, name, field_size = field
# TODO: handle regular struct sizes later?
# or merge back into Unpackable?
if not isinstance(field_size, int):
# look up the bit size from an earlier field
field_size = self.__dict__[field_size]
if fmt == 'UB':
self.__dict__[name] = bs.read('uint:%d' % field_size)
elif fmt == 'SB':
self.__dict__[name] = bs.read('int:%d' % field_size)
else:
raise ValueError("unknown fmt " + fmt)
self.size_in_bits += field_size
bs.bytealign()
self._size = bs.bytepos - offset
def fixRIC(fileName, recordNumbers, correctRIC, recLen):
numberRecords = os.stat(fileName).st_size / recLen
fileHandle = bitstring.ConstBitStream(filename=fileName)
newFileHandle = open(fileName + '.rw', 'wb')
for recordCount in xrange(0, numberRecords):
data = fileHandle.read(recLen * 8)
if recordCount + 1 in recordNumbers and len(recordNumbers) > 0:
dataStart = data[44 * 8:46 * 8].peek('uint:16')
if debug:
print 'Here is the data start: ' + str(dataStart)
print 'Here is the RIC location: ' + str(data[
(dataStart + 8) * 8:(dataStart + 12) * 8].peek('int:32'))
newdata = data[0:(dataStart + 8) * 8] + bitstring.pack(
'int:32', correctRIC[0]) + data[(dataStart + 12) * 8:]
newdata.tofile(newFileHandle)
recordNumbers.pop(0)
correctRIC.pop(0)
if debug:
print 'Here is how many bad records we have left: ' + str(
len(recordNumbers))
else:
data.tofile(newFileHandle)
newFileHandle.close()
if '.rw' in fileName:
shutil.move(fileName + '.rw', fileName)
return
def testConstBitStreamCopy(self):
cbs = bitstring.ConstBitStream(100)
cbs.pos = 50
cbs_copy = copy.copy(cbs)
self.assertEqual(cbs_copy.pos, 0)
self.assertTrue(cbs._datastore is cbs_copy._datastore)
self.assertTrue(cbs == cbs_copy)
def fixSampleCnt(fileName, recordNumbers, correctNumbers, recLen):
numberRecords = os.stat(fileName).st_size / recLen
fileHandle = bitstring.ConstBitStream(filename=fileName)
newFileHandle = open(fileName + '.rw', 'wb')
for recordCount in xrange(0, numberRecords):
data = fileHandle.read(recLen * 8)
if recordCount + 1 in recordNumbers and len(recordNumbers) > 0:
if debug:
print 'Here is the sample count: ' + str(
data[30 * 8:32 * 8].peek('uint:16'))
newdata = data[0:30 * 8] + bitstring.pack(
'uint:16', correctNumbers[0]) + data[32 * 8:]
newdata.tofile(newFileHandle)
recordNumbers.pop(0)
correctNumbers.pop(0)
if debug:
print 'Here is how many bad records we have left: ' + str(
len(recordNumbers))
else:
data.tofile(newFileHandle)
newFileHandle.close()
if '.rw' in fileName:
shutil.move(fileName + '.rw', fileName)
return
def read_record_ETL4(f, pos):
f = bitstring.ConstBitStream(filename=f)
f.bytepos = pos * 2952
r = f.readlist(
'2*uint:36,uint:8,pad:28,uint:8,pad:28,4*uint:6,pad:12,15*uint:36,pad:1008,bytes:21888'
)
"""
print('Serial Data Number:', r[0])
print('Serial Sheet Number:', r[1])
print('JIS Code:', r[2])
print('EBCDIC Code:', r[3])
print('4 Character Code:', ''.join([t56s[c] for c in r[4:8]]))
print('Evaluation of Individual Character Image:', r[8])
print('Evaluation of Character Group:', r[9])
print('Sample Position Y on Sheet:', r[10])
print('Sample Position X on Sheet:', r[11])
print('Male-Female Code:', r[12])
print('Age of Writer:', r[13])
print('Industry Classification Code:', r[14])
print('Occupation Classifiaction Code:', r[15])
print('Sheet Gatherring Date:', r[16])
print('Scanning Date:', r[17])
print('Number of X-Axis Sampling Points:', r[18])
print('Number of Y-Axis Sampling Points:', r[19])
print('Number of Levels of Pixel:', r[20])
print('Magnification of Scanning Lens:', r[21])
print('Serial Data Number (old):', r[22])
"""
return r
def testAppendToBits(self):
a = Bits(BitArray())
with self.assertRaises(AttributeError):
a.append('0b1')
self.assertEqual(type(a), Bits)
b = bitstring.ConstBitStream(bitstring.BitStream())
self.assertEqual(type(b), bitstring.ConstBitStream)
def main(source, destination):
t56s = '0123456789[#@:>? ABCDEFGHI&.](< JKLMNOPQR-$*);\'|/STUVWXYZ ,%="!'
def T56(c):
return t56s[c]
with codecs.open('co59-utf8.txt', 'r', 'utf-8') as co59f:
print(co59f)
co59t = co59f.read()
co59l = co59t.split()
CO59 = {}
for c in co59l:
ch = c.split(':')
co = ch[1].split(',')
CO59[(int(co[0]),int(co[1]))] = ch[0]
files = glob.glob(source)
# filename = 'E:\kanji dataset\extract\ETL2\ETL2_1'
filename = source
#skip = 0
f = bitstring.ConstBitStream(filename=filename)
for skip in range(11420):
f.pos = skip * 6 * 3660
r = f.readlist('int:36,uint:6,pad:30,6*uint:6,6*uint:6,pad:24,2*uint:6,pad:180,bytes:2700')
# print(r[0], T56(r[1]), "".join(map(T56, r[2:8])), "".join(map(T56, r[8:14])), CO59[tuple(r[14:16])])
iF = Image.frombytes('F', (60,60), r[16], 'bit', 6)
iP = iF.convert('L')
imagefilepath = destination+ '/%s' % CO59[tuple(r[14:16])]
if not os.path.exists(imagefilepath):
os.makedirs(imagefilepath)
fn = '%s/{:d}.png'.format(r[0]) % imagefilepath
iP.save(fn, 'PNG', bits=6)
enhancer = ImageEnhance.Brightness(iP)
iE = enhancer.enhance(4)
iE.save(fn, 'PNG')
def last(self):
self.pos = self.filesize - 188
self.tsopen.seek(self.pos)
self.bytes = self.tsopen.read(188)
self.bits = bitstring.ConstBitStream(bytes=self.bytes, length=1504)
self.packetnum = self.totalpackets
tsdata = self.decodets()
return tsdata
def decompress(inBytes, startOffset=0):
# Define some useful constants.
SEARCH_LOG2 = 8
SEARCH_SIZE = 2**SEARCH_LOG2
LOOKAHEAD_LOG2 = 4
LOOKAHEAD_SIZE = 2**LOOKAHEAD_LOG2
BIT_PASTCOPY = 0
BIT_LITERAL = 1
# Prepare to read the compressed bytes.
inStream = bitstring.ConstBitStream(bytes=inBytes)
inStream.bytepos = startOffset
# Allocate memory for the decompression process.
decompSize = inStream.read('uintle:16')
decomp = bytearray([0x00] * decompSize)
decompPos = 0
window = bytearray([0x20] * SEARCH_SIZE)
windowPos = 0xEF
# Main decompression loop.
while decompPos < decompSize:
nextCommand = inStream.read('bool')
if nextCommand == BIT_PASTCOPY:
# 0: Pastcopy case.
copySource = inStream.read(SEARCH_LOG2).uint
copyLength = inStream.read(LOOKAHEAD_LOG2).uint
copyLength += 2
# Truncate copies that would exceed "decompSize" bytes.
if (decompPos + copyLength) >= decompSize:
copyLength = decompSize - decompPos
for i in range(copyLength):
decomp[decompPos] = window[copySource]
decompPos += 1
window[windowPos] = window[copySource]
windowPos += 1
windowPos %= SEARCH_SIZE
copySource += 1
copySource %= SEARCH_SIZE
elif nextCommand == BIT_LITERAL:
# 1: Literal case.
literalByte = inStream.read('uint:8')
decomp[decompPos] = literalByte
decompPos += 1
window[windowPos] = literalByte
windowPos += 1
windowPos %= SEARCH_SIZE
# Calculate the end offset.
inStream.bytealign()
endOffset = inStream.bytepos
# Return the decompressed data and end offset.
return (decomp, endOffset)
def UnpackMessageFromSocket( sock ):
zeroLenChunkMax = 2
# Read the header bytes to get the messageID and length.
headerBytes = (16+32) >> 3
zeroLenChunkCount = 0
chunks = []
messageLen = 0
while messageLen < headerBytes:
# print headerBytes, messageLen, headerBytes - messageLen
chunk = sock.recv( headerBytes - messageLen )
if not chunk:
zeroLenChunkCount += 1
if zeroLenChunkCount < zeroLenChunkMax:
continue
raise RuntimeError( 'LLRP socket connection broken' )
chunks.append( chunk )
messageLen += len(chunk)
zeroLenChunkCount = 0
# Convert to a BitStream to get the message TypeCode and Length.
s = bitstring.ConstBitStream( bytes=b''.join(chunks) )
TypeCode = s.read('uintbe:16')
TypeCode &= ((1<<10)-1)
Length = s.read('uintbe:32')
# print( 'UnpackMessageFromSocket: TypeCode={} Length={} {}'.format(TypeCode, Length, _messageClassFromTypeCode[TypeCode].__name__) )
# Read the remaining message based on the Length.
zeroLenChunkCount = 0
while messageLen < Length:
chunk = sock.recv( Length - messageLen )
if not chunk:
zeroLenChunkCount += 1
if zeroLenChunkCount < zeroLenChunkMax:
continue
raise RuntimeError( 'LLRP socket connection broken' )
chunks.append( chunk )
messageLen += len(chunk)
zeroLenChunkCount = 0
# Convert the full message to a BitStream and parse it.
s = bitstring.ConstBitStream( bytes=b''.join(chunks) )
return _MessagePackUnpackLookup[TypeCode].unpack( s )
def bitstream_from_file_object(file_input, begin_at_offset_in_bytes: int = 0):
"""
:param file_input: opened file that contains CSME code binaries
(i.e. the input file provided by the user on the command line)
:param begin_at_offset_in_bytes: where to fast-forward seek to in the file before attempting a search
:return: a bitstring.ConstBitStream of the specified file's data, cued to the specified offset location
"""
return bitstring.ConstBitStream(file_input,
offset=begin_at_offset_in_bytes * 8)
def receive(self):
self.header = self.read(9)
log.info('header: %s', self.header)
parsed = bitstring.ConstBitStream(self.header).readlist(KP_MSG_HEADER)
size = parsed[-1]
self.payload = self.read(size * 5)
log.info('payload: %s', self.payload)
self.footer = self.read(4)
log.info('footer: %s', self.footer)
def verify(self, msg, signature):
hashedMsg = WinternitzPlusOTS.sha256Bytes(msg)
hashedBitString = bitstring.ConstBitStream(hashedMsg)
msgPubKey = []
# Apply logic here
if msgPubKey == self.publicKey:
return True
else:
return False
def reverse_bytewise(bitstream, dbg=False):
result = []
if dbg:
print(bitstream.bin)
for byte in bitstream.tobytes():
if dbg:
print(hex(byte))
result.append(reverse_byte(byte))
reverse_bytes = bitstring.ConstBitStream(bytes=result)
return reverse_bytes
def position_data_index(fn):
"""Given the filename of a position file, return the bit index of the first
data record past the header.
"""
hdr_end = "%%ENDHEADER\r\n"
token = bstr.ConstBitArray(bytearray(hdr_end))
bits = bstr.ConstBitStream(filename=fn)
bits.find(token)
bits.read('bytes:%d' % len(hdr_end))
return bits.pos
def __init__(self, filename, data):
super(AppleSoft, self).__init__()
self.filename = filename
data = bitstring.ConstBitStream(data)
# TODO: assert length is met
self.length = data.read('uintle:16')
self.lines = []
self.program = {}
last_line_number = -1
last_memory = 0x801
while data:
next_memory, line_number = data.readlist('uintle:16, uintle:16')
if not next_memory:
break
line = []
bytes_read = 4
while True:
token = data.read('uint:8')
bytes_read += 1
if token == 0:
self.lines.append(line_number)
self.program[line_number] = ''.join(line)
break
if token >= 0x80:
try:
line.append(' ' + TOKENS[token] + ' ')
except KeyError:
self.anomalies.append(
anomaly.Anomaly(
self, anomaly.CORRUPTION,
'Line number %d contains unexpected token: %02X'
% (line_number, token)))
else:
line.append(chr(token))
if last_memory + bytes_read != next_memory:
self.anomalies.append(
anomaly.Anomaly(
self, anomaly.UNUSUAL, "%x + %x == %x != %x (gap %d)" %
(last_memory, bytes_read, last_memory + bytes_read,
next_memory, next_memory - last_memory - bytes_read)))
if line_number <= last_line_number:
self.anomalies.append(
anomaly.Anomaly(
self, anomaly.UNUSUAL, "%d <= %d: %s" %
(line_number, last_line_number, ''.join(line))))
last_line_number = line_number
last_memory = next_memory
def decode(file_name):
# Reading bits
b = bs.ConstBitStream(filename=file_name)
bitSequence = []
try:
print(chr(27) + "[2J ", end='')
print('READING FILE ...')
while True:
bitSequence.append(b.read('uint:1'))
except bs.ReadError:
pass
# Cutting padding
padding = int("".join(str(x) for x in bitSequence[0:8]), 2)
bitSequence = bitSequence[8:len(bitSequence) - padding + 1]
code_length = 8
i = 0
root = HeapNode('NYT', 0)
heap = [root]
currentNode = root
output = ''
last_progress = -1
while i <= len(bitSequence):
progress = round(i * 100 / len(bitSequence))
if progress > last_progress:
last_progress = progress
print(chr(27) + "[2J ", end='')
print(f"PROGRESS = {round(i*100/len(bitSequence))}%")
# Node is a leaf
if currentNode.left is None and currentNode.right is None:
# Add new letter
if currentNode.char == 'NYT':
char = int(''.join(map(str, bitSequence[i:i + code_length])),
2)
i += code_length
# Letter is in the heap
else:
char = currentNode.char
add_to_heap(char, heap)
currentNode = root
output += chr(char)
# Traversing tree looking for a leaf
else:
if i != len(bitSequence):
bit = bitSequence[i]
currentNode = currentNode.left if bit == 0 else currentNode.right
i += 1
return output
def load_entries_to_memory(self):
file_name = 'raw_data/etl2_entries.obj'
try:
file_handler = open(file_name, 'rb')
entries = pickle.load(file_handler)
print('restored pickled etl2 data')
return entries
except:
print('processing raw etl2 data')
entries = []
image_size = (60, 60)
bits_per_pixel = 6
for file_directory, num_items in self.files:
file = bitstring.ConstBitStream(filename=file_directory)
# loop through the items in each file
for item_index in range(num_items):
file.pos = item_index * 6 * 3660
item_data = file.readlist(
'int:36,uint:6,pad:30,6*uint:6,6*uint:6,pad:24,2*uint:6,pad:180,bytes:2700'
)
# specifications about each item's data
# http://etlcdb.db.aist.go.jp/?page_id=1721
# 0 -> serial index
# 1 -> source, in T56
# 2:8 -> name of type of character, kanji, kana, in T56
# 8:14 -> name of font type, in T56
# 14:16 -> label
# 16 -> image bits
# print item_data[0], T56(r[1]), "".join(map(T56, item_data[2:8])), "".join(map(T56, r[8:14])), CO59[tuple(r[14:16])])
# save only the label & image
label = Etl2Dataset.CO59[tuple(item_data[14:16])]
# image is grayscale, use otsu's algorithm to binarize it
pil_image = Image.frombytes('F', image_size, item_data[16],
'bit', bits_per_pixel)
# np_image = np.array(pil_image)
# global_threshold = filters.threshold_otsu(np_image)
# binarized_image = np_image > global_threshold
# fromarray '1' is buggy, convert array to 0 & 255 uint8,
# then build image with PIL as 'L' & convert to '1'
# pil_image = Image.fromarray((binarized_image * 255).astype(np.uint8), mode='L')
entries.append(
CharacterEntry(pil_image=pil_image, label=label))
# save the data to file so we don't have to load it again
file_handler = open(file_name, 'wb')
pickle.dump(entries, file_handler)
return entries
def __init__(self, irsb, data, max_inst, max_bytes, bytes_offset, opt_level=None,
traceflags=None, allow_lookback=None):
super(GymratLifter, self).__init__(irsb, data, max_inst, max_bytes,
bytes_offset, opt_level, traceflags, allow_lookback)
self.logger = logging.getLogger('lifter')
self.logger.setLevel(logging.DEBUG)
if 'CData' in str(type(data)):
thedata = "".join([chr(data[x]) for x in range(max_bytes)])
else:
thedata = data
self.bitstrm = bitstring.ConstBitStream(bytes=thedata)
def testBaselineMemory(self):
try:
import pympler.asizeof.asizeof as size
except ImportError:
return
# These values might be platform dependent, so don't fret too much.
self.assertEqual(size(bitstring.ConstBitStream([0])), 64)
self.assertEqual(size(bitstring.Bits([0])), 64)
self.assertEqual(size(bitstring.BitStream([0])), 64)
self.assertEqual(size(bitstring.BitArray([0])), 64)
from bitstring.bitstore import ByteStore
self.assertEqual(size(ByteStore(bytearray())), 100)
