def detect_local_source(path, content, mime_type=None, encoding=None):
# TODO: may add sample_size
filename = os.path.basename(path)
parts = filename.split('.')
extension = parts[-1] if len(parts) > 1 else None
if magic is not None:
detected = magic.detect_from_content(content)
encoding = detected.encoding or encoding
mime_name = detected.name
mime_type = detected.mime_type or mime_type
else:
encoding = chardet.detect(content)['encoding'] or encoding
mime_name = None
mime_type = mime_type or mimetypes.guess_type(filename)[0]
plugin_name = plugin_name_by_mime_type(mime_type, mime_name, extension)
if encoding == 'binary':
encoding = None
return Source(uri=path,
plugin_name=plugin_name,
encoding=encoding)
def download_file(uri, verify_ssl):
response = requests.get(uri, verify=verify_ssl)
content = response.content
if magic is not None:
encoding = magic.detect_from_content(content).encoding
else:
encoding = response.encoding
# TODO: try to guess with uri.split('/')[-1].split('.')[-1].lower()
# TODO: try to guess with file-magic lib
try:
content_type = response.headers['content-type']
plugin_name = content_type.split('/')[-1].split(';')[0].lower()
except (KeyError, IndexError):
try:
plugin_name = uri.split('/')[-1].split('.')[-1].lower()
except IndexError:
raise RuntimeError('Could not identify file type.')
tmp = tempfile.NamedTemporaryFile()
filename = '{}.{}'.format(tmp.name, plugin_name)
tmp.close()
with open(filename, 'wb') as fobj:
fobj.write(content)
return {
'filename': filename,
'encoding': encoding,
}
def get_magics(path):
"""Get file format and encoding.
The magic library is not really good at detecting text file-based format
like CSV, JSON, YAML or, XML so we only use it to detect binary format and
the encoding.
Support both file-magic and magic as both as shipped under the same name
in various distributions.
"""
if not MAGIC:
return None, None
with open(path, 'rb') as file:
header = file.read(2048)
res = magic.detect_from_content(header)
mime_to_format = {
'application/pgp': 'gpg',
'application/x-sqlite3': 'sqlite3'
}
name_to_format = {'KDBX': 'kdbx', 'openssl': 'openssl', 'PGP': 'gpg'}
frmt = mime_to_format.get(res.mime_type, None)
for name in name_to_format:
if name in res.name:
frmt = name_to_format[name]
encoding = None # res.encoding
if 'UTF-8 Unicode (with BOM)' in res.name:
encoding = 'utf-8-sig'
return frmt, encoding
def readdir(self, path, fh):
dirents = ['.', '..']
if path in '/':
for gpx in self.track_dir:
dirents.extend([gpx])
else:
tid = self.track_dir[path[1:]][0]
data = requests.get(_urlget % tid, auth=(self.user, self.password))
logging.debug(data.headers['Content-Disposition'])
filename = re.sub(r'.*filename="([^"]+?)".*', r'\1',
data.headers['Content-Disposition'])
logging.debug('filename: {}'.format(filename))
extension = re.match(r'([^.]+)(\..*)$', filename).groups()[1]
self.track_dir[path[1:]][1] = data
detected_type = magic.detect_from_content(data.content).mime_type
type2fun = {'application/x-bzip2': bz2.decompress,
'application/x-gzip': gzip.decompress,
'text/xml': bytes,
'text/plain': bytes}
conv_fun = type2fun.get(detected_type, None)
if conv_fun:
timedate_data = re.search(b'<time>([^T]+)T([^:]+:[^:]+):',
type2fun[detected_type](data.content))
date_data = timedate_data.group(1)
time_data = timedate_data.group(2)
else:
time_data = b'notime'
dirents.append('{0}_{1}_{2}{3}'.format(tid,
date_data.decode('ascii'),
time_data.decode('ascii'),
extension))
for r in dirents:
yield r
def get_content_type(response: 'Response') -> str:
"""Get content type from ``response``.
Args:
response (:class:`requests.Response`): Response object.
Returns:
The content type from ``response``.
Note:
If the ``Content-Type`` header is not defined in ``response``,
the function will utilise |magic|_ to detect its content type.
.. |Response| replace:: ``requests.Response``.
.. _Response: https://requests.readthedocs.io/en/latest/api/index.html#requests.Response
.. |magic| replace:: ``magic``
.. _magic: https://pypi.org/project/python-magic/
"""
ct_type = response.headers.get('Content-Type')
if ct_type is None:
try:
ct_type = magic.detect_from_content(response.content).mime_type
except Exception:
ct_type = '(null)'
return ct_type.casefold().split(';', maxsplit=1)[0].strip()
def get_mimetype_and_encoding_for_content(content):
"""Function that returns the mime type and the encoding associated to
a content buffer using the magic module under the hood.
Args:
content (bytes): a content buffer
Returns:
A tuple (mimetype, encoding), for instance ('text/plain', 'us-ascii'),
associated to the provided content.
"""
while True:
try:
magic_result = magic.detect_from_content(content)
mime_type = magic_result.mime_type
encoding = magic_result.encoding
break
except Exception:
# workaround an issue with the magic module who can fail
# if detect_from_content is called multiple times in
# a short amount of time
reload(magic)
return mime_type, encoding
def detect_local_source(path, content, mime_type=None, encoding=None):
# TODO: may add sample_size
filename = os.path.basename(path)
parts = filename.split('.')
extension = parts[-1] if len(parts) > 1 else None
if magic is not None:
detected = magic.detect_from_content(content)
encoding = detected.encoding or encoding
mime_name = detected.name
mime_type = detected.mime_type or mime_type
else:
encoding = chardet.detect(content)['encoding'] or encoding
mime_name = None
mime_type = mime_type or mimetypes.guess_type(filename)[0]
plugin_name = plugin_name_by_mime_type(mime_type, mime_name, extension)
if encoding == 'binary':
encoding = None
return Source(uri=path,
plugin_name=plugin_name,
encoding=encoding)
def TypeOfBuffer(buf: bytearray, default: str = None) -> str:
try:
r = magic.detect_from_content(buf)
return r.mime_type
except:
pass
return default
def _getInstalledKernel(self):
# Could we maybe remove the dependency for the "magic" module with a struct?
# http://lxr.linux.no/#linux+v2.6.39/Documentation/x86/boot.txt
# https://stackoverflow.com/a/11179559/733214
try:
len(self.cfg)
except TypeError:
raise RuntimeError('Tried to find the isKernel with no config set up and parsed')
for f in self.cfg.findall('{0}fileChecks/{0}file'.format(self.ns)):
isKernel = (True
if f.attrib.get('isKernel', 'false').lower() in ('true', '1')
else
False)
if isKernel:
self.kernelFile = f.text
if self.kernelFile:
with open(os.path.join('/boot', self.kernelFile), 'rb') as fh:
magicname = magic.detect_from_content(fh.read())
names = [i.strip().split(None, 1) for i in magicname.name.split(',') if i.strip() != '']
for n in names:
if len(n) != 2:
continue
k, v = n
# Note: this only grabs the version number.
# If we want to get e.g. the build user/machine, date, etc.,
# then we need to do a join. Shouldn't be necessary, though.
if k.lower() == 'version':
self.installedKernVer = v.split(None, 1)[0]
return()
def download_file(uri, verify_ssl):
response = requests.get(uri, verify=verify_ssl)
content = response.content
if magic is not None:
encoding = magic.detect_from_content(content).encoding
else:
encoding = response.encoding
# TODO: try to guess with uri.split('/')[-1].split('.')[-1].lower()
# TODO: try to guess with file-magic lib
try:
content_type = response.headers['content-type']
plugin_name = content_type.split('/')[-1].split(';')[0].lower()
except (KeyError, IndexError):
try:
plugin_name = uri.split('/')[-1].split('.')[-1].lower()
except IndexError:
raise RuntimeError('Could not identify file type.')
tmp = tempfile.NamedTemporaryFile()
filename = '{}.{}'.format(tmp.name, plugin_name)
tmp.close()
with open(filename, 'wb') as fobj:
fobj.write(content)
return {'filename': filename, 'encoding': encoding, }
def test_detect_from_content(self):
# differ from upstream by opening file in binary mode,
# this avoids hitting a bug in python3+libfile bindings
# see https://github.com/ahupp/python-magic/issues/152
# for a similar issue
with open(self.filename, 'rb') as fobj:
result = magic.detect_from_content(fobj.read(4096))
self.assert_result(result)
def check_mime_type(data: bytes, valid_types: List[str]) -> bool:
detected = magic.detect_from_content(data)
mime_type = cast(str, detected.mime_type)
if mime_type in valid_types:
return True
return False
def get_buffer_mime_type(buffer):
try:
if hasattr(magic, 'detect_from_content'):
# Using file-magic module: https://github.com/file/file
return magic.detect_from_content(buffer[:128]).mime_type
else:
# Using python-magic module: https://github.com/ahupp/python-magic
return magic.from_buffer(buffer[:128], mime=True)
except Exception:
return ''
def magic_type(self, data, isdata=False):
try:
if isdata:
magictype = magic.detect_from_content(data[0:512]).name
else:
magictype = magic.detect_from_filename(data).name
except NameError:
magictype = 'Error - file-magic library required.'
except Exception as e:
magictype = 'Error getting magic type - %s' % e
return magictype
def guessWithMagic(content):
result = magic.detect_from_content(content)
return result.__dict__
# #worth to look into
# #https://bitbucket.org/Telofy/utilofies/csvprofiler/0d8cdc3ae5a0a08e7fb5906d96f0d8e2284751d1/utilofies/bslib.py?at=master#cl-15
# def intelligent_decode(fname):
# """ One problem remains in the latest version of UnicodeDammit, namely
# that pages that have beautifully declared encodings but contain one
# small erroneous byte sequence somewhere will fail to be decoded with
# the mostly correct encodings, while Windows-1252 somehow succeeds, but
# completely mucks up all umlauts and ligatures. Hence I want to remove
# Windows-1252 from the potential encodings.
#
# I don't fall back on cchardet just yet.
# """
# detector = bs4.dammit.EncodingDetector(fname)
# # Fall back on forcing it to UTF-8 only if no other encodings
# # could be found. (I use override_encodings for the HTTP encoding,
# # which seems at least less reliable to me than the declared encoding.)
# potential_encodings = \
# filter(bool, [detector.sniffed_encoding, detector.declared_encoding]
# + list(detector.override_encodings)) \
# or ['utf-8']
# contains_replacement_characters = False
# tried_encodings = []
# unicode_markup = None
# original_encoding = None
# for encoding in potential_encodings:
# tried_encodings.append(encoding)
# try:
# unicode_markup = detector.markup.decode(encoding)
# except Exception as excp:
# #logger.info('Unsuccessfully tried encoding %s: %r', encoding, excp)
# print 'Unsuccessfully tried encoding %s: %r', encoding, excp
# if unicode_markup is not None:
# original_encoding = encoding
# break
# if unicode_markup is None:
# # Whatever!
# unicode_markup = detector.markup.decode(
# potential_encodings[0], 'replace')
# original_encoding = potential_encodings[0]
# contains_replacement_characters = True
# return type(b'MockDammit', (object,), {
# 'contains_replacement_characters': contains_replacement_characters,
# 'original_encoding': original_encoding,
# 'detector': detector,
# 'is_html': detector.is_html,
# 'markup': detector.markup,
# 'tried_encodings': tried_encodings,
# 'unicode_markup': unicode_markup})
def _index_content(self, fileid, path_to_file, mimetype, writer):
"""Index one file.
"""
if not mimetype in EXTRACTORS:
content = "Missing extractor for {}".format(mimetype)
else:
with open(path_to_file, 'rb') as f:
document_bytes = f.read()
magic = detect_from_content(document_bytes)
content = EXTRACTORS[mimetype](path_to_file, document_bytes, magic)
writer.add_document(fileid=fileid, content=content)
def guessWithMagic(content):
result = magic.detect_from_content(content)
return result.__dict__
# #worth to look into
# #https://bitbucket.org/Telofy/utilofies/csvprofiler/0d8cdc3ae5a0a08e7fb5906d96f0d8e2284751d1/utilofies/bslib.py?at=master#cl-15
# def intelligent_decode(fname):
# """ One problem remains in the latest version of UnicodeDammit, namely
# that pages that have beautifully declared encodings but contain one
# small erroneous byte sequence somewhere will fail to be decoded with
# the mostly correct encodings, while Windows-1252 somehow succeeds, but
# completely mucks up all umlauts and ligatures. Hence I want to remove
# Windows-1252 from the potential encodings.
#
# I don't fall back on cchardet just yet.
# """
# detector = bs4.dammit.EncodingDetector(fname)
# # Fall back on forcing it to UTF-8 only if no other encodings
# # could be found. (I use override_encodings for the HTTP encoding,
# # which seems at least less reliable to me than the declared encoding.)
# potential_encodings = \
# filter(bool, [detector.sniffed_encoding, detector.declared_encoding]
# + list(detector.override_encodings)) \
# or ['utf-8']
# contains_replacement_characters = False
# tried_encodings = []
# unicode_markup = None
# original_encoding = None
# for encoding in potential_encodings:
# tried_encodings.append(encoding)
# try:
# unicode_markup = detector.markup.decode(encoding)
# except Exception as excp:
# #logger.info('Unsuccessfully tried encoding %s: %r', encoding, excp)
# print 'Unsuccessfully tried encoding %s: %r', encoding, excp
# if unicode_markup is not None:
# original_encoding = encoding
# break
# if unicode_markup is None:
# # Whatever!
# unicode_markup = detector.markup.decode(
# potential_encodings[0], 'replace')
# original_encoding = potential_encodings[0]
# contains_replacement_characters = True
# return type(b'MockDammit', (object,), {
# 'contains_replacement_characters': contains_replacement_characters,
# 'original_encoding': original_encoding,
# 'detector': detector,
# 'is_html': detector.is_html,
# 'markup': detector.markup,
# 'tried_encodings': tried_encodings,
# 'unicode_markup': unicode_markup})
def get_file_mime(fobj) -> str:
"""
:param fobj:
- new/replace - django.core.files.uploadedfile.InMemoryUploadedFile (for newly created)
- edit (w/o changing file itself) - django.core.files.base.File
:return: file mimetype
:note: https://stackoverflow.com/questions/4853581/django-get-uploaded-file-type-mimetype
"""
pos = fobj.tell()
fobj.seek(0)
mime = magic.detect_from_content(fobj.read(1024)).mime_type
fobj.seek(pos)
return mime
def constraint(value):
if hasattr(magic, 'detect_from_content'):
# Hack for temporary compatibility with both python 'magic' modules...
mime_type = magic.detect_from_content(str(value)).mime_type
else:
mime_type = magic.from_buffer(str(value), mime=True)
if mime_type in allowed_mime_types:
return None
else:
return _("Detected data type %(detected)s. Expected %(expected)s.",
detected=mime_type,
expected=', '.join(allowed_mime_types))
def get_mime_type(data: bytes) -> typing.Union[typing.Tuple[str, str], typing.Tuple[None, None]]:
"""Get mime-type information based on the provided bytes object.
Args:
data: Binary data.
Returns:
typing.Tuple[str, str]: Identified mime information and mime-type. If **magic** is not available, returns *None, None*.
E.g. *"ELF 64-bit LSB shared object, x86-64, version 1 (SYSV)", "application/x-sharedlib"*
"""
if magic is None:
return None, None
detected = magic.detect_from_content(data)
return detected.name, detected.mime_type
def get_image_metadata_from_file(file_like):
"""
Receive a valid image file and returns a 2-tuple of two strings:
[0]: Image format (i.e. 'jpg', 'gif' or 'png')
[1]: InMemoryUploadedFile-friendly save format (i.e. 'image/jpeg')
image_format, in_memory_file_type
"""
if hasattr(magic, 'from_buffer'):
mime_type = magic.from_buffer(file_like.read(1024), mime=True)
else:
info = magic.detect_from_content(file_like.read(1024))
mime_type = info.mime_type
file_like.seek(0)
image_format = MIME_TYPE_TO_PIL_IDENTIFIER[mime_type]
return image_format, mime_type
def _analyze_file(self, path_to_file):
content = open(path_to_file, 'rb').read()
md5hash = hashlib.md5(content).hexdigest()
magic = detect_from_content(content)
filetype = magic.mime_type
orig_name = os.path.basename(path_to_file)
_, orig_ext = os.path.splitext(orig_name)
if filetype == 'application/octet-stream' and magic.name == 'Microsoft OOXML':
if orig_ext == '.pptx':
filetype = 'application/vnd.openxmlformats-officedocument.presentationml.presentation'
elif orig_ext == '.docx':
filetype = 'application/vnd.openxmlformats-officedocument.wordprocessingml.document'
elif orig_ext == '.xlsx':
filetype = 'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet'
return {'md5hash': md5hash, 'mimetype': filetype, 'text': orig_name}
def to_python(self, data):
f = super().to_python(data)
if f is None:
return None
if f.size > self.max_upload_size:
raise forms.ValidationError(_('File is too big.'), code='size')
content_type = magic.detect_from_content(f.read(1024)).mime_type
if content_type not in self.content_types:
raise forms.ValidationError(
_('Filetype not supported.'), code='content_type'
)
f.seek(0)
return f
def get_magic_content_type(input): # pylint: disable=redefined-builtin
"""Get content-type based on magic library as *bytes*
As libmagic bindings are provided via several 'magic' packages, we try them in order
"""
if magic is not None:
if hasattr(input, 'seek'):
input.seek(0)
if hasattr(input, 'read'):
input = input.read()
if hasattr(magic, 'detect_from_content'):
result = magic.detect_from_content(input) # pylint: disable=no-member
if result:
return result.mime_type
elif hasattr(magic, 'from_buffer'):
return magic.from_buffer(input, mime=True)
return None
def get_file_kernel_ver(kpath):
# Gets the version of a kernel file.
kpath = os.path.abspath(os.path.expanduser(kpath))
_kinfo = {}
with open(kpath, 'rb') as f:
_m = magic.detect_from_content(f.read())
for i in _m.name.split(','):
l = i.strip().split()
# Note: this only grabs the version number.
# If we want to get e.g. the build user/machine, date, etc.,
# then we need to join l[1:].
# We technically don't even need a dict, either. We can just iterate.
# TODO.
_kinfo[l[0].lower()] = (l[1] if len(l) > 1 else None)
if 'version' not in _kinfo:
raise RuntimeError(
'Cannot deterimine the version of {0}'.format(kpath))
else:
return (_kinfo['version'])
def _file_magic(file: File) -> FileMagic:
"""Returns the file magic namedtuple from the respective file."""
if isinstance(file, bytes):
return detect_from_content(file[:1024]) # Fix issue #350.
if isinstance(file, str):
return _file_magic(Path(file))
if isinstance(file, Path):
if file.is_file():
return detect_from_filename(str(file))
raise FileNotFoundError(str(file))
if isinstance(file, FILE_LIKE_OBJECTS):
return detect_from_fobj(file)
raise TypeError(f'Cannot read MIME type from {type(file)}.')
def extract_images(child, namespaces, start=0):
"Extract draw:image with binary-data and replace by href"
import magic
images = []
for i, image in enumerate(
child.xpath('//draw:image', namespaces=namespaces), start):
binary_data, = image.xpath('./office:binary-data',
namespaces=namespaces)
data = base64.b64decode(binary_data.text)
if hasattr(magic, 'from_buffer'):
mime_type = magic.from_buffer(data, mime=True)
else:
# Not python-magic but file-magic
mime_type = magic.detect_from_content(data).mime_type
name = 'Pictures/image%s%s' % (i, mimetypes.guess_extension(mime_type))
image.remove(binary_data)
xlink_ns = namespaces['xlink']
image.attrib['{%s}href' % xlink_ns] = name
images.append((name, data, mime_type))
return images
def get_magics(path):
"""Get file format and encoding.
The magic library is not really good at detecting text file-based format
like CSV, JSON, YAML or, XML so we only use it to detect binary format and
the encoding.
Support both file-magic and python-magic as both are shipped under the same
name in various distributions.
"""
if not MAGIC:
return None, None
with open(path, 'rb') as file:
header = file.read(2048)
if hasattr(magic, 'detect_from_content'): # file-magic
res = magic.detect_from_content(header)
mime_type = res.mime_type
magic_name = res.name
else: # python-magic
mime_type = magic.from_buffer(header, mime=True)
magic_name = magic.from_buffer(header)
mime_to_format = {
'application/pgp': 'gpg',
'application/x-sqlite3': 'sqlite3'
}
name_to_format = {'KDBX': 'kdbx', 'openssl': 'openssl', 'PGP': 'gpg'}
frmt = mime_to_format.get(mime_type, None)
for name, plain_format in name_to_format.items():
if name in magic_name:
frmt = plain_format
encoding = None
if 'UTF-8 Unicode (with BOM)' in magic_name:
encoding = 'utf-8-sig'
return frmt, encoding
def extract_dir(data, start, outdir, file_number=65536, dir_number=65536):
global header_fmt, header_len
orig_start = start
while (file_number > 0 or dir_number > 0) and start < len(data):
header = struct.unpack(header_fmt, data[start:start + header_len])
start += header_len
filename = data[start:start + header[0]]
if b'\x00' in filename:
# this means end too
break
filename = filename.decode("ascii")
start += header[0]
if magic.detect_from_content(
data[start:start + header[3]]).mime_type != "application/zlib":
# dir
if dir_number == 0:
raise Exception(
"invalid directory number for directory \"%s\"" % (outdir))
n_file_number = header[3] // 0x10000
n_dir_number = header[3] & 0xffff
n_path = os.path.join(outdir, filename)
os.mkdir(n_path)
start += extract_dir(data, start, n_path, n_file_number,
n_dir_number)
dir_number -= 1
else:
# file
if file_number == 0:
raise Exception("invalid file number for directory \"%s\"" %
(outdir))
filedata = data[start:start + header[3]]
start += header[3]
with open(os.path.join(outdir, filename), "wb") as f:
f.write(zlib.decompress(filedata))
file_number -= 1
return start - orig_start
def get_file_type(filename_or_file):
"""
Get mime_type and encoding of file `filename_or_file`.
Handles both magic libraries.
:param filename_or_file: filename or open file
:type filename_or_file: str or file
:return: mime_type and encoding of `filename_or_file`
:rtype: FileType
"""
if hasattr(filename_or_file, 'seek'):
old_pos = filename_or_file.tell()
txt = filename_or_file.read()
filename_or_file.seek(old_pos)
elif isinstance(filename_or_file, string_types):
with open(filename_or_file, 'rb') as fp:
txt = fp.read()
else:
raise ValueError(
'Argument "filename_or_file" has unknown type {!r}.'.format(
type(filename_or_file)))
if hasattr(magic, 'from_file'):
mime = magic.Magic(mime=True, mime_encoding=True).from_buffer(txt)
mime_type, charset = mime.split(';')
encoding = charset.split('=')[-1]
text = magic.Magic().from_buffer(txt)
elif hasattr(magic, 'detect_from_filename'):
fm = magic.detect_from_content(txt)
mime_type = fm.mime_type
encoding = fm.encoding
text = fm.name
else:
raise RuntimeError('Unknown version or type of "magic" library.')
# auto detect utf-8 with BOM
if encoding == 'utf-8' and txt.startswith(codecs.BOM_UTF8):
encoding = 'utf-8-sig'
return FileType(mime_type, encoding, text)
def decompressor(self):
# TODO: use mime module as fallback?
# https://docs.python.org/3/library/mimetypes.html
# VERY less-than-ideal since it won't work without self.args['logfile']
# (and has iffy detection at best, since it relies on file extensions).
# Determine what decompressor to use, if we need to.
if has_magic:
_mime = magic.detect_from_content(self.data).mime_type
self.decompress = cmprsn_map[_mime]
if self.decompress:
import importlib
decmp = importlib.import_module(self.decompress)
self.raw = decmp.decompress(self.data)
else:
# Assume that it's text and that it isn't compressed.
# We'll get a UnicodeDecodeError exception if it isn't.
pass
try:
self.raw = self.data.decode('utf-8')
except UnicodeDecodeError:
pass
self.data = self.raw
return ()
def detect_ext(self, data):
import magic
fmagic = magic.detect_from_content(data)
if fmagic.mime_type.startswith('text/'):
if fmagic.encoding == 'unknown-8bit':
ext = '.bin'
else:
text = data.decode(fmagic.encoding)
if '@return' in text or '*start' in text or '.ks' in text or '[w]' in text:
ext = '.ks'
elif '.tjs' in text or '%[' in text or '];' in text:
ext = '.tjs'
else:
ext = '.txt'
else:
ext = mimetypes.guess_extension(fmagic.mime_type)
if ext == '.jpeg':
ext = '.jpg'
elif ext == '.oga':
ext = '.ogg'
elif ext == '.asf':
ext = '.wmv'
return ext
def extract_dir(data, start, outdir, file_number=65536, dir_number=65536):
global header_fmt, header_len
orig_start = start
while (file_number > 0 or dir_number > 0) and start < len(data):
header = struct.unpack(header_fmt, data[start:start+header_len])
start += header_len
filename = data[start:start+header[0]]
if b'\x00' in filename:
# this means end too
break
filename = filename.decode("ascii")
start += header[0]
if magic.detect_from_content(data[start:start+header[3]]).mime_type != "application/zlib":
# dir
if dir_number == 0:
raise Exception("invalid directory number for directory \"%s\"" % (outdir))
n_file_number = header[3] // 0x10000
n_dir_number = header[3] & 0xffff
n_path = os.path.join(outdir, filename)
os.mkdir(n_path)
start += extract_dir(data, start, n_path, n_file_number, n_dir_number)
dir_number -= 1
else:
# file
if file_number == 0:
raise Exception("invalid file number for directory \"%s\"" % (outdir))
filedata = data[start:start+header[3]]
start += header[3]
with open(os.path.join(outdir, filename), "wb") as f:
f.write(zlib.decompress(filedata))
file_number -= 1
return start - orig_start
def test_detect_from_content(self):
with open(self.filename) as fobj:
result = magic.detect_from_content(fobj.read(4096))
self.assert_result(result)
def get_content_mimetype(content: bytes) -> str:
""" MIME Type of content retrieved from magic headers """
detected_mime = magic.detect_from_content(content).mime_type
return MIME_OVERRIDES.get(detected_mime, detected_mime)