def create_dataset_series(path: str, preprocess: Callable[[str],
str]) -> Iterable:
"""Create dataset series.
Arguments:
path: The path of the file with the data
preprocess: Preprocessor function
Returns:
The dataset series.
"""
filetyper = magic.Magic(mime=True)
filetyper.flags |= magic.MAGIC_SYMLINK
filetyper.cookie = magic.magic_open(filetyper.flags)
magic.magic_load(filetyper.cookie, None)
log("Loading {}".format(path))
file_type = filetyper.from_file(path)
if file_type.startswith('text/'):
reader = PlainTextFileReader(path)
for line in reader.read():
yield preprocess(line)
elif file_type == 'application/gzip' or file_type == 'application/x-gzip':
gzreader = GZipReader(path)
for line in gzreader.read():
yield preprocess(line)
elif file_type == 'application/octet-stream':
return np.load(path)
else:
raise Exception("Unsupported data type: {}, file {}".format(
file_type, path))
def _load_magic_file(self):
self.magic_file = ':'.join(
(constants.MAGIC_RULE_PATH, '/usr/share/file/magic.mgc'))
if self.use_cache:
self.log.info("Checking for custom magic file...")
with get_cachestore('system',
config=self.config,
datastore=self.datastore) as cache:
try:
custom_magic = "/tmp/custom.magic"
cache.download('custom_magic', custom_magic)
self.magic_file = ':'.join(
(custom_magic, '/usr/share/file/magic.mgc'))
self.log.info("Custom magic file loaded!")
except FileStoreException:
self.log.info("No custom magic file found.")
with self.lock:
self.file_type = magic.magic_open(magic.MAGIC_CONTINUE +
magic.MAGIC_RAW)
magic.magic_load(self.file_type, self.magic_file)
self.mime_type = magic.magic_open(magic.MAGIC_CONTINUE +
magic.MAGIC_RAW +
magic.MAGIC_MIME)
magic.magic_load(self.mime_type, self.magic_file)
def get_file_mime_type(name):
m = magic.Magic()
m.cookie = magic.magic_open(magic.MAGIC_NONE | magic.MAGIC_MIME
| magic.MAGIC_MIME_ENCODING)
magic.magic_load(m.cookie, None)
filetype = m.from_file(name)
return filetype
def get_mime_type(content):
# try to fixup encoding
if hasattr(magic, "open"):
m = magic.open(magic.MAGIC_MIME)
m.load()
filetype = m.buffer(content)
else:
m = magic.Magic()
m.cookie = magic.magic_open(magic.MAGIC_NONE | magic.MAGIC_MIME
| magic.MAGIC_MIME_ENCODING)
magic.magic_load(m.cookie, None)
filetype = m.from_buffer(content)
return filetype.split('; ', 1)
def get_cleaned_text_file_content(uploaded_file):
"""Read uploaded file, try to fix up encoding to UTF-8 and
transform line endings into Unix style, then return the content as
a UTF-8 string. Errors are reported as
django.core.exceptions.ValidationError exceptions."""
if not uploaded_file:
return u""
if uploaded_file.size and uploaded_file.size > 10 * 1000 * 1000:
raise ValidationError("Text file too large (size %s)." %
uploaded_file.size)
content = "".join(uploaded_file.chunks())
# try to fixup encoding
import magic
if hasattr(magic, "open"):
m = magic.open(magic.MAGIC_MIME)
m.load()
filetype = m.buffer(content)
else:
m = magic.Magic()
m.cookie = magic.magic_open(magic.MAGIC_NONE | magic.MAGIC_MIME
| magic.MAGIC_MIME_ENCODING)
magic.magic_load(m.cookie, None)
filetype = m.from_buffer(content)
if not filetype.startswith("text"):
raise ValidationError(
"Uploaded file does not appear to be a text file.")
match = re.search("charset=([\w-]+)", filetype)
if not match:
raise ValidationError("File has unknown encoding.")
encoding = match.group(1)
if "ascii" not in encoding:
try:
content = content.decode(encoding)
except Exception as e:
raise ValidationError(
"Error decoding file (%s). Try submitting with UTF-8 encoding or remove non-ASCII characters."
% str(e))
# turn line-endings into Unix style
content = content.replace("\r\n", "\n").replace("\r", "\n")
return content.encode("utf-8")
def parse_file_charset(self):
import magic
self.fd.file.seek(0)
content = self.fd.file.read(4096)
if hasattr(magic, "open"):
m = magic.open(magic.MAGIC_MIME)
m.load()
filetype = m.buffer(content)
else:
m = magic.Magic()
m.cookie = magic.magic_open(magic.MAGIC_NONE | magic.MAGIC_MIME | magic.MAGIC_MIME_ENCODING)
magic.magic_load(m.cookie, None)
filetype = m.from_buffer(content)
if not 'ascii' in filetype:
self.parsed_info.add_error('A plain text ASCII document must be submitted.')
def put_identify_custom_magic_file(**_):
"""
Save a new version of identify's custom LibMagic file
Variables:
None
Arguments:
None
Data Block:
<current custom.magic file>
Result example:
{"success": True}
"""
data = request.json.encode('utf-8')
magic_file = None
try:
with tempfile.NamedTemporaryFile(delete=False) as tmp:
magic_file = tmp.name
tmp.write(data)
try:
test = magic.magic_open(magic.MAGIC_CONTINUE + magic.MAGIC_RAW)
magic.magic_load(test, magic_file)
except magic.MagicException:
return make_api_response(
{'success': False},
"The magic file you have submitted is invalid.", 400)
finally:
if magic_file and os.path.exists(magic_file):
os.unlink(magic_file)
with forge.get_cachestore('system', config=config,
datastore=STORAGE) as cache:
if hashlib.sha256(data).hexdigest() == get_sha256_for_file(
constants.MAGIC_RULE_PATH):
cache.delete('custom_magic')
else:
cache.save('custom_magic', data, ttl=ADMIN_FILE_TTL, force=True)
# Notify components watching to reload magic file
event_sender.send('identify', 'magic')
return make_api_response({'success': True})
def calc_magic(stream):
# Missing python-magic features:
# - magic_descriptor (https://github.com/ahupp/python-magic/pull/227)
# - direct support for symlink flag
magic_cookie = magic.magic_open(magic.MAGIC_SYMLINK)
magic.magic_load(magic_cookie, None)
try:
fd_path = get_fd_path(stream)
if fd_path:
return magic.maybe_decode(magic.magic_file(magic_cookie, fd_path))
else:
# Handle BytesIO in-memory streams
stream.seek(0, os.SEEK_SET)
return magic.maybe_decode(magic.magic_buffer(magic_cookie, stream.read()))
finally:
magic.magic_close(magic_cookie)
return None
def parse_file_charset(self):
import magic
self.fd.file.seek(0)
content = self.fd.file.read(4096)
if hasattr(magic, "open"):
m = magic.open(magic.MAGIC_MIME)
m.load()
filetype = m.buffer(content)
else:
m = magic.Magic()
m.cookie = magic.magic_open(magic.MAGIC_NONE | magic.MAGIC_MIME
| magic.MAGIC_MIME_ENCODING)
magic.magic_load(m.cookie, None)
filetype = m.from_buffer(content)
if not 'ascii' in filetype:
self.parsed_info.add_error(
'A plain text document must be submitted.')
def parse_file_charset(self):
import magic
self.fd.file.seek(0)
content = self.fd.file.read()
if hasattr(magic, "open"):
m = magic.open(magic.MAGIC_MIME)
m.load()
filetype = m.buffer(content)
else:
m = magic.Magic()
m.cookie = magic.magic_open(magic.MAGIC_NONE | magic.MAGIC_MIME | magic.MAGIC_MIME_ENCODING)
magic.magic_load(m.cookie, None)
filetype = m.from_buffer(content)
if not 'ascii' in filetype and not 'utf-8' in filetype:
self.parsed_info.add_error('A plain text ASCII document is required. '
'Found an unexpected encoding: "%s". '
'You probably have one or more non-ascii characters in your file.' % filetype
)
def get_cleaned_text_file_content(uploaded_file):
"""Read uploaded file, try to fix up encoding to UTF-8 and
transform line endings into Unix style, then return the content as
a UTF-8 string. Errors are reported as
django.forms.ValidationError exceptions."""
if not uploaded_file:
return u""
if uploaded_file.size and uploaded_file.size > 10 * 1000 * 1000:
raise django.forms.ValidationError("Text file too large (size %s)." % uploaded_file.size)
content = "".join(uploaded_file.chunks())
# try to fixup encoding
import magic
if hasattr(magic, "open"):
m = magic.open(magic.MAGIC_MIME)
m.load()
filetype = m.buffer(content)
else:
m = magic.Magic()
m.cookie = magic.magic_open(magic.MAGIC_NONE | magic.MAGIC_MIME | magic.MAGIC_MIME_ENCODING)
magic.magic_load(m.cookie, None)
filetype = m.from_buffer(content)
if not filetype.startswith("text"):
raise django.forms.ValidationError("Uploaded file does not appear to be a text file.")
match = re.search("charset=([\w-]+)", filetype)
if not match:
raise django.forms.ValidationError("File has unknown encoding.")
encoding = match.group(1)
if "ascii" not in encoding:
try:
content = content.decode(encoding)
except Exception as e:
raise django.forms.ValidationError("Error decoding file (%s). Try submitting with UTF-8 encoding or remove non-ASCII characters." % str(e))
# turn line-endings into Unix style
content = content.replace("\r\n", "\n").replace("\r", "\n")
return content.encode("utf-8")
from typing import List, Any, Callable import magic # tests: lint, mypy # pylint: disable=invalid-name Reader = Callable[[List[str]], Any] FILETYPER = magic.Magic(mime=True) FILETYPER.flags |= magic.MAGIC_SYMLINK FILETYPER.cookie = magic.magic_open(FILETYPER.flags) magic.magic_load(FILETYPER.cookie, None)
custom = re.compile(r'^custom: ', re.IGNORECASE)
ssdeep_from_file = None
magic_lock = None
file_type = None
mime_type = None
if platform.system() != 'Windows':
import magic
magic_lock = threading.Lock()
file_type = magic.magic_open(magic.MAGIC_CONTINUE + magic.MAGIC_RAW)
magic.magic_load(file_type, constants.RULE_PATH)
mime_type = magic.magic_open(magic.MAGIC_CONTINUE + magic.MAGIC_MIME)
magic.magic_load(mime_type, constants.RULE_PATH)
try:
# noinspection PyUnresolvedReferences
import ssdeep # ssdeep requires apt-get cython and pip install ssdeep
ssdeep_from_file = ssdeep.hash_from_file
except ImportError:
pass # ssdeep_from_file will be None if we fail to import ssdeep.
# Translate the match object into a sub-type label.
def subtype(label):
for entry in sl_patterns:
def __init__(self,magic_file):
self.cookie = magic.magic_open(magic.MAGIC_COMPRESS|magic.MAGIC_MIME|magic.MAGIC_CONTINUE|magic.MAGIC_PRESERVE_ATIME|magic.MAGIC_ERROR|magic.MAGIC_MIME_ENCODING)
# lolololo
magic.magic_load(self.cookie,magic_file.encode('utf-8'))
self.thread = threading.currentThread()