def test_check_html():
html_text = '<p>\n<a href="url">Link</a>\n</p>\n'
assert check_html(html_text, file_path=False)
# Test a non-HTML binary string
assert not check_html(b'No HTML here\nSecond line\n', file_path=False)
with tempfile.NamedTemporaryFile(mode='w') as tmp:
tmp.write(html_text)
tmp.flush()
assert check_html(tmp.name)
# Test a non-UTF8 binary file
with tempfile.NamedTemporaryFile(mode='wb') as tmpb:
tmpb.write(b'\x1f\x8b')
tmpb.flush()
assert not check_html(tmpb.name)
def handle_uploaded_dataset_file(filename,
datatypes_registry,
ext='auto',
is_multi_byte=False):
is_valid, ext = handle_compressed_file(filename,
datatypes_registry,
ext=ext)
if not is_valid:
raise InappropriateDatasetContentError(
'The compressed uploaded file contains inappropriate content.')
if ext in AUTO_DETECT_EXTENSIONS:
ext = guess_ext(filename,
sniff_order=datatypes_registry.sniff_order,
is_multi_byte=is_multi_byte)
if check_binary(filename):
if not Binary.is_ext_unsniffable(
ext) and not datatypes_registry.get_datatype_by_extension(
ext).sniff(filename):
raise InappropriateDatasetContentError(
'The binary uploaded file contains inappropriate content.')
elif check_html(filename):
raise InappropriateDatasetContentError(
'The uploaded file contains inappropriate HTML content.')
return ext
def check_file_content_for_html_and_images(file_path):
message = ''
if checkers.check_html(file_path):
message = 'The file "%s" contains HTML content.\n' % str(file_path)
elif checkers.check_image(file_path):
message = 'The file "%s" contains image content.\n' % str(file_path)
return message
def check_file_content_for_html_and_images(file_path):
message = ''
if checkers.check_html(file_path):
message = 'The file "%s" contains HTML content.\n' % str(file_path)
elif checkers.check_image(file_path):
message = 'The file "%s" contains image content.\n' % str(file_path)
return message
def handle_uploaded_dataset_file_internal(
filename,
datatypes_registry,
ext='auto',
tmp_prefix='sniff_upload_',
tmp_dir=None,
in_place=False,
check_content=True,
is_binary=None,
auto_decompress=True,
uploaded_file_ext=None,
convert_to_posix_lines=None,
convert_spaces_to_tabs=None,
):
is_valid, ext, converted_path, compressed_type = handle_compressed_file(
filename,
datatypes_registry,
ext=ext,
tmp_prefix=tmp_prefix,
tmp_dir=tmp_dir,
in_place=in_place,
check_content=check_content,
auto_decompress=auto_decompress,
)
try:
if not is_valid:
if is_tar(converted_path):
raise InappropriateDatasetContentError('TAR file uploads are not supported')
raise InappropriateDatasetContentError('The uploaded compressed file contains invalid content')
# This needs to be checked again after decompression
is_binary = check_binary(converted_path)
if not is_binary and (convert_to_posix_lines or convert_spaces_to_tabs):
# Convert universal line endings to Posix line endings, spaces to tabs (if desired)
if convert_spaces_to_tabs:
convert_fxn = convert_newlines_sep2tabs
else:
convert_fxn = convert_newlines
line_count, _converted_path = convert_fxn(converted_path, in_place=in_place, tmp_dir=tmp_dir, tmp_prefix=tmp_prefix)
if not in_place:
if converted_path and filename != converted_path:
os.unlink(converted_path)
converted_path = _converted_path
if ext in AUTO_DETECT_EXTENSIONS:
ext = guess_ext(converted_path, sniff_order=datatypes_registry.sniff_order, is_binary=is_binary)
if not is_binary and check_content and check_html(converted_path):
raise InappropriateDatasetContentError('The uploaded file contains invalid HTML content')
except Exception:
if filename != converted_path:
os.unlink(converted_path)
raise
return ext, converted_path, compressed_type
def handle_uploaded_dataset_file_internal(
filename,
datatypes_registry,
ext='auto',
tmp_prefix='sniff_upload_',
tmp_dir=None,
in_place=False,
check_content=True,
is_binary=None,
auto_decompress=True,
uploaded_file_ext=None,
convert_to_posix_lines=None,
convert_spaces_to_tabs=None,
):
is_valid, ext, converted_path, compressed_type = handle_compressed_file(
filename,
datatypes_registry,
ext=ext,
tmp_prefix=tmp_prefix,
tmp_dir=tmp_dir,
in_place=in_place,
check_content=check_content,
auto_decompress=auto_decompress,
)
try:
if not is_valid:
if is_tar(converted_path):
raise InappropriateDatasetContentError('TAR file uploads are not supported')
raise InappropriateDatasetContentError('The uploaded compressed file contains invalid content')
# This needs to be checked again after decompression
is_binary = check_binary(converted_path)
if not is_binary and (convert_to_posix_lines or convert_spaces_to_tabs):
# Convert universal line endings to Posix line endings, spaces to tabs (if desired)
if convert_spaces_to_tabs:
convert_fxn = convert_newlines_sep2tabs
else:
convert_fxn = convert_newlines
line_count, _converted_path = convert_fxn(converted_path, in_place=in_place, tmp_dir=tmp_dir, tmp_prefix=tmp_prefix)
if not in_place:
if converted_path and filename != converted_path:
os.unlink(converted_path)
converted_path = _converted_path
if ext in AUTO_DETECT_EXTENSIONS:
ext = guess_ext(converted_path, sniff_order=datatypes_registry.sniff_order, is_binary=is_binary)
if not is_binary and check_content and check_html(converted_path):
raise InappropriateDatasetContentError('The uploaded file contains invalid HTML content')
except Exception:
if filename != converted_path:
os.unlink(converted_path)
raise
return ext, converted_path, compressed_type
def handle_uploaded_dataset_file(filename, datatypes_registry, ext='auto', is_multi_byte=False):
is_valid, ext = handle_compressed_file(filename, datatypes_registry, ext=ext)
if not is_valid:
raise InappropriateDatasetContentError('The compressed uploaded file contains inappropriate content.')
if ext in AUTO_DETECT_EXTENSIONS:
ext = guess_ext(filename, sniff_order=datatypes_registry.sniff_order, is_multi_byte=is_multi_byte)
if check_binary(filename):
if not Binary.is_ext_unsniffable(ext) and not datatypes_registry.get_datatype_by_extension(ext).sniff(filename):
raise InappropriateDatasetContentError('The binary uploaded file contains inappropriate content.')
elif check_html(filename):
raise InappropriateDatasetContentError('The uploaded file contains inappropriate HTML content.')
return ext
def is_data_index_sample_file( file_path ):
"""
Attempt to determine if a .sample file is appropriate for copying to ~/tool-data when
a tool shed repository is being installed into a Galaxy instance.
"""
# Currently most data index files are tabular, so check that first. We'll assume that
# if the file is tabular, it's ok to copy.
if is_column_based( file_path ):
return True
# If the file is any of the following, don't copy it.
if checkers.check_html( file_path ):
return False
if checkers.check_image( file_path ):
return False
if checkers.check_binary( name=file_path ):
return False
if checkers.is_bz2( file_path ):
return False
if checkers.is_gzip( file_path ):
return False
if checkers.check_zip( file_path ):
return False
# Default to copying the file if none of the above are true.
return True
def is_data_index_sample_file(file_path):
"""
Attempt to determine if a .sample file is appropriate for copying to ~/tool-data when
a tool shed repository is being installed into a Galaxy instance.
"""
# Currently most data index files are tabular, so check that first. We'll assume that
# if the file is tabular, it's ok to copy.
if is_column_based(file_path):
return True
# If the file is any of the following, don't copy it.
if checkers.check_html(file_path):
return False
if checkers.check_image(file_path):
return False
if checkers.check_binary(name=file_path):
return False
if checkers.is_bz2(file_path):
return False
if checkers.is_gzip(file_path):
return False
if checkers.is_zip(file_path):
return False
# Default to copying the file if none of the above are true.
return True
def add_file(dataset, registry, json_file, output_path):
data_type = None
line_count = None
converted_path = None
stdout = None
link_data_only = dataset.get('link_data_only', 'copy_files')
run_as_real_user = in_place = dataset.get('in_place', True)
purge_source = dataset.get('purge_source', True)
# in_place is True if there is no external chmod in place,
# however there are other instances where modifications should not occur in_place:
# when a file is added from a directory on the local file system (ftp import folder or any other path).
if dataset.type in ('server_dir', 'path_paste', 'ftp_import'):
in_place = False
check_content = dataset.get('check_content' , True)
auto_decompress = dataset.get('auto_decompress', True)
try:
ext = dataset.file_type
except AttributeError:
file_err('Unable to process uploaded file, missing file_type parameter.', dataset, json_file)
return
if dataset.type == 'url':
try:
page = urlopen(dataset.path) # page will be .close()ed by sniff methods
temp_name, dataset.is_multi_byte = sniff.stream_to_file(page, prefix='url_paste', source_encoding=util.get_charset_from_http_headers(page.headers))
except Exception as e:
file_err('Unable to fetch %s\n%s' % (dataset.path, str(e)), dataset, json_file)
return
dataset.path = temp_name
# See if we have an empty file
if not os.path.exists(dataset.path):
file_err('Uploaded temporary file (%s) does not exist.' % dataset.path, dataset, json_file)
return
if not os.path.getsize(dataset.path) > 0:
file_err('The uploaded file is empty', dataset, json_file)
return
if not dataset.type == 'url':
# Already set is_multi_byte above if type == 'url'
try:
dataset.is_multi_byte = multi_byte.is_multi_byte(codecs.open(dataset.path, 'r', 'utf-8').read(100))
except UnicodeDecodeError as e:
dataset.is_multi_byte = False
# Is dataset an image?
i_ext = get_image_ext(dataset.path)
if i_ext:
ext = i_ext
data_type = ext
# Is dataset content multi-byte?
elif dataset.is_multi_byte:
data_type = 'multi-byte char'
ext = sniff.guess_ext(dataset.path, registry.sniff_order, is_multi_byte=True)
# Is dataset content supported sniffable binary?
else:
# FIXME: This ignores the declared sniff order in datatype_conf.xml
# resulting in improper behavior
type_info = Binary.is_sniffable_binary(dataset.path)
if type_info:
data_type = type_info[0]
ext = type_info[1]
if not data_type:
root_datatype = registry.get_datatype_by_extension(dataset.file_type)
if getattr(root_datatype, 'compressed', False):
data_type = 'compressed archive'
ext = dataset.file_type
else:
# See if we have a gzipped file, which, if it passes our restrictions, we'll uncompress
is_gzipped, is_valid = check_gzip(dataset.path, check_content=check_content)
if is_gzipped and not is_valid:
file_err('The gzipped uploaded file contains inappropriate content', dataset, json_file)
return
elif is_gzipped and is_valid and auto_decompress:
if link_data_only == 'copy_files':
# We need to uncompress the temp_name file, but BAM files must remain compressed in the BGZF format
CHUNK_SIZE = 2 ** 20 # 1Mb
fd, uncompressed = tempfile.mkstemp(prefix='data_id_%s_upload_gunzip_' % dataset.dataset_id, dir=os.path.dirname(output_path), text=False)
gzipped_file = gzip.GzipFile(dataset.path, 'rb')
while 1:
try:
chunk = gzipped_file.read(CHUNK_SIZE)
except IOError:
os.close(fd)
os.remove(uncompressed)
file_err('Problem decompressing gzipped data', dataset, json_file)
return
if not chunk:
break
os.write(fd, chunk)
os.close(fd)
gzipped_file.close()
# Replace the gzipped file with the decompressed file if it's safe to do so
if not in_place:
dataset.path = uncompressed
else:
shutil.move(uncompressed, dataset.path)
os.chmod(dataset.path, 0o644)
dataset.name = dataset.name.rstrip('.gz')
data_type = 'gzip'
if not data_type:
# See if we have a bz2 file, much like gzip
is_bzipped, is_valid = check_bz2(dataset.path, check_content)
if is_bzipped and not is_valid:
file_err('The gzipped uploaded file contains inappropriate content', dataset, json_file)
return
elif is_bzipped and is_valid and auto_decompress:
if link_data_only == 'copy_files':
# We need to uncompress the temp_name file
CHUNK_SIZE = 2 ** 20 # 1Mb
fd, uncompressed = tempfile.mkstemp(prefix='data_id_%s_upload_bunzip2_' % dataset.dataset_id, dir=os.path.dirname(output_path), text=False)
bzipped_file = bz2.BZ2File(dataset.path, 'rb')
while 1:
try:
chunk = bzipped_file.read(CHUNK_SIZE)
except IOError:
os.close(fd)
os.remove(uncompressed)
file_err('Problem decompressing bz2 compressed data', dataset, json_file)
return
if not chunk:
break
os.write(fd, chunk)
os.close(fd)
bzipped_file.close()
# Replace the bzipped file with the decompressed file if it's safe to do so
if not in_place:
dataset.path = uncompressed
else:
shutil.move(uncompressed, dataset.path)
os.chmod(dataset.path, 0o644)
dataset.name = dataset.name.rstrip('.bz2')
data_type = 'bz2'
if not data_type:
# See if we have a zip archive
is_zipped = check_zip(dataset.path)
if is_zipped and auto_decompress:
if link_data_only == 'copy_files':
CHUNK_SIZE = 2 ** 20 # 1Mb
uncompressed = None
uncompressed_name = None
unzipped = False
z = zipfile.ZipFile(dataset.path)
for name in z.namelist():
if name.endswith('/'):
continue
if unzipped:
stdout = 'ZIP file contained more than one file, only the first file was added to Galaxy.'
break
fd, uncompressed = tempfile.mkstemp(prefix='data_id_%s_upload_zip_' % dataset.dataset_id, dir=os.path.dirname(output_path), text=False)
if sys.version_info[:2] >= (2, 6):
zipped_file = z.open(name)
while 1:
try:
chunk = zipped_file.read(CHUNK_SIZE)
except IOError:
os.close(fd)
os.remove(uncompressed)
file_err('Problem decompressing zipped data', dataset, json_file)
return
if not chunk:
break
os.write(fd, chunk)
os.close(fd)
zipped_file.close()
uncompressed_name = name
unzipped = True
else:
# python < 2.5 doesn't have a way to read members in chunks(!)
try:
outfile = open(uncompressed, 'wb')
outfile.write(z.read(name))
outfile.close()
uncompressed_name = name
unzipped = True
except IOError:
os.close(fd)
os.remove(uncompressed)
file_err('Problem decompressing zipped data', dataset, json_file)
return
z.close()
# Replace the zipped file with the decompressed file if it's safe to do so
if uncompressed is not None:
if not in_place:
dataset.path = uncompressed
else:
shutil.move(uncompressed, dataset.path)
os.chmod(dataset.path, 0o644)
dataset.name = uncompressed_name
data_type = 'zip'
if not data_type:
# TODO refactor this logic. check_binary isn't guaranteed to be
# correct since it only looks at whether the first 100 chars are
# printable or not. If someone specifies a known unsniffable
# binary datatype and check_binary fails, the file gets mangled.
if check_binary(dataset.path) or Binary.is_ext_unsniffable(dataset.file_type):
# We have a binary dataset, but it is not Bam, Sff or Pdf
data_type = 'binary'
# binary_ok = False
parts = dataset.name.split(".")
if len(parts) > 1:
ext = parts[-1].strip().lower()
if check_content and not Binary.is_ext_unsniffable(ext):
file_err('The uploaded binary file contains inappropriate content', dataset, json_file)
return
elif Binary.is_ext_unsniffable(ext) and dataset.file_type != ext:
err_msg = "You must manually set the 'File Format' to '%s' when uploading %s files." % (ext.capitalize(), ext)
file_err(err_msg, dataset, json_file)
return
if not data_type:
# We must have a text file
if check_content and check_html(dataset.path):
file_err('The uploaded file contains inappropriate HTML content', dataset, json_file)
return
if data_type != 'binary':
if link_data_only == 'copy_files' and data_type not in ('gzip', 'bz2', 'zip'):
# Convert universal line endings to Posix line endings if to_posix_lines is True
# and the data is not binary or gzip-, bz2- or zip-compressed.
if dataset.to_posix_lines:
tmpdir = output_adjacent_tmpdir(output_path)
tmp_prefix = 'data_id_%s_convert_' % dataset.dataset_id
if dataset.space_to_tab:
line_count, converted_path = sniff.convert_newlines_sep2tabs(dataset.path, in_place=in_place, tmp_dir=tmpdir, tmp_prefix=tmp_prefix)
else:
line_count, converted_path = sniff.convert_newlines(dataset.path, in_place=in_place, tmp_dir=tmpdir, tmp_prefix=tmp_prefix)
if dataset.file_type == 'auto':
ext = sniff.guess_ext(dataset.path, registry.sniff_order)
else:
ext = dataset.file_type
data_type = ext
# Save job info for the framework
if ext == 'auto' and data_type == 'binary':
ext = 'data'
if ext == 'auto' and dataset.ext:
ext = dataset.ext
if ext == 'auto':
ext = 'data'
datatype = registry.get_datatype_by_extension(ext)
if dataset.type in ('server_dir', 'path_paste') and link_data_only == 'link_to_files':
# Never alter a file that will not be copied to Galaxy's local file store.
if datatype.dataset_content_needs_grooming(dataset.path):
err_msg = 'The uploaded files need grooming, so change your <b>Copy data into Galaxy?</b> selection to be ' + \
'<b>Copy files into Galaxy</b> instead of <b>Link to files without copying into Galaxy</b> so grooming can be performed.'
file_err(err_msg, dataset, json_file)
return
if link_data_only == 'copy_files' and converted_path:
# Move the dataset to its "real" path
try:
shutil.move(converted_path, output_path)
except OSError as e:
# We may not have permission to remove converted_path
if e.errno != errno.EACCES:
raise
elif link_data_only == 'copy_files':
if purge_source and not run_as_real_user:
# if the upload tool runs as a real user the real user
# can't move dataset.path as this path is owned by galaxy.
shutil.move(dataset.path, output_path)
else:
shutil.copy(dataset.path, output_path)
# Write the job info
stdout = stdout or 'uploaded %s file' % data_type
info = dict(type='dataset',
dataset_id=dataset.dataset_id,
ext=ext,
stdout=stdout,
name=dataset.name,
line_count=line_count)
if dataset.get('uuid', None) is not None:
info['uuid'] = dataset.get('uuid')
json_file.write(dumps(info) + "\n")
if link_data_only == 'copy_files' and datatype and datatype.dataset_content_needs_grooming(output_path):
# Groom the dataset content if necessary
datatype.groom_dataset_content(output_path)
# We have a binary dataset, but it is not Bam, Sff or Pdf
data_type = 'binary'
# binary_ok = False
parts = dataset.name.split( "." )
if len( parts ) > 1:
ext = parts[-1].strip().lower()
if not Binary.is_ext_unsniffable(ext):
file_err( 'The uploaded binary file contains inappropriate content', dataset, json_file )
return
elif Binary.is_ext_unsniffable(ext) and dataset.file_type != ext:
err_msg = "You must manually set the 'File Format' to '%s' when uploading %s files." % ( ext.capitalize(), ext )
file_err( err_msg, dataset, json_file )
return
if not data_type:
# We must have a text file
if check_html( dataset.path ):
file_err( 'The uploaded file contains inappropriate HTML content', dataset, json_file )
return
if data_type != 'binary':
if link_data_only == 'copy_files':
if dataset.type in ( 'server_dir', 'path_paste' ) and data_type not in [ 'gzip', 'bz2', 'zip' ]:
in_place = False
# Convert universal line endings to Posix line endings, but allow the user to turn it off,
# so that is becomes possible to upload gzip, bz2 or zip files with binary data without
# corrupting the content of those files.
if dataset.to_posix_lines:
tmpdir = output_adjacent_tmpdir( output_path )
tmp_prefix = 'data_id_%s_convert_' % dataset.dataset_id
if dataset.space_to_tab:
line_count, converted_path = sniff.convert_newlines_sep2tabs( dataset.path, in_place=in_place, tmp_dir=tmpdir, tmp_prefix=tmp_prefix )
else:
def _resolve_src(item):
converted_path = None
name, path = _has_src_to_path(item)
dbkey = item.get("dbkey", "?")
requested_ext = item.get("ext", "auto")
info = item.get("info", None)
object_id = item.get("object_id", None)
link_data_only = upload_config.link_data_only
if "link_data_only" in item:
# Allow overriding this on a per file basis.
link_data_only = _link_data_only(item)
to_posix_lines = upload_config.get_option(item, "to_posix_lines")
space_to_tab = upload_config.get_option(item, "space_to_tab")
in_place = item.get("in_place", False)
purge_source = item.get("purge_source", True)
# Follow upload.py logic but without the auto-decompress logic.
registry = upload_config.registry
check_content = upload_config.check_content
data_type, ext = None, requested_ext
is_binary = check_binary(path)
if is_binary:
data_type, ext = handle_sniffable_binary_check(
data_type, ext, path, registry)
if data_type is None:
root_datatype = registry.get_datatype_by_extension(ext)
if getattr(root_datatype, 'compressed', False):
data_type = 'compressed archive'
ext = ext
elif is_binary:
data_type, ext = handle_unsniffable_binary_check(
data_type, ext, path, name, is_binary, requested_ext,
check_content, registry)
if not data_type and check_content and check_html(path):
raise UploadProblemException(
'The uploaded file contains inappropriate HTML content')
if data_type != 'binary':
if not link_data_only:
if to_posix_lines:
if space_to_tab:
line_count, converted_path = sniff.convert_newlines_sep2tabs(
path, in_place=in_place, tmp_dir=".")
else:
line_count, converted_path = sniff.convert_newlines(
path, in_place=in_place, tmp_dir=".")
else:
if space_to_tab:
line_count, converted_path = sniff.sep2tabs(
path, in_place=in_place, tmp_dir=".")
if requested_ext == 'auto':
ext = sniff.guess_ext(converted_path or path,
registry.sniff_order)
else:
ext = requested_ext
data_type = ext
if ext == 'auto' and data_type == 'binary':
ext = 'data'
if ext == 'auto' and requested_ext:
ext = requested_ext
if ext == 'auto':
ext = 'data'
datatype = registry.get_datatype_by_extension(ext)
if link_data_only:
# Never alter a file that will not be copied to Galaxy's local file store.
if datatype.dataset_content_needs_grooming(path):
err_msg = 'The uploaded files need grooming, so change your <b>Copy data into Galaxy?</b> selection to be ' + \
'<b>Copy files into Galaxy</b> instead of <b>Link to files without copying into Galaxy</b> so grooming can be performed.'
raise UploadProblemException(err_msg)
# If this file is not in the workdir make sure it gets there.
if not link_data_only and converted_path:
path = upload_config.ensure_in_working_directory(
converted_path, purge_source, in_place)
elif not link_data_only:
path = upload_config.ensure_in_working_directory(
path, purge_source, in_place)
if not link_data_only and datatype and datatype.dataset_content_needs_grooming(
path):
# Groom the dataset content if necessary
datatype.groom_dataset_content(path)
rval = {
"name": name,
"filename": path,
"dbkey": dbkey,
"ext": ext,
"link_data_only": link_data_only
}
if info is not None:
rval["info"] = info
if object_id is not None:
rval["object_id"] = object_id
return rval
def add_file(dataset, registry, json_file, output_path):
data_type = None
line_count = None
converted_path = None
stdout = None
link_data_only = dataset.get('link_data_only', 'copy_files') != 'copy_files'
# run_as_real_user is estimated from galaxy config (external chmod indicated of inputs executed)
# If this is True we always purge supplied upload inputs so they are cleaned up and we reuse their
# paths during data conversions since this user already owns that path.
# Older in_place check for upload jobs created before 18.01, TODO remove in 19.XX. xref #5206
run_as_real_user = dataset.get('run_as_real_user', False) or dataset.get("in_place", False)
# purge_source defaults to True unless this is an FTP import and
# ftp_upload_purge has been overridden to False in Galaxy's config.
# We set purge_source to False if:
# - the job does not have write access to the file, e.g. when running as the
# real user
# - the files are uploaded from external paths.
purge_source = dataset.get('purge_source', True) and not run_as_real_user and dataset.type not in ('server_dir', 'path_paste')
# in_place is True unless we are running as a real user or importing external paths (i.e.
# this is a real upload and not a path paste or ftp import).
# in_place should always be False if running as real user because the uploaded file will
# be owned by Galaxy and not the user and it should be False for external paths so Galaxy doesn't
# modify files not controlled by Galaxy.
in_place = not run_as_real_user and dataset.type not in ('server_dir', 'path_paste', 'ftp_import')
# Base on the check_upload_content Galaxy config option and on by default, this enables some
# security related checks on the uploaded content, but can prevent uploads from working in some cases.
check_content = dataset.get('check_content' , True)
# auto_decompress is a request flag that can be swapped off to prevent Galaxy from automatically
# decompressing archive files before sniffing.
auto_decompress = dataset.get('auto_decompress', True)
try:
ext = dataset.file_type
except AttributeError:
raise UploadProblemException('Unable to process uploaded file, missing file_type parameter.')
if dataset.type == 'url':
try:
page = urlopen(dataset.path) # page will be .close()ed by sniff methods
temp_name = sniff.stream_to_file(page, prefix='url_paste', source_encoding=util.get_charset_from_http_headers(page.headers))
except Exception as e:
raise UploadProblemException('Unable to fetch %s\n%s' % (dataset.path, str(e)))
dataset.path = temp_name
# See if we have an empty file
if not os.path.exists(dataset.path):
raise UploadProblemException('Uploaded temporary file (%s) does not exist.' % dataset.path)
if not os.path.getsize(dataset.path) > 0:
raise UploadProblemException('The uploaded file is empty')
# Is dataset content supported sniffable binary?
is_binary = check_binary(dataset.path)
if is_binary:
# Sniff the data type
guessed_ext = sniff.guess_ext(dataset.path, registry.sniff_order)
# Set data_type only if guessed_ext is a binary datatype
datatype = registry.get_datatype_by_extension(guessed_ext)
if isinstance(datatype, Binary):
data_type = guessed_ext
ext = guessed_ext
if not data_type:
root_datatype = registry.get_datatype_by_extension(dataset.file_type)
if getattr(root_datatype, 'compressed', False):
data_type = 'compressed archive'
ext = dataset.file_type
else:
# See if we have a gzipped file, which, if it passes our restrictions, we'll uncompress
is_gzipped, is_valid = check_gzip(dataset.path, check_content=check_content)
if is_gzipped and not is_valid:
raise UploadProblemException('The gzipped uploaded file contains inappropriate content')
elif is_gzipped and is_valid and auto_decompress:
if not link_data_only:
# We need to uncompress the temp_name file, but BAM files must remain compressed in the BGZF format
CHUNK_SIZE = 2 ** 20 # 1Mb
fd, uncompressed = tempfile.mkstemp(prefix='data_id_%s_upload_gunzip_' % dataset.dataset_id, dir=os.path.dirname(output_path), text=False)
gzipped_file = gzip.GzipFile(dataset.path, 'rb')
while 1:
try:
chunk = gzipped_file.read(CHUNK_SIZE)
except IOError:
os.close(fd)
os.remove(uncompressed)
raise UploadProblemException('Problem decompressing gzipped data')
if not chunk:
break
os.write(fd, chunk)
os.close(fd)
gzipped_file.close()
# Replace the gzipped file with the decompressed file if it's safe to do so
if not in_place:
dataset.path = uncompressed
else:
shutil.move(uncompressed, dataset.path)
os.chmod(dataset.path, 0o644)
dataset.name = dataset.name.rstrip('.gz')
data_type = 'gzip'
if not data_type:
# See if we have a bz2 file, much like gzip
is_bzipped, is_valid = check_bz2(dataset.path, check_content)
if is_bzipped and not is_valid:
raise UploadProblemException('The gzipped uploaded file contains inappropriate content')
elif is_bzipped and is_valid and auto_decompress:
if not link_data_only:
# We need to uncompress the temp_name file
CHUNK_SIZE = 2 ** 20 # 1Mb
fd, uncompressed = tempfile.mkstemp(prefix='data_id_%s_upload_bunzip2_' % dataset.dataset_id, dir=os.path.dirname(output_path), text=False)
bzipped_file = bz2.BZ2File(dataset.path, 'rb')
while 1:
try:
chunk = bzipped_file.read(CHUNK_SIZE)
except IOError:
os.close(fd)
os.remove(uncompressed)
raise UploadProblemException('Problem decompressing bz2 compressed data')
if not chunk:
break
os.write(fd, chunk)
os.close(fd)
bzipped_file.close()
# Replace the bzipped file with the decompressed file if it's safe to do so
if not in_place:
dataset.path = uncompressed
else:
shutil.move(uncompressed, dataset.path)
os.chmod(dataset.path, 0o644)
dataset.name = dataset.name.rstrip('.bz2')
data_type = 'bz2'
if not data_type:
# See if we have a zip archive
is_zipped = check_zip(dataset.path)
if is_zipped and auto_decompress:
if not link_data_only:
CHUNK_SIZE = 2 ** 20 # 1Mb
uncompressed = None
uncompressed_name = None
unzipped = False
z = zipfile.ZipFile(dataset.path)
for name in z.namelist():
if name.endswith('/'):
continue
if unzipped:
stdout = 'ZIP file contained more than one file, only the first file was added to Galaxy.'
break
fd, uncompressed = tempfile.mkstemp(prefix='data_id_%s_upload_zip_' % dataset.dataset_id, dir=os.path.dirname(output_path), text=False)
if sys.version_info[:2] >= (2, 6):
zipped_file = z.open(name)
while 1:
try:
chunk = zipped_file.read(CHUNK_SIZE)
except IOError:
os.close(fd)
os.remove(uncompressed)
raise UploadProblemException('Problem decompressing zipped data')
if not chunk:
break
os.write(fd, chunk)
os.close(fd)
zipped_file.close()
uncompressed_name = name
unzipped = True
else:
# python < 2.5 doesn't have a way to read members in chunks(!)
try:
with open(uncompressed, 'wb') as outfile:
outfile.write(z.read(name))
uncompressed_name = name
unzipped = True
except IOError:
os.close(fd)
os.remove(uncompressed)
raise UploadProblemException('Problem decompressing zipped data')
z.close()
# Replace the zipped file with the decompressed file if it's safe to do so
if uncompressed is not None:
if not in_place:
dataset.path = uncompressed
else:
shutil.move(uncompressed, dataset.path)
os.chmod(dataset.path, 0o644)
dataset.name = uncompressed_name
data_type = 'zip'
if not data_type:
if is_binary or registry.is_extension_unsniffable_binary(dataset.file_type):
# We have a binary dataset, but it is not Bam, Sff or Pdf
data_type = 'binary'
parts = dataset.name.split(".")
if len(parts) > 1:
ext = parts[-1].strip().lower()
is_ext_unsniffable_binary = registry.is_extension_unsniffable_binary(ext)
if check_content and not is_ext_unsniffable_binary:
raise UploadProblemException('The uploaded binary file contains inappropriate content')
elif is_ext_unsniffable_binary and dataset.file_type != ext:
err_msg = "You must manually set the 'File Format' to '%s' when uploading %s files." % (ext, ext)
raise UploadProblemException(err_msg)
if not data_type:
# We must have a text file
if check_content and check_html(dataset.path):
raise UploadProblemException('The uploaded file contains inappropriate HTML content')
if data_type != 'binary':
if not link_data_only and data_type not in ('gzip', 'bz2', 'zip'):
# Convert universal line endings to Posix line endings if to_posix_lines is True
# and the data is not binary or gzip-, bz2- or zip-compressed.
if dataset.to_posix_lines:
tmpdir = output_adjacent_tmpdir(output_path)
tmp_prefix = 'data_id_%s_convert_' % dataset.dataset_id
if dataset.space_to_tab:
line_count, converted_path = sniff.convert_newlines_sep2tabs(dataset.path, in_place=in_place, tmp_dir=tmpdir, tmp_prefix=tmp_prefix)
else:
line_count, converted_path = sniff.convert_newlines(dataset.path, in_place=in_place, tmp_dir=tmpdir, tmp_prefix=tmp_prefix)
if dataset.file_type == 'auto':
ext = sniff.guess_ext(converted_path or dataset.path, registry.sniff_order)
else:
ext = dataset.file_type
data_type = ext
# Save job info for the framework
if ext == 'auto' and data_type == 'binary':
ext = 'data'
if ext == 'auto' and dataset.ext:
ext = dataset.ext
if ext == 'auto':
ext = 'data'
datatype = registry.get_datatype_by_extension(ext)
if dataset.type in ('server_dir', 'path_paste') and link_data_only:
# Never alter a file that will not be copied to Galaxy's local file store.
if datatype.dataset_content_needs_grooming(dataset.path):
err_msg = 'The uploaded files need grooming, so change your <b>Copy data into Galaxy?</b> selection to be ' + \
'<b>Copy files into Galaxy</b> instead of <b>Link to files without copying into Galaxy</b> so grooming can be performed.'
raise UploadProblemException(err_msg)
if not link_data_only and converted_path:
# Move the dataset to its "real" path
try:
shutil.move(converted_path, output_path)
except OSError as e:
# We may not have permission to remove converted_path
if e.errno != errno.EACCES:
raise
elif not link_data_only:
if purge_source:
shutil.move(dataset.path, output_path)
else:
shutil.copy(dataset.path, output_path)
# Write the job info
stdout = stdout or 'uploaded %s file' % data_type
info = dict(type='dataset',
dataset_id=dataset.dataset_id,
ext=ext,
stdout=stdout,
name=dataset.name,
line_count=line_count)
if dataset.get('uuid', None) is not None:
info['uuid'] = dataset.get('uuid')
json_file.write(dumps(info) + "\n")
if not link_data_only and datatype and datatype.dataset_content_needs_grooming(output_path):
# Groom the dataset content if necessary
datatype.groom_dataset_content(output_path)
def handle_uploaded_dataset_file_internal(
filename: str,
datatypes_registry,
ext: str = 'auto',
tmp_prefix: Optional[str] = 'sniff_upload_',
tmp_dir: Optional[str] = None,
in_place: bool = False,
check_content: bool = True,
is_binary: Optional[bool] = None,
auto_decompress: bool = True,
uploaded_file_ext: Optional[str] = None,
convert_to_posix_lines: Optional[bool] = None,
convert_spaces_to_tabs: Optional[bool] = None,
) -> HandleUploadedDatasetFileInternalResponse:
is_valid, ext, converted_path, compressed_type = handle_compressed_file(
filename,
datatypes_registry,
ext=ext,
tmp_prefix=tmp_prefix,
tmp_dir=tmp_dir,
in_place=in_place,
check_content=check_content,
auto_decompress=auto_decompress,
)
converted_newlines = False
converted_spaces = False
try:
if not is_valid:
if is_tar(converted_path):
raise InappropriateDatasetContentError(
'TAR file uploads are not supported')
raise InappropriateDatasetContentError(
'The uploaded compressed file contains invalid content')
# This needs to be checked again after decompression
is_binary = check_binary(converted_path)
guessed_ext = ext
if ext in AUTO_DETECT_EXTENSIONS:
guessed_ext = guess_ext(converted_path,
sniff_order=datatypes_registry.sniff_order,
is_binary=is_binary)
guessed_datatype = datatypes_registry.get_datatype_by_extension(
guessed_ext)
if not is_binary and guessed_datatype.is_binary:
# It's possible to have a datatype that is binary but not within the first 1024 bytes,
# so check_binary might return a false negative. This is for instance true for PDF files
is_binary = True
if not is_binary and (convert_to_posix_lines
or convert_spaces_to_tabs):
# Convert universal line endings to Posix line endings, spaces to tabs (if desired)
convert_fxn = convert_function(convert_to_posix_lines,
convert_spaces_to_tabs)
line_count, _converted_path, converted_newlines, converted_spaces = convert_fxn(
converted_path,
in_place=in_place,
tmp_dir=tmp_dir,
tmp_prefix=tmp_prefix)
if not in_place:
if converted_path and filename != converted_path:
os.unlink(converted_path)
assert _converted_path
converted_path = _converted_path
if ext in AUTO_DETECT_EXTENSIONS:
ext = guess_ext(converted_path,
sniff_order=datatypes_registry.sniff_order,
is_binary=is_binary)
else:
ext = guessed_ext
if not is_binary and check_content and check_html(converted_path):
raise InappropriateDatasetContentError(
'The uploaded file contains invalid HTML content')
except Exception:
if filename != converted_path:
os.unlink(converted_path)
raise
return HandleUploadedDatasetFileInternalResponse(ext, converted_path,
compressed_type,
converted_newlines,
converted_spaces)
def add_file(dataset, registry, json_file, output_path):
data_type = None
line_count = None
converted_path = None
stdout = None
link_data_only = dataset.get('link_data_only', 'copy_files')
in_place = dataset.get('in_place', True)
purge_source = dataset.get('purge_source', True)
try:
ext = dataset.file_type
except AttributeError:
file_err(
'Unable to process uploaded file, missing file_type parameter.',
dataset, json_file)
return
if dataset.type == 'url':
try:
page = urlopen(
dataset.path) # page will be .close()ed by sniff methods
temp_name, dataset.is_multi_byte = sniff.stream_to_file(
page,
prefix='url_paste',
source_encoding=util.get_charset_from_http_headers(
page.headers))
except Exception as e:
file_err('Unable to fetch %s\n%s' % (dataset.path, str(e)),
dataset, json_file)
return
dataset.path = temp_name
# See if we have an empty file
if not os.path.exists(dataset.path):
file_err('Uploaded temporary file (%s) does not exist.' % dataset.path,
dataset, json_file)
return
if not os.path.getsize(dataset.path) > 0:
file_err('The uploaded file is empty', dataset, json_file)
return
if not dataset.type == 'url':
# Already set is_multi_byte above if type == 'url'
try:
dataset.is_multi_byte = multi_byte.is_multi_byte(
codecs.open(dataset.path, 'r', 'utf-8').read(100))
except UnicodeDecodeError as e:
dataset.is_multi_byte = False
# Is dataset an image?
i_ext = get_image_ext(dataset.path)
if i_ext:
ext = i_ext
data_type = ext
# Is dataset content multi-byte?
elif dataset.is_multi_byte:
data_type = 'multi-byte char'
ext = sniff.guess_ext(dataset.path,
registry.sniff_order,
is_multi_byte=True)
# Is dataset content supported sniffable binary?
else:
# FIXME: This ignores the declared sniff order in datatype_conf.xml
# resulting in improper behavior
type_info = Binary.is_sniffable_binary(dataset.path)
if type_info:
data_type = type_info[0]
ext = type_info[1]
if not data_type:
root_datatype = registry.get_datatype_by_extension(dataset.file_type)
if getattr(root_datatype, 'compressed', False):
data_type = 'compressed archive'
ext = dataset.file_type
else:
# See if we have a gzipped file, which, if it passes our restrictions, we'll uncompress
is_gzipped, is_valid = check_gzip(dataset.path)
if is_gzipped and not is_valid:
file_err(
'The gzipped uploaded file contains inappropriate content',
dataset, json_file)
return
elif is_gzipped and is_valid:
if link_data_only == 'copy_files':
# We need to uncompress the temp_name file, but BAM files must remain compressed in the BGZF format
CHUNK_SIZE = 2**20 # 1Mb
fd, uncompressed = tempfile.mkstemp(
prefix='data_id_%s_upload_gunzip_' %
dataset.dataset_id,
dir=os.path.dirname(output_path),
text=False)
gzipped_file = gzip.GzipFile(dataset.path, 'rb')
while 1:
try:
chunk = gzipped_file.read(CHUNK_SIZE)
except IOError:
os.close(fd)
os.remove(uncompressed)
file_err('Problem decompressing gzipped data',
dataset, json_file)
return
if not chunk:
break
os.write(fd, chunk)
os.close(fd)
gzipped_file.close()
# Replace the gzipped file with the decompressed file if it's safe to do so
if dataset.type in ('server_dir',
'path_paste') or not in_place:
dataset.path = uncompressed
else:
shutil.move(uncompressed, dataset.path)
os.chmod(dataset.path, 0o644)
dataset.name = dataset.name.rstrip('.gz')
data_type = 'gzip'
if not data_type and bz2 is not None:
# See if we have a bz2 file, much like gzip
is_bzipped, is_valid = check_bz2(dataset.path)
if is_bzipped and not is_valid:
file_err(
'The gzipped uploaded file contains inappropriate content',
dataset, json_file)
return
elif is_bzipped and is_valid:
if link_data_only == 'copy_files':
# We need to uncompress the temp_name file
CHUNK_SIZE = 2**20 # 1Mb
fd, uncompressed = tempfile.mkstemp(
prefix='data_id_%s_upload_bunzip2_' %
dataset.dataset_id,
dir=os.path.dirname(output_path),
text=False)
bzipped_file = bz2.BZ2File(dataset.path, 'rb')
while 1:
try:
chunk = bzipped_file.read(CHUNK_SIZE)
except IOError:
os.close(fd)
os.remove(uncompressed)
file_err(
'Problem decompressing bz2 compressed data',
dataset, json_file)
return
if not chunk:
break
os.write(fd, chunk)
os.close(fd)
bzipped_file.close()
# Replace the bzipped file with the decompressed file if it's safe to do so
if dataset.type in ('server_dir',
'path_paste') or not in_place:
dataset.path = uncompressed
else:
shutil.move(uncompressed, dataset.path)
os.chmod(dataset.path, 0o644)
dataset.name = dataset.name.rstrip('.bz2')
data_type = 'bz2'
if not data_type:
# See if we have a zip archive
is_zipped = check_zip(dataset.path)
if is_zipped:
if link_data_only == 'copy_files':
CHUNK_SIZE = 2**20 # 1Mb
uncompressed = None
uncompressed_name = None
unzipped = False
z = zipfile.ZipFile(dataset.path)
for name in z.namelist():
if name.endswith('/'):
continue
if unzipped:
stdout = 'ZIP file contained more than one file, only the first file was added to Galaxy.'
break
fd, uncompressed = tempfile.mkstemp(
prefix='data_id_%s_upload_zip_' %
dataset.dataset_id,
dir=os.path.dirname(output_path),
text=False)
if sys.version_info[:2] >= (2, 6):
zipped_file = z.open(name)
while 1:
try:
chunk = zipped_file.read(CHUNK_SIZE)
except IOError:
os.close(fd)
os.remove(uncompressed)
file_err(
'Problem decompressing zipped data',
dataset, json_file)
return
if not chunk:
break
os.write(fd, chunk)
os.close(fd)
zipped_file.close()
uncompressed_name = name
unzipped = True
else:
# python < 2.5 doesn't have a way to read members in chunks(!)
try:
outfile = open(uncompressed, 'wb')
outfile.write(z.read(name))
outfile.close()
uncompressed_name = name
unzipped = True
except IOError:
os.close(fd)
os.remove(uncompressed)
file_err(
'Problem decompressing zipped data',
dataset, json_file)
return
z.close()
# Replace the zipped file with the decompressed file if it's safe to do so
if uncompressed is not None:
if dataset.type in ('server_dir',
'path_paste') or not in_place:
dataset.path = uncompressed
else:
shutil.move(uncompressed, dataset.path)
os.chmod(dataset.path, 0o644)
dataset.name = uncompressed_name
data_type = 'zip'
if not data_type:
# TODO refactor this logic. check_binary isn't guaranteed to be
# correct since it only looks at whether the first 100 chars are
# printable or not. If someone specifies a known unsniffable
# binary datatype and check_binary fails, the file gets mangled.
if check_binary(dataset.path) or Binary.is_ext_unsniffable(
dataset.file_type):
# We have a binary dataset, but it is not Bam, Sff or Pdf
data_type = 'binary'
# binary_ok = False
parts = dataset.name.split(".")
if len(parts) > 1:
ext = parts[-1].strip().lower()
if not Binary.is_ext_unsniffable(ext):
file_err(
'The uploaded binary file contains inappropriate content',
dataset, json_file)
return
elif Binary.is_ext_unsniffable(
ext) and dataset.file_type != ext:
err_msg = "You must manually set the 'File Format' to '%s' when uploading %s files." % (
ext.capitalize(), ext)
file_err(err_msg, dataset, json_file)
return
if not data_type:
# We must have a text file
if check_html(dataset.path):
file_err(
'The uploaded file contains inappropriate HTML content',
dataset, json_file)
return
if data_type != 'binary':
if link_data_only == 'copy_files':
if dataset.type in ('server_dir',
'path_paste') and data_type not in [
'gzip', 'bz2', 'zip'
]:
in_place = False
# Convert universal line endings to Posix line endings, but allow the user to turn it off,
# so that is becomes possible to upload gzip, bz2 or zip files with binary data without
# corrupting the content of those files.
if dataset.to_posix_lines:
tmpdir = output_adjacent_tmpdir(output_path)
tmp_prefix = 'data_id_%s_convert_' % dataset.dataset_id
if dataset.space_to_tab:
line_count, converted_path = sniff.convert_newlines_sep2tabs(
dataset.path,
in_place=in_place,
tmp_dir=tmpdir,
tmp_prefix=tmp_prefix)
else:
line_count, converted_path = sniff.convert_newlines(
dataset.path,
in_place=in_place,
tmp_dir=tmpdir,
tmp_prefix=tmp_prefix)
if dataset.file_type == 'auto':
ext = sniff.guess_ext(dataset.path, registry.sniff_order)
else:
ext = dataset.file_type
data_type = ext
# Save job info for the framework
if ext == 'auto' and dataset.ext:
ext = dataset.ext
if ext == 'auto':
ext = 'data'
datatype = registry.get_datatype_by_extension(ext)
if dataset.type in ('server_dir',
'path_paste') and link_data_only == 'link_to_files':
# Never alter a file that will not be copied to Galaxy's local file store.
if datatype.dataset_content_needs_grooming(dataset.path):
err_msg = 'The uploaded files need grooming, so change your <b>Copy data into Galaxy?</b> selection to be ' + \
'<b>Copy files into Galaxy</b> instead of <b>Link to files without copying into Galaxy</b> so grooming can be performed.'
file_err(err_msg, dataset, json_file)
return
if link_data_only == 'copy_files' and dataset.type in (
'server_dir',
'path_paste') and data_type not in ['gzip', 'bz2', 'zip']:
# Move the dataset to its "real" path
if converted_path is not None:
shutil.copy(converted_path, output_path)
try:
os.remove(converted_path)
except:
pass
else:
# This should not happen, but it's here just in case
shutil.copy(dataset.path, output_path)
elif link_data_only == 'copy_files':
if purge_source:
shutil.move(dataset.path, output_path)
else:
shutil.copy(dataset.path, output_path)
# Write the job info
stdout = stdout or 'uploaded %s file' % data_type
info = dict(type='dataset',
dataset_id=dataset.dataset_id,
ext=ext,
stdout=stdout,
name=dataset.name,
line_count=line_count)
if dataset.get('uuid', None) is not None:
info['uuid'] = dataset.get('uuid')
json_file.write(dumps(info) + "\n")
if link_data_only == 'copy_files' and datatype.dataset_content_needs_grooming(
output_path):
# Groom the dataset content if necessary
datatype.groom_dataset_content(output_path)
if len(parts) > 1:
ext = parts[-1].strip().lower()
if not Binary.is_ext_unsniffable(ext):
file_err(
'The uploaded binary file contains inappropriate content',
dataset, json_file)
return
elif Binary.is_ext_unsniffable(
ext) and dataset.file_type != ext:
err_msg = "You must manually set the 'File Format' to '%s' when uploading %s files." % (
ext.capitalize(), ext)
file_err(err_msg, dataset, json_file)
return
if not data_type:
# We must have a text file
if check_html(dataset.path):
file_err(
'The uploaded file contains inappropriate HTML content',
dataset, json_file)
return
if data_type != 'binary':
if link_data_only == 'copy_files':
if dataset.type in ('server_dir',
'path_paste') and data_type not in [
'gzip', 'bz2', 'zip'
]:
in_place = False
# Convert universal line endings to Posix line endings, but allow the user to turn it off,
# so that is becomes possible to upload gzip, bz2 or zip files with binary data without
# corrupting the content of those files.
if dataset.to_posix_lines:
