def test_make_filename(self):
"""download.py: Test directory structure creation from make_filename"""
dest = dl.make_filename(name='full_path.html', dir=self.test_dir)
self.assertEqual(dest,os.path.join(self.test_dir,'full_path.html'))
long_new_path = os.path.join(self.test_dir,'alpha','beta','gamma')
# First the negative test - assert it fails to create new directories
with self.assertRaises(EnvironmentError):
dest = dl.make_filename(name='foobar.html', dir=long_new_path)
self.assertFalse(os.path.isdir(long_new_path))
# Then the positive - assert it succeeds when makedirs is on
dest = dl.make_filename(name='foobar.html', dir=long_new_path,
makedirs=True)
self.assertTrue(os.path.isdir(long_new_path))
def download(cls,name, store=False,silent=True, retries=0):
"""Download a reference genome of the given name, and return a GRCGenome
Fetches the named reference assembly from the web, and creates a new
GRCGenome object to handle it.
If store is False (default), the data will be kept in a temporary file,
and will be destroyed as soon as the object is released. If True,
the entire assembly will be saved in the current
directory - ValueError will be raised if this file seems to already
exist. The resulting file will have the '.assembly' suffix, and may be
either a .zip, a .tar, a .tar.gz, or a .tar.bz2 file. See
``tigerlily.utility.archive.Archive`` for more information.
If store is a string, it will be assumed to be a path to a
directory (trailing slash optional) in which the .tar.gz archive should
be stored. (Again, ValueError will be raised if the file already
exists.) If necessary, any intermiediate directories will be created.
If silent is False, status messages will be printed using print() to
keep the user informed of the progress. This is usually very important
in command line applications as the reference archives are about 900 MB
in size and may take minutes or hours to download depending on the
internet connection.
Because of the large size of these files, it is highly recommended that
the store option be set. Please do not use Tiger Lily to abuse the
UCSC Genome Browser group's generosity in hosting these large files to
the general public.
>>> refgen = GRCGenome.download('test1')
>>> refgen2 = GRCGenome.download('test1',store=True)
>>> import os
>>> os.path.isfile('test1.assembly')
True
>>> os.unlink('test1.assembly')
Only supported reference genome assemblies are allowed, otherwise
ValueError will be raised.
>>> GRCGenome.download('invalid')
Traceback (most recent call last):
...
ValueError: Unknown or unsupported reference genome specified
When downloading the assembly files the download will check with stored
md5 values and compare to see if the download was completed correctly.
if the assembly is not correctly downloaded after specified amount of
retries then an exeption will be thrown and abort the download.
"""
if name not in SUPPORTED_ASSEMBLIES:
raise ValueError('Unknown or unsupported reference genome'
' specified')
url = SUPPORTED_ASSEMBLIES[name]
client = ConsoleDownloader()
if store and store is True:
filename = make_filename(name='{}.assembly'.format(name))
elif store:
name,dir = os.path.split(store)
filename = make_filename(name=name,dir=dir,makedirs=True)
else:
temp = tempfile.NamedTemporaryFile()
filename = temp.name
client.retrieve(url[0], filename=filename, silent=silent)
if url[1] != None:
infile = open(filename,'rb')
content = infile.read()
infile.close()
md5 = hashlib.md5(content).hexdigest()
if md5 == url[1]:
return GRCGenome.load_archive(Archive(filepath=filename))
else:
os.remove(filename)
if retries > 0:
return GRCGenome.download(name,store,silent,retries-1)
else:
raise EnvironmentError('MD5sum failed 5 tries, download aborted')
else:
return GRCGenome.load_archive(Archive(filepath=filename))
