def get_eggnog_proteins_file():
return pjoin(get_data_path(), "e5.proteomes.faa")
def get_eggnog_taxid_info_file():
return pjoin(get_data_path(), "e5.taxid_info.tsv")
def download_og2level():
url = 'http://eggnogdb.embl.de/download/emapperdb-%s/og2level.tsv.gz' % (
DATABASE_VERSION)
cmd = 'cd %s && wget -O og2level.tsv.gz %s' % (get_data_path(), url)
run(cmd)
sys.exit(1)
if (args.taxids is not None and
args.taxids != "") and (args.taxa is not None and args.taxa != ""):
print(colorify(f'Use either --taxids or --taxa, not both', 'red'))
sys.exit(1)
##
if "EGGNOG_DATA_DIR" in os.environ:
set_data_path(os.environ["EGGNOG_DATA_DIR"])
if args.data_dir:
set_data_path(args.data_dir)
data_path = get_data_path()
# http://eggnog5.embl.de/download/eggnog_5.0/e5.proteomes.faa
if not pexists(get_eggnog_proteins_file()):
if args.allyes or ask(
f"Download eggnog5 proteins to {data_path}? ~9GB (It is required to create new databases)"
) == 'y':
print(
colorify(
f'Downloading eggnog5 proteins file to {data_path}...',
'green'))
download_proteins(data_path)
else:
print(
colorify(
f'eggnog5 proteins file was not found. Use --data_dir to specify another data path, or allow the download',
def download_diamond_db():
url = 'http://eggnogdb.embl.de/download/emapperdb-%s/eggnog_proteins.dmnd.gz' % (
DATABASE_VERSION)
cmd = 'cd %s && wget -nH --user-agent=Mozilla/5.0 --relative --no-parent --reject "index.html*" --cut-dirs=4 -e robots=off -O eggnog_proteins.dmnd.gz %s && echo Decompressing... && gunzip eggnog_proteins.dmnd.gz %s' % (
get_data_path(), url, gunzip_flag())
run(cmd)
def download_groups():
url = 'http://eggnogdb.embl.de/download/emapperdb-%s/OG_fasta.tar.gz' % (
DATABASE_VERSION)
cmd = 'cd %s && wget -nH --user-agent=Mozilla/5.0 --relative --no-parent --reject "index.html*" --cut-dirs=4 -e robots=off -O OG_fasta.tar.gz %s && echo Decompressing... && tar -zxf OG_fasta.tar.gz && rm OG_fasta.tar.gz' % (
get_data_path(), url)
run(cmd)
args = parser.parse_args()
if "EGGNOG_DATA_DIR" in os.environ:
set_data_path(os.environ["EGGNOG_DATA_DIR"])
if args.data_dir:
set_data_path(args.data_dir)
# if args.force or not pexists(pjoin(get_data_path(), 'og2level.tsv.gz')):
# print colorify('Downloading "og2level.tsv.gz" at %s' %get_data_path(), 'green')
# download_og2level()
# if 'all' in args.dbs:
# args.dbs = EGGNOG_DATABASES
if args.force or not pexists(pjoin(get_data_path(), 'eggnog.db')):
if args.allyes or ask("Download main annotation database?") == 'y':
print colorify(
'Downloading "eggnog.db" at %s...' % get_data_path(), 'green')
download_annotations()
else:
print 'Skipping'
else:
if not args.quiet:
print colorify(
'Skipping eggnog.db database (already present). Use -f to force download',
'lblue')
# if args.force or not pexists(pjoin(get_data_path(), 'OG_fasta')):
# if args.allyes or ask("Download OG fasta files for annotation refinement (~20GB after decompression)?") == 'y':
parser.add_argument('-q', action="store_true", dest='quiet',
help='quiet_mode')
parser.add_argument("--data_dir", metavar='', type=existing_dir,
help='Directory to use for DATA_PATH.')
args = parser.parse_args()
if "EGGNOG_DATA_DIR" in os.environ:
set_data_path(os.environ["EGGNOG_DATA_DIR"])
if args.data_dir:
set_data_path(args.data_dir)
data_path = get_data_path()
##
# Annotation DB
if args.force or not pexists(get_eggnogdb_file()):
if args.allyes or ask("Download main annotation database?") == 'y':
print(colorify(f'Downloading "eggnog.db" at {data_path}...', 'green'))
download_annotations(data_path)
else:
print('Skipping')
else:
if not args.quiet:
print(colorify('Skipping eggnog.db database (already present). Use -f to force download', 'lblue'))
##
parser.add_argument('-q',
action="store_true",
dest='quiet',
help='quiet_mode')
parser.add_argument("--data_dir",
metavar='',
type=existing_dir,
help='Directory to use for DATA_PATH.')
args = parser.parse_args()
if args.data_dir:
set_data_path(args.data_dir)
if args.force or not pexists(pjoin(get_data_path(), 'og2level.tsv.gz')):
print colorify('Downloading "og2level.tsv.gz" at %s' % get_data_path(),
'green')
download_og2level()
if 'all' in args.dbs:
args.dbs = EGGNOG_DATABASES
if args.force or not pexists(pjoin(get_data_path(), 'eggnog.db')):
if args.allyes or ask("Download main annotation database?") == 'y':
print colorify(
'Downloading "eggnog.db" at %s...' % get_data_path(), 'green')
download_annotations()
else:
print 'Skipping'
