import sys
import logging
-logger = logging.getLogger(os.path.basename(sys.argv[0]))
+import arvados
+import arvados.commands._util as arv_cmd
+
+logger = logging.getLogger('arvados.arv-get')
+
+def abort(msg, code=1):
+ print >>sys.stderr, "arv-get:", msg
+ exit(code)
parser = argparse.ArgumentParser(
- description='Copy data from Keep to a local file or pipe.')
+ description='Copy data from Keep to a local file or pipe.',
+ parents=[arv_cmd.retry_opt])
parser.add_argument('locator', type=str,
help="""
Collection locator, optionally with a file path or prefix.
group.add_argument('--progress', action='store_true',
help="""
Display human-readable progress on stderr (bytes and, if possible,
-percentage of total data size). This is the default behavior when
-stderr is a tty and stdout is not a tty.
+percentage of total data size). This is the default behavior when it
+is not expected to interfere with the output: specifically, stderr is
+a tty _and_ either stdout is not a tty, or output is being written to
+named files rather than stdout.
""")
group.add_argument('--no-progress', action='store_true',
help="""
args.destination[-1] == os.path.sep):
args.destination = os.path.join(args.destination,
os.path.basename(args.locator))
- logger.debug("Appended source file name to destination directory: %s" %
+ logger.debug("Appended source file name to destination directory: %s",
args.destination)
-# Turn on --progress by default if stderr is a tty and stdout isn't.
-if (not (args.batch_progress or args.no_progress)
- and os.isatty(sys.stderr.fileno())
- and not os.isatty(sys.stdout.fileno())):
- args.progress = True
-
if args.destination == '-':
args.destination = '/dev/stdout'
if args.destination == '/dev/stdout':
else:
args.destination = args.destination.rstrip(os.sep)
+# Turn on --progress by default if stderr is a tty and output is
+# either going to a named file, or going (via stdout) to something
+# that isn't a tty.
+if (not (args.batch_progress or args.no_progress)
+ and sys.stderr.isatty()
+ and (args.destination != '/dev/stdout'
+ or not sys.stdout.isatty())):
+ args.progress = True
-import arvados
r = re.search(r'^(.*?)(/.*)?$', args.locator)
collection = r.group(1)
get_prefix = r.group(2)
if args.r and not get_prefix:
get_prefix = os.sep
+api_client = arvados.api('v1')
+reader = arvados.CollectionReader(collection, num_retries=args.retries)
-todo = []
-todo_bytes = 0
if not get_prefix:
- try:
- if not args.n:
- if not args.f and os.path.exists(args.destination):
- logger.error('Local file %s already exists' % args.destination)
- sys.exit(1)
- with open(args.destination, 'wb') as f:
- try:
- c = arvados.api('v1').collections().get(
- uuid=collection).execute()
- manifest = c['manifest_text']
- except Exception as e:
- logging.warning(
- "API lookup failed for collection %s (%s: %s)" %
- (collection, type(e), str(e)))
- manifest = arvados.Keep.get(collection)
- f.write(manifest)
- sys.exit(0)
- except arvados.errors.NotFoundError as e:
- logger.error(e)
- sys.exit(1)
-
-reader = arvados.CollectionReader(collection)
+ if not args.n:
+ open_flags = os.O_CREAT | os.O_WRONLY
+ if not args.f:
+ open_flags |= os.O_EXCL
+ try:
+ out_fd = os.open(args.destination, open_flags)
+ with os.fdopen(out_fd, 'wb') as out_file:
+ out_file.write(reader.manifest_text())
+ except (IOError, OSError) as error:
+ abort("can't write to '{}': {}".format(args.destination, error))
+ except (arvados.errors.ApiError, arvados.errors.KeepReadError) as error:
+ abort("failed to download '{}': {}".format(collection, error))
+ sys.exit(0)
+
+reader.normalize()
# Scan the collection. Make an array of (stream, file, local
# destination filename) tuples, and add up total size to extract.
-
+todo = []
+todo_bytes = 0
try:
for s in reader.all_streams():
for f in s.all_files():
os.path.join(s.name(), f.name())[len(get_prefix)+1:])
if (not (args.n or args.f or args.skip_existing) and
os.path.exists(dest_path)):
- logger.error('Local file %s already exists' % dest_path)
- sys.exit(1)
+ abort('Local file %s already exists.' % (dest_path,))
else:
if os.path.join(s.name(), f.name()) != '.' + get_prefix:
continue
todo += [(s, f, dest_path)]
todo_bytes += f.size()
except arvados.errors.NotFoundError as e:
- logger.error(e)
- sys.exit(1)
+ abort(e)
# Read data, and (if not -n) write to local file(s) or pipe.
digestor = None
if not args.n:
if args.skip_existing and os.path.exists(outfilename):
- logger.debug('Local file %s exists. Skipping.' % outfilename)
+ logger.debug('Local file %s exists. Skipping.', outfilename)
continue
elif not args.f and (os.path.isfile(outfilename) or
os.path.isdir(outfilename)):
# Good thing we looked again: apparently this file wasn't
# here yet when we checked earlier.
- logger.error('Local file %s already exists' % outfilename)
- sys.exit(1)
+ abort('Local file %s already exists.' % (outfilename,))
if args.r:
arvados.util.mkdir_dash_p(os.path.dirname(outfilename))
try:
outfile = open(outfilename, 'wb')
- except Exception as e:
- logger.error('Open(%s) failed: %s' % (outfilename, e))
- sys.exit(1)
+ except Exception as error:
+ abort('Open(%s) failed: %s' % (outfilename, error))
if args.hash:
digestor = hashlib.new(args.hash)
try:
sys.stderr.write("%s %s/%s\n"
% (digestor.hexdigest(), s.name(), f.name()))
except KeyboardInterrupt:
- if outfile:
+ if outfile and outfilename != '/dev/stdout':
os.unlink(outfilename)
break
projects / arvados.git / blobdiff