def du(summarize, human_readable, s3_paths):
"""
display disk usage statistics
"""
size_formatter = magic.human_bytes if human_readable else lambda s: s
for s3_path in s3_paths:
totals = {}
# In recursive tallying of totals this key is the stop condition.
stop_path = '/' + s3_util.bucket_and_key(s3_path)[1].rstrip('/')
def tally(path_segment, size):
if not summarize or path_segment == stop_path:
# If the summarize option is present, we only care about the grand total.
if path_segment not in totals:
totals[path_segment] = 0
totals[path_segment] += size
if path_segment != stop_path:
parent_dir = path_segment.rsplit('/', 1)[0]
if parent_dir == '':
# Edge case when totalling the entire bucket.
parent_dir = '/'
tally(parent_dir, size)
for obj in s3_util.keys(s3_path):
# usage for all given paths, and recursively for directories (excludes individual files)
dbg('adding {}', obj)
dir_key = '/' + obj['key'].rsplit('/', 1)[0]
tally(dir_key, obj['len'])
for path, total in sorted(totals.items()):
out('{}\t{}', size_formatter(total), path)
def keys(s3_paths):
"""
Prefix key scan.
Outputs keys in the format s3://bucket/key
"""
if not len(s3_paths):
s3_paths = ['']
for s3_path in s3_paths:
for obj in s3_util.keys(s3_path):
out('s3://{}/{}', obj['bucket'], obj['key'])
def scank(bucket, start, end):
"""
Lexical scan within a bucket.
This scans over keys within a particular bucket where each key k: START_KEY_OR_PREFIX <= k < END_KEY_OR_PREFIX
Outputs keys in the format s3://bucket/key
Example: Scan all keys in mybucket that are lexically equal or greater than 'folder/a' to the end of the bucket.
zs3 scank mybucket folder/a
Authors note: This is particularly useful if you have keyed objects with a date component
(e.g. mybucket/logs/2015-05-10). You could list a range say for everything that happened in 2014 and 2015 with
`zs3 scank mybucket/logs/2014-01-01 /logs/2016-01-01`
"""
for obj in s3_util.scank(bucket, start, end):
out('s3://{}/{}', obj['bucket'], obj['key'])
def scanb(start, end):
"""
Lexical scan across buckets.
This scan traverses across multiple buckets for keys where each key k: START_KEY_OR_PREFIX <= k < END_KEY_OR_PREFIX
Outputs keys in the format s3://bucket/key
Example: Scan all keys that are lexically equal or greater than 'mybucket_a/logs/' and less than
'mybucket_c/logs/2016'. If there are three buckets mybucket_a, mybucket_b, and mybucket_c, this lists
everything in mybucket_a equal to or greater than logs/, everything in mybucket_b, and all keys less than
'logs/2016' in mybucket_c.
zs3 scanb mybucket_a/logs/ mybucket_c/logs/2016
Authors note: I'm not totally sure why would need to page across buckets.
"""
for obj in s3_util.scanb(start, end):
out('s3://{}/{}', obj['bucket'], obj['key'])
def ls(color, long_format, s3_paths):
"""
list directory contents.
zs3 ls [bucket][/key_or_prefix]...
"""
if color:
# Click actually does some of this for us, but keep it in here to be explicit and to mirror `ls --color`
if color.lower() == 'always':
color = True
elif color.lower() == 'auto':
color = sys.stdout.isatty()
if not len(s3_paths):
s3_paths = ['']
for s3_path in s3_paths:
# ls normally prints names aligned to a grid
objs = list(s3_util.ls(s3_path))
if not objs:
_, key = s3_util.bucket_and_key(s3_path)
if key:
raise ClickException("ls {}: No such file or directory".format(s3_path))
else:
# TODO same sort of exception if the bucket doesn't exist
continue
objs = sorted(objs, key=itemgetter('path'))
if long_format:
def out_gen():
for obj in objs:
filename = os.path.basename(obj['path'])
if color and obj['is_dir']:
out_key = Style.BRIGHT + Fore.BLUE + filename + Style.RESET_ALL
else:
out_key = filename
if obj['is_dir']:
yield ['?' * 10,
'??',
'?????',
'?????',
'0',
'??? ?? ??:??',
out_key]
else:
yield ['?' * 10,
'??',
'?????',
'?????',
'{}'.format(obj['size']),
'??? ?? ??:??',
out_key]
out(tabulate(out_gen(), tablefmt='plain'))
else: # not long_format
def out_gen():
for obj in objs:
filename = os.path.basename(obj['path'])
if color and obj['is_dir']:
yield Style.BRIGHT + Fore.BLUE + filename + Style.RESET_ALL
else:
yield filename
out_grid(list(out_gen()))
