Update bigtables tool: Sort by size with indexes, not just data.
import sys
class ProgressWrapper:
def __init__(self, f, size=0):
self.f = f
self.size = size
self.readsize = 0
self.cycles = 0
self.print_cycles = 200
def humanize(self, bytes):
if bytes > 1024**3:
return '%.1fG' % (bytes / 1024.**3)
if bytes > 1024**2:
return '%.1fM' % (bytes / 1024.**2)
if bytes > 1024:
return '%.1fk' % (bytes / 1024.)
return '%d' % bytes
def write(self, data):
self.size += len(data)
if self.cycles == 0:
print(' read %s \r' % self.humanize(self.size), end='')
sys.stdout.flush()
self.cycles = self.print_cycles * 200
else:
self.cycles -= 1
return self.f.write(data)
def read(self, size):
self.readsize += size
if self.cycles == 0:
if self.size > 0:
percent = self.readsize * 100. / self.size
else:
percent = 100
if percent > 100:
percent = 100
print(' written %s / %s (%.1f%%) \r' % (
self.humanize(self.readsize),
self.humanize(self.size),
percent), end='')
sys.stdout.flush()
self.cycles = self.print_cycles
else:
self.cycles -= 1
return self.f.read(size)
def close(self):
self.f.close()