1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
|
# This work is licensed under the GNU GPLv2 or later.
# See the COPYING file in the top-level directory.
# This file is part of urlgrabber, a high-level cross-protocol url-grabber
# Copyright 2002-2004 Michael D. Stenner, Ryan Tomayko
# This code is all straight from python-urlgrabber, which we historically
# used the system installed version of. But since the project is in
# maintenance mode upstream, and eventually we want to switch to python3,
# we are just copying this for now.
import sys
import time
import math
import fcntl
import struct
import termios
# Code from https://mail.python.org/pipermail/python-list/2000-May/033365.html
def terminal_width(fd=1):
""" Get the real terminal width """
try:
buf = 'abcdefgh'
buf = fcntl.ioctl(fd, termios.TIOCGWINSZ, buf)
ret = struct.unpack('hhhh', buf)[1]
if ret == 0:
return 80
# Add minimum too?
return ret
except: # IOError
return 80
_term_width_val = None
_term_width_last = None
def terminal_width_cached(fd=1, cache_timeout=1.000):
""" Get the real terminal width, but cache it for a bit. """
global _term_width_val
global _term_width_last
now = time.time()
if _term_width_val is None or (now - _term_width_last) > cache_timeout:
_term_width_val = terminal_width(fd)
_term_width_last = now
return _term_width_val
class TerminalLine:
""" Help create dynamic progress bars, uses terminal_width_cached(). """
def __init__(self, min_rest=0, beg_len=None, fd=1, cache_timeout=1.000):
if beg_len is None:
beg_len = min_rest
self._min_len = min_rest
self._llen = terminal_width_cached(fd, cache_timeout)
if self._llen < beg_len:
self._llen = beg_len
self._fin = False
def __len__(self):
""" Usable length for elements. """
return self._llen - self._min_len
def rest_split(self, fixed, elements=2):
""" After a fixed length, split the rest of the line length among
a number of different elements (default=2). """
if self._llen < fixed:
return 0
return (self._llen - fixed) // elements
def add(self, element, full_len=None):
""" If there is room left in the line, above min_len, add element.
Note that as soon as one add fails all the rest will fail too. """
if full_len is None:
full_len = len(element)
if len(self) < full_len:
self._fin = True
if self._fin:
return ''
self._llen -= len(element)
return element
def rest(self):
""" Current rest of line, same as .rest_split(fixed=0, elements=1). """
return self._llen
class BaseMeter:
def __init__(self):
self.update_period = 0.3 # seconds
self.filename = None
self.url = None
self.basename = None
self.text = None
self.size = None
self.start_time = None
self.fsize = None
self.last_amount_read = 0
self.last_update_time = None
self.re = RateEstimator()
def start(self, filename=None, url=None, basename=None,
size=None, now=None, text=None):
self.filename = filename
self.url = url
self.basename = basename
self.text = text
#size = None ######### TESTING
self.size = size
if not size is None:
self.fsize = format_number(size) + 'B'
if now is None:
now = time.time()
self.start_time = now
self.re.start(size, now)
self.last_amount_read = 0
self.last_update_time = now
self._do_start(now)
def _do_start(self, now=None):
pass
def update(self, amount_read, now=None):
# for a real gui, you probably want to override and put a call
# to your mainloop iteration function here
if now is None:
now = time.time()
if (not self.last_update_time or
(now >= self.last_update_time + self.update_period)):
self.re.update(amount_read, now)
self.last_amount_read = amount_read
self.last_update_time = now
self._do_update(amount_read, now)
def _do_update(self, amount_read, now=None):
pass
def end(self, amount_read, now=None):
if now is None:
now = time.time()
self.re.update(amount_read, now)
self.last_amount_read = amount_read
self.last_update_time = now
self._do_end(amount_read, now)
def _do_end(self, amount_read, now=None):
pass
# This is kind of a hack, but progress is gotten from grabber which doesn't
# know about the total size to download. So we do this so we can get the data
# out of band here. This will be "fixed" one way or anther soon.
_text_meter_total_size = 0
_text_meter_sofar_size = 0
def text_meter_total_size(size, downloaded=0):
global _text_meter_total_size
global _text_meter_sofar_size
_text_meter_total_size = size
_text_meter_sofar_size = downloaded
#
# update: No size (minimal: 17 chars)
# -----------------------------------
# <text> <rate> | <current size> <elapsed time>
# 8-48 1 8 3 6 1 9 5
#
# Order: 1. <text>+<current size> (17)
# 2. +<elapsed time> (10, total: 27)
# 3. + ( 5, total: 32)
# 4. +<rate> ( 9, total: 41)
#
# update: Size, Single file
# -------------------------
# <text> <pc> <bar> <rate> | <current size> <eta time> ETA
# 8-25 1 3-4 1 6-16 1 8 3 6 1 9 1 3 1
#
# Order: 1. <text>+<current size> (17)
# 2. +<eta time> (10, total: 27)
# 3. +ETA ( 5, total: 32)
# 4. +<pc> ( 4, total: 36)
# 5. +<rate> ( 9, total: 45)
# 6. +<bar> ( 7, total: 52)
#
# update: Size, All files
# -----------------------
# <text> <total pc> <pc> <bar> <rate> | <current size> <eta time> ETA
# 8-22 1 5-7 1 3-4 1 6-12 1 8 3 6 1 9 1 3 1
#
# Order: 1. <text>+<current size> (17)
# 2. +<eta time> (10, total: 27)
# 3. +ETA ( 5, total: 32)
# 4. +<total pc> ( 5, total: 37)
# 4. +<pc> ( 4, total: 41)
# 5. +<rate> ( 9, total: 50)
# 6. +<bar> ( 7, total: 57)
#
# end
# ---
# <text> | <current size> <elapsed time>
# 8-56 3 6 1 9 5
#
# Order: 1. <text> ( 8)
# 2. +<current size> ( 9, total: 17)
# 3. +<elapsed time> (10, total: 27)
# 4. + ( 5, total: 32)
#
def _term_add_bar(tl, bar_max_length, pc):
blen = bar_max_length
bar = '='*int(blen * pc)
if (blen * pc) - int(blen * pc) >= 0.5:
bar += '-'
return tl.add(' [%-*.*s]' % (blen, blen, bar))
def _term_add_end(tl, osize, size):
if osize: # osize should be None or >0, but that's been broken.
if size > osize: # Is ??? better? Really need something to say < vs >.
return tl.add(' !!! '), True
elif size != osize:
return tl.add(' ... '), True
return tl.add(' ' * 5), False
class TextMeter(BaseMeter):
def __init__(self, fo=sys.stderr):
BaseMeter.__init__(self)
self.fo = fo
def _do_update(self, amount_read, now=None):
etime = self.re.elapsed_time()
fread = format_number(amount_read)
#self.size = None
if self.text is not None:
text = self.text
else:
text = self.basename
ave_dl = format_number(self.re.average_rate())
sofar_size = None
if _text_meter_total_size:
sofar_size = _text_meter_sofar_size + amount_read
sofar_pc = (sofar_size * 100) // _text_meter_total_size
# Include text + ui_rate in minimal
tl = TerminalLine(8, 8+1+8)
# For big screens, make it more readable.
use_hours = bool(tl._llen > 80)
ui_size = tl.add(' | %5sB' % fread)
if self.size is None:
ui_time = tl.add(' %s' % format_time(etime, use_hours))
ui_end = tl.add(' ' * 5)
ui_rate = tl.add(' %5sB/s' % ave_dl)
out = '%-*.*s%s%s%s%s\r' % (tl.rest(), tl.rest(), text,
ui_rate, ui_size, ui_time, ui_end)
else:
rtime = self.re.remaining_time()
frtime = format_time(rtime, use_hours)
frac = self.re.fraction_read()
ui_time = tl.add(' %s' % frtime)
ui_end = tl.add(' ETA ')
if sofar_size is None:
ui_sofar_pc = ''
else:
ui_sofar_pc = tl.add(' (%i%%)' % sofar_pc,
full_len=len(" (100%)"))
ui_pc = tl.add(' %2i%%' % (frac*100))
ui_rate = tl.add(' %5sB/s' % ave_dl)
# Make text grow a bit before we start growing the bar too
blen = 4 + tl.rest_split(8 + 8 + 4)
ui_bar = _term_add_bar(tl, blen, frac)
out = '\r%-*.*s%s%s%s%s%s%s%s\r' % (tl.rest(), tl.rest(), text,
ui_sofar_pc, ui_pc, ui_bar,
ui_rate,ui_size,ui_time, ui_end)
self.fo.write(out)
self.fo.flush()
def _do_end(self, amount_read, now=None):
global _text_meter_total_size
global _text_meter_sofar_size
total_size = format_number(amount_read)
if self.text is not None:
text = self.text
else:
text = self.basename
tl = TerminalLine(8)
# For big screens, make it more readable.
use_hours = bool(tl._llen > 80)
ui_size = tl.add(' | %5sB' % total_size)
ui_time = tl.add(' %s' % format_time(self.re.elapsed_time(), use_hours))
ui_end, not_done = _term_add_end(tl, self.size, amount_read)
out = '\r%-*.*s%s%s%s\n' % (tl.rest(), tl.rest(), text,
ui_size, ui_time, ui_end)
self.fo.write(out)
self.fo.flush()
# Don't add size to the sofar size until we have all of it.
# If we don't have a size, then just pretend/hope we got all of it.
if not_done:
return
if _text_meter_total_size:
_text_meter_sofar_size += amount_read
if _text_meter_total_size <= _text_meter_sofar_size:
_text_meter_total_size = 0
_text_meter_sofar_size = 0
text_progress_meter = TextMeter
######################################################################
# support classes and functions
class RateEstimator:
def __init__(self, timescale=5.0):
self.timescale = timescale
self.total = None
self.start_time = None
self.last_update_time = None
self.last_amount_read = 0
self.ave_rate = None
def start(self, total=None, now=None):
if now is None:
now = time.time()
self.total = total
self.start_time = now
self.last_update_time = now
self.last_amount_read = 0
self.ave_rate = None
def update(self, amount_read, now=None):
if now is None:
now = time.time()
# libcurl calls the progress callback when fetching headers
# too, thus amount_read = 0 .. hdr_size .. 0 .. content_size.
# Occasionally we miss the 2nd zero and report avg speed < 0.
# Handle read_diff < 0 here. BZ 1001767.
if amount_read == 0 or amount_read < self.last_amount_read:
# if we just started this file, all bets are off
self.last_update_time = now
self.last_amount_read = amount_read
self.ave_rate = None
return
#print 'times', now, self.last_update_time
time_diff = now - self.last_update_time
read_diff = amount_read - self.last_amount_read
# First update, on reget is the file size
if self.last_amount_read:
self.last_update_time = now
self.ave_rate = self._temporal_rolling_ave(\
time_diff, read_diff, self.ave_rate, self.timescale)
self.last_amount_read = amount_read
#print 'results', time_diff, read_diff, self.ave_rate
#####################################################################
# result methods
def average_rate(self):
"get the average transfer rate (in bytes/second)"
return self.ave_rate
def elapsed_time(self):
"the time between the start of the transfer and the most recent update"
return self.last_update_time - self.start_time
def remaining_time(self):
"estimated time remaining"
if not self.ave_rate or not self.total:
return None
return (self.total - self.last_amount_read) / self.ave_rate
def fraction_read(self):
"""the fraction of the data that has been read
(can be None for unknown transfer size)"""
if self.total is None:
return None
elif self.total == 0:
return 1.0
else:
return float(self.last_amount_read) / self.total
#########################################################################
# support methods
def _temporal_rolling_ave(self, time_diff, read_diff, last_ave, timescale):
"""a temporal rolling average performs smooth averaging even when
updates come at irregular intervals. This is performed by scaling
the "epsilon" according to the time since the last update.
Specifically, epsilon = time_diff / timescale
As a general rule, the average will take on a completely new value
after 'timescale' seconds."""
epsilon = time_diff / timescale
if epsilon > 1:
epsilon = 1.0
return self._rolling_ave(time_diff, read_diff, last_ave, epsilon)
def _rolling_ave(self, time_diff, read_diff, last_ave, epsilon):
"""perform a "rolling average" iteration
a rolling average "folds" new data into an existing average with
some weight, epsilon. epsilon must be between 0.0 and 1.0 (inclusive)
a value of 0.0 means only the old value (initial value) counts,
and a value of 1.0 means only the newest value is considered."""
try:
recent_rate = read_diff / time_diff
except ZeroDivisionError:
recent_rate = None
if last_ave is None:
return recent_rate
elif recent_rate is None:
return last_ave
# at this point, both last_ave and recent_rate are numbers
return epsilon * recent_rate + (1 - epsilon) * last_ave
def _round_remaining_time(self, rt, start_time=15.0):
"""round the remaining time, depending on its size
If rt is between n*start_time and (n+1)*start_time round downward
to the nearest multiple of n (for any counting number n).
If rt < start_time, round down to the nearest 1.
For example (for start_time = 15.0):
2.7 -> 2.0
25.2 -> 25.0
26.4 -> 26.0
35.3 -> 34.0
63.6 -> 60.0
"""
if rt < 0:
return 0.0
shift = int(math.log(rt / start_time) / math.log(2))
rt = int(rt)
if shift <= 0:
return rt
return float(int(rt) >> shift << shift)
def format_time(seconds, use_hours=0):
if seconds is None or seconds < 0:
if use_hours:
return '--:--:--'
else:
return '--:--'
elif seconds == float('inf'):
return 'Infinite'
else:
seconds = int(seconds)
minutes = seconds // 60
seconds = seconds % 60
if use_hours:
hours = minutes // 60
minutes = minutes % 60
return '%02i:%02i:%02i' % (hours, minutes, seconds)
else:
return '%02i:%02i' % (minutes, seconds)
def format_number(number, SI=0, space=' '):
"""Turn numbers into human-readable metric-like numbers"""
symbols = ['', # (none)
'k', # kilo
'M', # mega
'G', # giga
'T', # tera
'P', # peta
'E', # exa
'Z', # zetta
'Y'] # yotta
if SI:
step = 1000.0
else:
step = 1024.0
thresh = 999
depth = 0
max_depth = len(symbols) - 1
number = number or 0
# we want numbers between 0 and thresh, but don't exceed the length
# of our list. In that event, the formatting will be screwed up,
# but it'll still show the right number.
while number > thresh and depth < max_depth:
depth = depth + 1
number = number / step
if isinstance(number, int):
# it's an int or a long, which means it didn't get divided,
# which means it's already short enough
fmt = '%i%s%s'
elif number < 9.95:
# must use 9.95 for proper sizing. For example, 9.99 will be
# rounded to 10.0 with the .1f format string (which is too long)
fmt = '%.1f%s%s'
else:
fmt = '%.0f%s%s'
return(fmt % (float(number or 0), space, symbols[depth]))
|