-
Notifications
You must be signed in to change notification settings - Fork 0
/
compute_production_stats.py
executable file
·523 lines (468 loc) · 21.9 KB
/
compute_production_stats.py
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
518
519
520
521
522
523
import os
import glob
import re
import sys
import socket
import couchdb
import logging
import argparse
import yaml
import json
import distance
import operator
import time
from datetime import date
from datetime import datetime
try:
import ConfigParser
except ImportError:
import configparser
CONFIG = {}
logger = logging.getLogger(__name__)
def setupServer(conf):
db_conf = conf['statusdb']
url="https://{0}:{1}@{2}".format(db_conf['username'], db_conf['password'], db_conf['url'])
return couchdb.Server(url)
def load_yaml_config(config_file):
"""Load YAML config file
:param str config_file: The path to the configuration file.
:returns: A dict of the parsed config file.
:rtype: dict
:raises IOError: If the config file cannot be opened.
"""
if type(config_file) is file:
CONFIG.update(yaml.load(config_file) or {})
return CONFIG
else:
try:
with open(config_file, 'r') as f:
content = yaml.load(f)
CONFIG.update(content)
return content
except IOError as e:
e.message = "Could not open configuration file \"{}\".".format(config_file)
raise e
def get_FC_type(FCid):
FC_type = ""
if "ST-" in FCid:
FC_type = "HiSeqX"
elif "000000000-" in FCid:
FC_type = "MiSeq"
else:
FC_type = "HiSeq2500"
return FC_type
def parse_flowcell_db():
couch = setupServer(CONFIG)
#fetch info about proejcts (reference type)
projectsDB = couch["projects"]
project_summary = projectsDB.view("project/summary")
projects = {}
for row in project_summary:
if "project_name" not in row.value:
print("somehting is wrong here... I guess I am going to fail")
if "reference_genome" in row.value:
projects[row.value["project_name"]] = row.value["reference_genome"]
else:
projects[row.value["project_name"]] = "None"
flowcell_db = couch["x_flowcells"]
flowcells = {}
instrument_types = ["HiSeqX", "MiSeq", "HiSeq2500"]
for instrument_type in instrument_types:
flowcells[instrument_type] = {}
for fc_doc in flowcell_db:
try:
samplesheet_csv = flowcell_db[fc_doc]["samplesheet_csv"]
except KeyError:
if "RunInfo" in flowcell_db[fc_doc]:
print("{}".format(flowcell_db[fc_doc]["RunInfo"]["Id"]))
continue
flowcell_id = flowcell_db[fc_doc]["RunInfo"]["Id"]
instrument_type = get_FC_type(flowcell_id)
if flowcell_id not in flowcells[instrument_type]:
flowcells[instrument_type][flowcell_id] = {}
for sample_lane in samplesheet_csv:
if instrument_type == "MiSeq":
lane = "1"
else:
lane = sample_lane["Lane"]
if lane not in flowcells[instrument_type][flowcell_id]:
flowcells[instrument_type][flowcell_id][lane] = {}
if "Sample_Project" in sample_lane:
project = sample_lane["Sample_Project"].strip()
elif "Project" in sample_lane:
project = sample_lane["Project"].strip()
else:
print("WRONG")
# now correct the project
if "." not in project:
project = project.replace("__", "_").replace("_", ".", 1)
if "." in project:
project_pieaces = project.split(".")
project = "{}.{}".format(project_pieaces[0].upper(), project_pieaces[1])
if project not in projects:
print("{} not found in projects".format(project))
continue
if project in flowcells[instrument_type][flowcell_id][lane]:
# cehck that it is the same
if projects[project] != flowcells[instrument_type][flowcell_id][lane][project]:
print("stange found same project but different organisms")
else:
flowcells[instrument_type][flowcell_id][lane][project] = projects[project]
#now in flowcell I have a full list of instrument/FCid/lanes/Project and organism
#now i can pull all the stats I want
date_low_limit = date(16,1,1)
date_upper_limit = date(17,1,1)
total_number_FC = 0
total_number_lanes = 0
total_number_human_lanes = 0
total_number_non_human_lanes = 0
total_number_mixed_lanes = 0
for instrument_type in flowcells:
instrument_number_FC = 0
instrument_number_lanes = 0
instrument_number_human_lanes = 0
instrument_number_non_human_lanes = 0
instrument_number_mixed_lanes = 0
for flowcell_id in flowcells[instrument_type]:
start_date_string = flowcell_id
year = start_date_string[0:2]
month = start_date_string[2:4]
day = start_date_string[4:6]
fc_date = date(int(year), int(month), int(day))
if fc_date >= date_low_limit and fc_date < date_upper_limit:
total_number_FC += 1
instrument_number_FC += 1
for lane in flowcells[instrument_type][flowcell_id]:
total_number_lanes += 1
instrument_number_lanes += 1
lane_type = ""
for project in flowcells[instrument_type][flowcell_id][lane]:
if lane_type == "":
lane_type = flowcells[instrument_type][flowcell_id][lane][project]
elif lane_type != flowcells[instrument_type][flowcell_id][lane][project]:
lane_type = "mixed"
else:
lane_type = flowcells[instrument_type][flowcell_id][lane][project]
if lane_type == "mixed":
total_number_mixed_lanes += 1
instrument_number_mixed_lanes += 1
elif lane_type == "hg19":
total_number_human_lanes += 1
instrument_number_human_lanes += 1
else:
total_number_non_human_lanes += 1
instrument_number_non_human_lanes += 1
else:
continue
print("{}".format(instrument_type))
print("\tNumber of FC: {}".format(instrument_number_FC))
print("\tNumber of lanes: {}".format(instrument_number_lanes))
print("\tNumber of Human lanes: {}".format(instrument_number_human_lanes))
print("\tNumber of Non-Human lanes: {}".format(instrument_number_non_human_lanes))
print("\tNumber of Mixed lanes: {}".format(instrument_number_mixed_lanes))
print("TOTAL")
print("\tNumber of FC: {}".format(total_number_FC))
print("\tNumber of lanes: {}".format(total_number_lanes))
print("\tNumber of Human lanes: {}".format(total_number_human_lanes))
print("\tNumber of Non-Human lanes: {}".format(total_number_non_human_lanes))
print("\tNumber of Mixed lanes: {}".format(total_number_mixed_lanes))
def instrument_usage():
couch = setupServer(CONFIG)
#fetch info about proejcts (reference type)
projectsDB = couch["projects"]
project_summary = projectsDB.view("project/summary")
projects = {}
instruments = {}
for row in project_summary:
if "close_date" not in row.value:
continue
if 'aborted' in row.value['details'] and row.value['details']['aborted']:
continue
year_close_date = int(row.value["close_date"].split("-")[0])
if year_close_date >= 2013:
if 'sequencing_platform' not in row.value['details']:
continue
else:
instrument = row.value['details']['sequencing_platform']
if 'sequencing_setup' not in row.value['details']:
continue
else:
sequencing_setup = row.value['details']['sequencing_setup']
if sequencing_setup == "special" or sequencing_setup == "Special":
continue
pattern = re.compile("^[0-9]+x[0-9]+")
if not pattern.match(sequencing_setup):
continue
if instrument not in instruments:
instruments[instrument] = {'number':1, 'setup': {sequencing_setup : 1}}
else:
instruments[instrument]['number'] += 1
if sequencing_setup not in instruments[instrument]['setup']:
instruments[instrument]['setup'][sequencing_setup] = 1
else:
instruments[instrument]['setup'][sequencing_setup] += 1
project_name = row.value["project_name"]
projects[project_name] = {'sequencing_platform': instrument,
'sequencing_setup' : sequencing_setup,
'samples_sequenced': set(),
'lanes': 0,
'sequencers' : set()
}
flowcell_db = couch["x_flowcells"]
project_sequenced = {}
instrument_runs_per_week = {}
for fc_doc in flowcell_db:
if 'RunInfo' not in flowcell_db[fc_doc]:
continue
instrument = flowcell_db[fc_doc]["RunInfo"]['Instrument']
if 'illumina' not in flowcell_db[fc_doc]:
print("Not illumina field found in doc")
continue
if 'Demultiplex_Stats' not in flowcell_db[fc_doc]['illumina']:
print("Not Demultiplex_Stats field found in doc")
continue
if 'Barcode_lane_statistics' not in flowcell_db[fc_doc]['illumina']['Demultiplex_Stats']:
print("Not Barcode_lane_statistics field found in doc")
continue
projects_in_lanes = {}
for sample_lane in flowcell_db[fc_doc]['illumina']['Demultiplex_Stats']['Barcode_lane_statistics']:
if sample_lane['Sample'] == 'unknown' :
continue
if "Sample_Project" in sample_lane:
project = sample_lane["Sample_Project"].strip()
elif "Project" in sample_lane:
project = sample_lane["Project"].strip()
else:
print("WRONG")
# now correct the project
if "." not in project:
project = project.replace("__", "_").replace("_", ".", 1)
if "." in project:
project_pieaces = project.split(".")
project = "{}.{}".format(project_pieaces[0].upper(), project_pieaces[1])
if project == "default":
continue
elif project not in projects:
continue
projects[project]['samples_sequenced'].update([sample_lane['Sample']])
projects[project]['sequencers'].update([instrument])
lane = sample_lane['Lane']
if lane not in projects_in_lanes:
projects_in_lanes[lane] = {}
if project not in projects_in_lanes[lane] :
projects_in_lanes[lane][project] = 1
else:
projects_in_lanes[lane][project] += 1
year = int("20" + flowcell_db[fc_doc]['RunInfo']['Date'][0:2])
month = int(flowcell_db[fc_doc]['RunInfo']['Date'][2:4])
day = int(flowcell_db[fc_doc]['RunInfo']['Date'][4:6])
date_seq = datetime(year , month , day )
if instrument not in instrument_runs_per_week:
instrument_runs_per_week[instrument] = {}
date_entry = "{}_{}".format(year, month) #date_seq.isocalendar()[1]) # year plus week number
if date_entry not in instrument_runs_per_week[instrument]:
instrument_runs_per_week[instrument][date_entry] = 1
else:
instrument_runs_per_week[instrument][date_entry] += 1
for lane in projects_in_lanes:
for project in projects_in_lanes[lane]:
projects[project]['lanes'] += 1
if 'date' in projects[project]:
if projects[project]['date'] < date_seq:
projects[project]['date'] = date_seq
else:
projects[project]['date'] = date_seq
flowcell_db = couch["flowcells"]
for fc_doc in flowcell_db:
if 'RunInfo' not in flowcell_db[fc_doc]:
continue
if 'Date' not in flowcell_db[fc_doc]['RunInfo']:
continue
year = int(flowcell_db[fc_doc]['RunInfo']['Date'][0:2])
if year < 13:
print("run {} too old".format(flowcell_db[fc_doc]['RunInfo']['Id']))
continue
if 'Instrument' not in flowcell_db[fc_doc]["RunInfo"]:
print("ERROR: Instrument not found in RunInfo: how is this possible?")
exit
instrument = flowcell_db[fc_doc]["RunInfo"]['Instrument']
if 'illumina' not in flowcell_db[fc_doc]:
print("Not illumina field found in doc {}".format(fc_doc))
continue
if 'Demultiplex_Stats' not in flowcell_db[fc_doc]['illumina']:
print("Not Demultiplex_Stats field found in doc {}".format(fc_doc))
continue
if 'Barcode_lane_statistics' not in flowcell_db[fc_doc]['illumina']['Demultiplex_Stats']:
print("Not Barcode_lane_statistics field found in doc {}".format(fc_doc))
continue
projects_in_lanes = {}
for sample_lane in flowcell_db[fc_doc]['illumina']['Demultiplex_Stats']['Barcode_lane_statistics']:
if sample_lane['Sample ID'] == 'unknown' :
continue
if "Description" in sample_lane:
project = sample_lane["Description"].strip()
elif "Project" in sample_lane:
project = sample_lane["Project"].strip()
else:
print("WRONG")
# now correct the project
if "." not in project:
project = project.replace("__", "_").replace("_", ".", 1)
if "." in project:
project_pieaces = project.split(".")
project = "{}.{}".format(project_pieaces[0].upper(), project_pieaces[1])
if project == "default":
continue
elif project not in projects:
continue
elif len(projects[project]['samples_sequenced']) > 0:
continue #it means I have already saw this i x_flowcell db
projects[project]['samples_sequenced'].update([sample_lane['Sample ID']])
projects[project]['sequencers'].update([instrument])
lane = sample_lane['Lane']
if lane not in projects_in_lanes:
projects_in_lanes[lane] = {}
if project not in projects_in_lanes[lane] :
projects_in_lanes[lane][project] = 1
else:
projects_in_lanes[lane][project] += 1
year = int("20" + flowcell_db[fc_doc]['RunInfo']['Date'][0:2])
month = int(flowcell_db[fc_doc]['RunInfo']['Date'][2:4])
day = int(flowcell_db[fc_doc]['RunInfo']['Date'][4:6])
date_seq = datetime(year , month , day )
if instrument not in instrument_runs_per_week:
instrument_runs_per_week[instrument] = {}
date_entry = "{}_{}".format(year, month) #date_seq.isocalendar()[1]) # year plus week number
if date_entry not in instrument_runs_per_week[instrument]:
instrument_runs_per_week[instrument][date_entry] = 1
else:
instrument_runs_per_week[instrument][date_entry] += 1
for lane in projects_in_lanes:
for project in projects_in_lanes[lane]:
projects[project]['lanes'] += 1
if 'date' in projects[project]:
if projects[project]['date'] < date_seq:
projects[project]['date'] = date_seq
else:
projects[project]['date'] = date_seq
years = [2013, 2014, 2015, 2016, 2017]
sys.stdout.write('date,')
for instrument in sorted(instrument_runs_per_week):
sys.stdout.write('{},'.format(instrument))
sys.stdout.write('\n')
for year in years:
for week in range(1,13):
date_to_search = "{}_{}".format(year, week)
sys.stdout.write('{},'.format(date_to_search))
for instrument in sorted(instrument_runs_per_week):
if date_to_search in instrument_runs_per_week[instrument]:
sys.stdout.write('{},'.format(instrument_runs_per_week[instrument][date_to_search]))
else:
sys.stdout.write('0,')
sys.stdout.write('\n')
sys.stdout.write('\n')
years = [2013, 2014, 2015, 2016, 2017]
sequencers_year_setup = {}
sequencers_setup = set([ projects[project]['sequencing_setup'] for project in projects if projects[project]['lanes'] > 0])
for project in projects:
if len(projects[project]['sequencers']) == 0:
continue
year = projects[project]['date'].year
samples = len(projects[project]['samples_sequenced'])
sequencers = projects[project]['sequencers']
sequencing_setup = projects[project]['sequencing_setup']
for sequencer in sequencers:
if sequencer not in sequencers_year_setup:
sequencers_year_setup[sequencer] = {}
if year not in sequencers_year_setup[sequencer]:
sequencers_year_setup[sequencer][year] = {}
if sequencing_setup not in sequencers_year_setup[sequencer][year]:
sequencers_year_setup[sequencer][year][sequencing_setup] = samples
else:
sequencers_year_setup[sequencer][year][sequencing_setup] += samples
for sequencer in sorted(sequencers_year_setup):
sys.stdout.write('{}\n'.format(sequencer))
for setup in sorted(sequencers_setup):
sys.stdout.write('{},'.format(setup))
for year in sorted(years):
if year not in sequencers_year_setup[sequencer]:
#this sequencer, this year had no runs
sys.stdout.write('0,')
else:
if setup not in sequencers_year_setup[sequencer][year]:
#it means that this sequencer this year has not such setup
sys.stdout.write('0,')
else:
sys.stdout.write('{},'.format(sequencers_year_setup[sequencer][year][setup]))
sys.stdout.write('\n')
def year_bp_production():
couch = setupServer(CONFIG)
db_names = ['flowcells', 'x_flowcells']
flowcells = {}
production_stats = {}
for db_name in db_names:
flowcell_db = couch[db_name]
for fc_doc in flowcell_db:
if 'RunInfo' not in flowcell_db[fc_doc]:
continue
if 'Flowcell' not in flowcell_db[fc_doc]['RunInfo']:
continue
fc_name = flowcell_db[fc_doc]['RunInfo']['Flowcell']
if fc_name in flowcells:
continue
else:
flowcells[fc_name] = 0
year = int(flowcell_db[fc_doc]['RunInfo']['Date'][0:2])
month = int(flowcell_db[fc_doc]['RunInfo']['Date'][2:4])
if year < 12:
continue
yield_MBases = 0
if 'illumina' not in flowcell_db[fc_doc]:
continue
if 'Demultiplex_Stats' not in flowcell_db[fc_doc]['illumina']:
continue
if db_name == "x_flowcells":
if 'Flowcell_stats' not in flowcell_db[fc_doc]['illumina']['Demultiplex_Stats']:
continue
if 'Yield (MBases)' not in flowcell_db[fc_doc]['illumina']['Demultiplex_Stats']['Flowcell_stats']:
continue
yield_MBases = int(flowcell_db[fc_doc]['illumina']['Demultiplex_Stats']['Flowcell_stats']['Yield (MBases)'].replace(',', ''))
else:
if 'Barcode_lane_statistics' not in flowcell_db[fc_doc]['illumina']['Demultiplex_Stats']:
continue
for sample in flowcell_db[fc_doc]['illumina']['Demultiplex_Stats']['Barcode_lane_statistics']:
yield_MBases += int(sample['Yield (Mbases)'].replace(',', ''))
if year not in production_stats:
production_stats[year] = month_production = [0]*12
production_stats[year][month-1] += yield_MBases
sys.stdout.write(',')
for year in sorted(production_stats):
sys.stdout.write('{},'.format(year))
sys.stdout.write('\n')
for month in range(0,12,1):
sys.stdout.write('{},'.format(month+1))
for year in sorted(production_stats):
sys.stdout.write('{},'.format(production_stats[year][month]))
sys.stdout.write('\n')
sys.stdout.write('\n')
def main(args):
configuration_file = args.config
load_yaml_config(configuration_file)
configuration_file = args.config
load_yaml_config(configuration_file)
if args.mode == 'production-stats':
projects = parse_flowcell_db()
if args.mode == 'instrument-usage':
instrument_usage()
if args.mode == 'year-stats':
year_bp_production()
if __name__ == '__main__':
parser = argparse.ArgumentParser("""This scripts queries statusdb x_flowcelldb and project database and fetches informations about what organisms have been sequenced. It can be run in the following modes:
- production-stats: for each instrument type it prints number of FCs, number of lanes, etc. It then prints a summary of all stats
- instrument-usage: for each instrument type and year it prints different run set-ups and samples run with that set-up
- year-stats: cumulative data production by month
""")
parser.add_argument('--config', help="configuration file", type=str, required=True)
parser.add_argument('--mode', help="define what action needs to be executed", type=str, required=True, choices=('production-stats', 'instrument-usage', 'year-stats'))
args = parser.parse_args()
main(args)