1
|
#!/usr/bin/env python
|
2
|
# Scrubs the taxonlabels in VegBIEN using TNRS.
|
3
|
|
4
|
# to estimate total runtime:
|
5
|
# bin/psql_vegbien <<<'SELECT COUNT(*) FROM tnrs_input_name'
|
6
|
# # names from above * 1.5 multiplier for scrubbing accepted names
|
7
|
# (the test_taxonomic_names sample from Brad produces 8 accepted names for
|
8
|
# 15 input names)
|
9
|
# * ((# ms/name from log file * 1 sec/1000 ms) + (# sec to run
|
10
|
# `SELECT * FROM "tnrs_input_name"` in log file / tnrs.max_names names/batch))
|
11
|
# * 1 hr / 3600 sec * 1 day / 24 hr = # days
|
12
|
|
13
|
import os.path
|
14
|
import sys
|
15
|
|
16
|
sys.path.append(os.path.dirname(__file__)+"/../lib")
|
17
|
|
18
|
import csvs
|
19
|
import opts
|
20
|
import profiling
|
21
|
import sql
|
22
|
import sql_gen
|
23
|
import sql_io
|
24
|
import strings
|
25
|
import tnrs
|
26
|
|
27
|
tnrs_input = sql_gen.Table('tnrs_input_name')
|
28
|
tnrs_data = sql_gen.Table('tnrs')
|
29
|
|
30
|
def main():
|
31
|
# Input
|
32
|
env_names = []
|
33
|
db_config = opts.get_env_vars(sql.db_config_names, None, env_names)
|
34
|
verbosity = float(opts.get_env_var('verbosity', 3, env_names))
|
35
|
if not 'engine' in db_config: raise SystemExit('Usage: '
|
36
|
+opts.env_usage(env_names)+' '+sys.argv[0]+' 2>>log')
|
37
|
|
38
|
def log(msg, level=1):
|
39
|
'''Higher level -> more verbose'''
|
40
|
if level <= verbosity:
|
41
|
sys.stderr.write(strings.to_raw_str(msg.rstrip('\n')+'\n'))
|
42
|
|
43
|
# Connect to DB
|
44
|
db = sql.connect(db_config, log_debug=log)
|
45
|
|
46
|
cumulative_tnrs_profiler = profiling.ItersProfiler(iter_text='name')
|
47
|
|
48
|
# Iterate over unscrubbed verbatim taxonlabels
|
49
|
while True:
|
50
|
# Fetch next set
|
51
|
cur = sql.select(db, tnrs_input, limit=tnrs.max_names, cacheable=False)
|
52
|
this_ct = cur.rowcount
|
53
|
log('Processing '+str(this_ct)+' taxonlabels')
|
54
|
if this_ct == 0: break
|
55
|
# otherwise, rows found
|
56
|
names = list(sql.values(cur))
|
57
|
|
58
|
def process():
|
59
|
# Run TNRS
|
60
|
log('Making TNRS request')
|
61
|
stream = tnrs.tnrs_request(names,
|
62
|
cumulative_profiler=cumulative_tnrs_profiler)
|
63
|
|
64
|
log('Storing TNRS response data')
|
65
|
sql_io.append_csv(db, tnrs_data, *csvs.reader_and_header(stream))
|
66
|
# start transaction *before* submitting data, so Time_submitted is
|
67
|
# correctly set to the submission time rather than the insertion time.
|
68
|
# these may differ by several minutes if TNRS is slow.
|
69
|
sql.with_savepoint(db, process)
|
70
|
|
71
|
main()
|