s396341/mass-scraper
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Fix get next utterance (by score).

Browse Source 
Switch to argparse in annotation_stats.py
This commit is contained in:
siulkilulki 2018-05-25 15:09:23 +02:00
parent 6a3819eb0a
commit 626307f135
2 changed files with 36 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

41
annotation_stats.py
View File

@ -1,6 +1,7 @@
#!/usr/bin/env python3 #!/usr/bin/env python3
import sys import sys
import argparse
import redis import redis
from extractor.find_hours import color_hour from extractor.find_hours import color_hour
import pickle import pickle
@ -77,18 +78,42 @@ def print_stats():
print('Annotated utterances: {}'.format(len(annotated))) print('Annotated utterances: {}'.format(len(annotated)))
def get_args():
parser = argparse.ArgumentParser()
subparser = parser.add_subparsers(dest='cmd')
parser_stats = subparser.add_parser('stats', help='Show annotation stats.')
parser_investigate = subparser.add_parser(
'investigate', help='investigate cookie.')
parser_investigate.add_argument('cookie', help='User cookie string')
parser_index = subparser.add_parser('index', help='Print utterance.')
parser_index.add_argument('index', type=int, help='Utterance index')
subparser.add_parser('ipdb', help='Get into ipdb.')
parser_exec = subparser.add_parser('exec', help='Execute redis command.')
parser_exec.add_argument(
'redis_command',
help=
'Redis command (lowercased). e.g. to get r.zrangebyscore("key", "-inf", "inf", start=0, num=1) pass \'zrangebyscore("key", "-inf", "inf", start=0, num=1)\''
)
return parser.parse_args()
def main(): def main():
if sys.argv[1] == 'stats': args = get_args()
if args.cmd == 'stats':
print_stats() print_stats()
elif sys.argv[1] == 'investigate': elif args.cmd == 'investigate':
investigate_by_cookie(sys.argv[2]) investigate_by_cookie(args.cookie)
elif sys.argv[1] == 'index': elif args.cmd == 'index':
pprint_utterance(int(sys.argv[2])) pprint_utterance(args.index)
elif sys.argv[1] == 'console': elif args.cmd == 'ipdb':
import ipdb import ipdb
ipdb.set_trace() ipdb.set_trace()
elif sys.argv[1] == 'exec': elif args.cmd == 'exec':
exec('print(r.{})'.format(sys.argv[2]), {'print': print, 'r': r}) exec('print(r.{})'.format(args.redis_command), {
'print': print,
'r': r
})
if __name__ == '__main__': if __name__ == '__main__':

6
webapp/app.py
View File

@ -65,13 +65,13 @@ def get_next(cookie_hash):
"""returns utterance with minmum annotations if that utterance """returns utterance with minmum annotations if that utterance
wasn't annotated by cookie_hash user wasn't annotated by cookie_hash user
or not yet annotated utterance by cookie_hash user""" or not yet annotated utterance by cookie_hash user"""
index = int(r.zrangebyscore(UTT_SCORES, '-inf', 'inf')[1]) index = int(r.zrangebyscore(UTT_SCORES, '-inf', 'inf', start=0, num=1)[0])
if r.exists(f'{cookie_hash}:{index}'): if r.exists(f'{cookie_hash}:{index}'):
index = find_not_annotated(cookie_hash) index = find_not_annotated(cookie_hash)
log('found unannotated index: {}'.format(index)) log('found unannotated index: {}'.format(index))
left_context, hour, right_context = get_utterance_for_web(index) left_context, hour, right_context = get_utterance_for_web(index)
# log('get_next index: {}, score: {}'.format(index, log('get_next index: {}, score: {}'.format(index,
# r.zscore(UTT_SCORES, index))) r.zscore(UTT_SCORES, index)))
return index, left_context, hour, right_context return index, left_context, hour, right_context