jpt-archive/oyster
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

random sampling

Browse Source
This commit is contained in:
James Turk 2012-05-15 17:40:46 -04:00
parent b4de2ee0f9
commit 4172b43c0f
1 changed files with 6 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
oyster/scripts/signal.py
View File

@ -1,6 +1,7 @@
#!/usr/bin/env python #!/usr/bin/env python
import argparse import argparse
import traceback import traceback
import random
from celery.execute import send_task from celery.execute import send_task
from celery import current_app from celery import current_app
@ -22,14 +23,15 @@ def main():
docs = kernel.db.tracked.find({'doc_class': args.doc_class, docs = kernel.db.tracked.find({'doc_class': args.doc_class,
'versions': {'$ne': []} 'versions': {'$ne': []}
}, timeout=False) }, timeout=False)
print '%s docs in %s' % (docs.count(), args.doc_class) total = docs.count()
print '{0} docs in {1}'.format(total, args.doc_class)
if args.sample: if args.sample:
print 'sampling 100 documents' limit = 100
docs = docs.limit(100) print 'sampling {0} documents'.format(limit)
docs = docs.limit(limit).offset(random.randint(0, total-limit))
args.immediate = True args.immediate = True
total = docs.count()
errors = 0 errors = 0
if args.immediate: if args.immediate: