[New-bugs-announce] [issue31889] difflib SequenceMatcher ratio() still have unpredictable behavior
Simon Descarpentries
report at bugs.python.org
Sat Oct 28 06:03:33 EDT 2017
New submission from Simon Descarpentries <siltaar at acoeuro.com>:
I, it's my 1st post here. I'm a French computer-science engineer with 10 years XP and manager at Acoeuro.com SSLL compagny. I suggested a better regexp integration on python-ideas a few months ago failing to convince getting things done.
Despite issue https://bugs.python.org/issue25391 closed in 2010, nothing seems visible on https://docs.python.org/3/library/difflib.html to help users anticipating that a string greatly matching at 199 characters length, won't at all at the 200th one. It's an inconsistent behavior that looks like a bug.
#!/usr/bin/env python3
from difflib import SequenceMatcher
a = 'ab'*400
b = 'ba'*400
for i in range(1, 400):
diff_ratio = SequenceMatcher(None, a=a[:i], b=b[:i]).ratio()
print('%3.i %.2f' % (i, diff_ratio), end=' ')
not i % 10 and print('')
EOF
At 199c I have a 0.99 ratio, and 0.00 at 200c. The results are nearly the same with strings made of emails like in https://github.com/Siltaar/drop_alternatives especially comparing strings like :
"performantsetdvelopperducontenusimilairepourvosprochainsTweets.Suiveznous at TwitterBusinesspourdautresinfosetactus.TesterlesPublicitsTwitterbusiness.twitter.com|@TwitterBusiness|Aide|SedsinscrireLemailsadresse at gggTwitter,Inc.MarketStreet,SuiteSanFrancisco,CA"
"rducontenusimilairepourvosprochainsTweets.Suiveznous at TwitterBusinesspourprofiterdautresinfosetactus.TesterlesPublicitésTwitterbusiness.twitter.com at TwitterBusinessAideSedésinscrireTwitterInternationalCompanyOneCumberlandPlace,FenianStreetDublin,DAXIRELAND"
Fortunately, I didn't experienced the problem using quick_ratio().
The documentation is not clear about ratio / quick_ratio / real_quick_ratio ; but looks like unstable. With in addition an inconsistent behavior it looks like worthing some fix.
----------
components: Library (Lib)
messages: 305153
nosy: Siltaar
priority: normal
severity: normal
status: open
title: difflib SequenceMatcher ratio() still have unpredictable behavior
type: behavior
versions: Python 2.7, Python 3.5, Python 3.6
_______________________________________
Python tracker <report at bugs.python.org>
<https://bugs.python.org/issue31889>
_______________________________________
More information about the New-bugs-announce
mailing list