Re: [RFR] python-levenshtein Description [ITP]
My gateway server's idea of an April Fools' Day prank is a hard
drive failure. It's a good thing I was building a replacement.
Nicolas François wrote:
> Justin B Rye wrote:
>> Description: extension for computing string similarities and edit distances
>> The Levenshtein module computes Levenshtein distances, similarity ratios,
>> generalized medians and set medians of ASCII or Unicode strings. Because
>> it's implemented in C, it's much faster than the corresponding Python
>> library functions and methods.
[...]
> I wonder if "ASCII or Unicode" is correct (I think it should work with any
> encoding). Unicode strings have a different type in Python, so the
> information that it support both is interesting (there are other python
> extension that do not support this feature).
>
> Would "Unicode and non-Unicode strings" be OK.
> (or "normal and Unicode strings")
I can't advise on the issue of which description is the best fit for
the way it works in Python, but I would suggest keeping the "or"
instead of another (potentially confusing) "and".
> The Levenshtein distance is the minimum number of insertion, deletion, or
> substitution of single characters to transform one string into the other.
Minimum number, but still plural! Make it:
The Levenshtein distance is the minimum number of single-character
insertions, deletions, and substitutions to transform one string into
another.
> It is useful in applications that need to determine how similar two
> strings are, such as spell checkers or fuzzy matching of gettext messages.
"Spell checkers" are (software) applications; "fuzzy matching of
gettext messages" isn't, so it's unbalanced. How about just:
It is useful for spell checking, or fuzzy matching of gettext messages.
(I'm astonished to find that "spellcheckers" is wrong - Google says
"Did you mean: spell checkers?" - even though they aren't checkers
for spells, they're things that perform spellchecking. Oh well...)
--
JBR with qualifications in linguistics, experience as a Debian
sysadmin, and probably no clue about this particular package
Reply to: