Re: [xsl] Spell Check Type Matching in XPath?

["C. M. Sperberg-McQueen cmsmcq@xxxxxxxxxxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>]

On Thu, 2022-04-21 at 19:01 +0000, Eliot Kimber
eliot.kimber@xxxxxxxxxxxxxx wrote:
> Ibm looking at Jenibs code now. Ibll see what I can do with it.
> B 
> The fact that this is the best there is (a MarkMail search basically
> brought me to Mikebs response below), it suggests that therebs not
> something more obvious that I simply failed to see.

Non-obvious (at least to me), but possibly faster, given that you know
already one of the strings to be matched, may be the symmetric-deletion
approach to edit distance described by Wolf Garber [1]. It allows a
fairly quick detection of whether the candidate string is within edit
distance  1 of the string you're looking to match -- if you adjust the
way you do it, you can detect strings within distance 2.

[1]
https://wolfgarbe.medium.com/1000x-faster-spelling-correction-algorithm-2012-8701fcd87a5f

Michael Sperberg-McQueen

B 
> However, Jenibs comments in her post about recursion suggests therebs
> a way to improve the code in XSLT 3/XPath 3, maybe something using
> iterateb&.
> B 
> Cheers,
> B 
> E.
> B 
> _____________________________________________
> Eliot Kimber
> Sr Staff Content Engineer
> O: 512 554 9368
> M: 512 554 9368
> servicenow.com
> LinkedInB |B TwitterB |B YouTubeB |B Facebook
> B 
> From:Michael Kay mike@xxxxxxxxxxxx
> <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>
> Date: Thursday, April 21, 2022 at 1:35 PM
> To: xsl-list <xsl-list@xxxxxxxxxxxxxxxxxxxxxx>
> Subject: Re: [xsl] Spell Check Type Matching in XPath?
> [External Email]
> B 
> Jeni Tennison's work on computing Levenshtein distance in XSLT may be
> relevant:
> B 
> http://www.jenitennison.com/2007/05/03/levenshtein-distance-in-xslt-2-0.html
> B 
> (It would be interesting to see it reworked for XSLT 3.0...)
> B 
> Search also for "Levenshtein distance XSLT" on Markmail.
> B 
> Michael Kay
> Saxonica
> 
> 
> > On 21 Apr 2022, at 18:57, Eliot Kimbereliot.kimber@xxxxxxxxxxxxxx
> > <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:
> > B 
> > Ibm writing a Schematron rule that tries to identify URLs where the
> > server component is close to, but not quite, bdocs.servicenow.comb,
> > i.e., bseivcenowb or bservcinowb or whatever. I also need to
> > eliminate servers that are not like servicenow, such as
> > bdocs.amazon.comb.
> > B 
> > Basically I want a the kind of fuzzy match on bservicenowb that
> > youbd get with normal spell checking.
> > B 
> > Ibm not seeing an easy way to do this in XSLT/XPath (in the context
> > of the XSLT Schematron engine in Oxygen XML).
> > B 
> > But I feel like Ibm missing some more-or-less obvious way to do
> > this with regular expression or maybe a fold or something (I can
> > use XPath 3).
> > B 
> > What am I missing?
> > B 
> > Thanks,
> > B 
> > E.
> > _____________________________________________
> > Eliot Kimber
> > Sr Staff Content Engineer
> > O: 512 554 9368
> > M: 512 554 9368
> > servicenow.com
> > LinkedInB |B TwitterB |B YouTubeB |B Facebook
> > XSL-List info and archive 
> > EasyUnsubscribeB (by email)
> B 
> XSL-List info and archive 
> EasyUnsubscribe (by email)
> XSL-List info and archiveEasyUnsubscribe(by email)