Gdtyjr

I would like to obtain the following behavior: If an acronym is only used once within a section/chapter/part, at this occurrence only the long name should be used. If however the acronym is used more than one, I want the long-short behavior.

I have the following non working MWE:

documentclassbook
usepackagexparse
ExplSyntaxOn
NewExpandableDocumentCommandfpcompare m m m 

 % #1 = test to perform
 % #2 = text for the true case
 % #3 = text for the false case
 fp_compare:nTF #1 #2 #3 

ExplSyntaxOff

usepackage[acronym,symbols,nomain,toc,sanitizesort,nogroupskip,xindy,numberedsection=autolabel]glossaries-extra
setglossarystylesuper
setabbreviationstyle[acronym]long-short

GlsXtrEnableEntryUnitCountingacronym1section
%renewcommand*glscgls%
%renewcommand*GlscGls%
%renewcommand*glsplcglspl%
%renewcommand*GlsplcGlspl%
%renewcommand*GLScGLS%
%renewcommand*GLSplcGLSpl%


makenoidxglossaries
newacronymwnWNwireless network

glsxtrnewsymbol[description=pi,type=symbols]symb:piensuremathpi%


usepackagexstring
renewcommand*gls[1]%
 IfEqglscategory#1acronym%to apply this to acronym only (and not symbols as well)
 fpcompareglsentrycurrcount#1 > 1glsentrydesc#1cgls#1% fpcompare to detect how often used 
 %
 cgls#1%
 %
%
renewcommand*Gls[1]%
 IfEqglscategory#1acronym%
 fpcompareglsentrycurrcount#1 > 1Glsentrydesc#1cgls#1%
 %
 cgls#1%
 %
%
renewcommand*glspl[1]%
 IfEqglscategory#1acronym%
 fpcompareglsentrycurrcount#1 > 1glsentryplural#1cglspl#1%
 %
 cglspl#1%
 %
%
renewcommand*Glspl[1]%
 IfEqglscategory#1acronym%
 fpcompareglsentrycurrcount#1 > 1Glsentryplural#1cglspl#1%
 %
 cglspl#1%
 %
%
renewcommand*GLS[1]%
 IfEqglscategory#1acronym%
 fpcompareglsentrycurrcount#1 > 1GLSentrydesc#1cgls#1%
 %
 cgls#1%
 %
%
renewcommand*GLSpl[1]%
 IfEqglscategory#1acronym%
 fpcompareglsentrycurrcount#1 > 1GLSentryplural#1cglspl#1%
 %
 cglspl#1%
 %
%

begindocument
 sectionA
 glswn\
 glsplwn\
 glsplwn\
 glsplwn\
 count: glsentrycurrcountwn\
 glssymb:pi
 sectionB 

 glsplwn\
 count: glsentrycurrcountwn

 sectionC
 Glsplwn\
 glswn\
 count: glsentrycurrcountwn
enddocument

which produces:

I however, would have expected/wanted a count of 4 in section A, and in section C, that at first the full description is printed instead of WNs

How to make this work reliably?

You could make acronyms section-specific by giving them section-specific names. Of course you could easily automate this:

documentclassbook
usepackage[xindy,style=long,numberline,savewrites=true,acronym,nomain]glossaries
usepackagepgffor

makeglossaries
glsenableentrycount 

%create a separate acrony for each section
letoldnewacronymnewacronym renewcommandnewacronym[3]foreach n in 0,...,10oldnewacronymcn-#1#2#3

%%patch all the commands to use the within-section version of the acronym
letoldcglscgls renewcommandcgls[1]oldcglssarabicsection-#1
letoldcglsplcglspl renewcommandcglspl[1]oldcglsplsarabicsection-#1
letoldcGlsplcGlspl renewcommandcGlspl[1]oldcGlsplsarabicsection-#1

newacronymwnWNwireless network

begindocument
 sectionA
 cglswn\
 cglsplwn\
 cglsplwn\
 cglsplwn\
 count: glsentrycurrcounts1-wn\

 sectionB 
 cglsplwn\
 count: glsentrycurrcounts2-wn

 sectionC
 cGlsplwn\
 cglswn\
 count: glsentrycurrcounts3-wn 
enddocument

Basically, when the acronym wn is defined, it creates separate acronyms for each section, naming them s1-wn and so on. Now when the acronym is called within section X, it "redirects" and uses sX-wn. The result:

Of course this would need some more tweaking:

1. So far it only works for sections, not chapters, but that should be straighforward


2. It is hard-wired to work for up to 10 sections. You could increase that number, introduce a new counter, or do some .aux-file-magic to automatize this to work for any number of chapters/section/subsections


3. It also needs some further tweaking to deal with symbols and so on.


4. I haven't tried it with glossaries-extra so far, because I don't have that installed.


5. I don't now how you want your list of abbreviations to appear. So for now my proposal ignores this, as did your question.

Credits: Part of my solution is taken from here