[NTG-context] Keeping track of hyphenations

Hans Hagen j.hagen at xs4all.nl
Fri Mar 19 22:41:51 CET 2021


On 3/19/2021 9:49 PM, denis.maier at ub.unibe.ch wrote:
> Wow, Hans, that looks awesome. Thanks a lot.
> 
> Regarding the first output: What are the numbers intended to mean?

The number of time that this word was used hyphenated (the same word can 
of course end up hyphenated differently in the paragraph).

> In the second example we're talking in terms of lines in the input file, right?

No, the linenumber but because I use a loop it's all the same line; in 
practice these numbers point you to the paragraph start which is good 
enough.

> If so, would it make sense to add some more context? Like so, or similar:
> 
> hyhenated       > around line 244 ( "our marvelous and everyday capacity to select") in file hup.tex: every-day

As said, normally you have that info (I can probably add the offset but 
i assume that one is not going to check every hyphenation.)

> In any case, that's already super conventient. I'll put together a script to compare hyphenations with a whitelist and upload that to the wiki once it's ready.
> 
> What's the performance hit of all this? (I mean should you disable this and only run when needed, or can you safely leave the trackers enabled?)

measures on my 2013 laptop: a 1000 times \samplefile{tufte}\par takes:

pagella:

1000 steps, 4.140 seconds, 0.004140215 per step
1000 steps, 4.222 seconds, 0.004221645 per step

runtime: 4.937 seconds, 250 processed pages, 250 shipped pages, 50.635 
pages/second

latin modern:

1000 steps, 2.947 seconds, 0.002947476 per step
1000 steps, 3.043 seconds, 0.003042808 per step

runtime: 3.549 seconds, 223 processed pages, 223 shipped pages, 62.840 
pages/second

so not really a problem and i can probaby make it a bit faster (but we 
need to keep up with the urban myth that context is the slower system)

(it's not something i'd turn on by default)


> Thanks again for your help!
> 
> Denis
> 
> -----Ursprüngliche Nachricht-----
> Von: Hans Hagen <j.hagen at xs4all.nl>
> Gesendet: Freitag, 19. März 2021 19:09
> An: Maier, Denis Christian (UB) <denis.maier at ub.unibe.ch>; ntg-context at ntg.nl
> Betreff: Re: AW: [NTG-context] Keeping track of hyphenations
> 
> On 3/19/2021 1:00 PM, denis.maier at ub.unibe.ch wrote:
>> ----Ursprüngliche Nachricht-----
>> Von: Hans Hagen <j.hagen at xs4all.nl>
>> Gesendet: Freitag, 19. März 2021 12:44
>> An: mailing list for ConTeXt users <ntg-context at ntg.nl>; Maier, Denis
>> Christian (UB) <denis.maier at ub.unibe.ch>
>> Betreff: Re: [NTG-context] Keeping track of hyphenations
>>
>> On 3/19/2021 11:47 AM, denis.maier at ub.unibe.ch wrote:
>>> Hi
>>>
>>> Is there  a way to conveniently track hyphenations ? I've had a look
>>> at https://wiki.contextgarden.net/Trackers
>>> <https://wiki.contextgarden.net/Trackers> but couldn't find anything.
>> depends on what you want to track
>>
>> Something like this: Add every hyphenated word to the log file, like
>> hyphenations > 'What-ever' , realpage 1, userpage 1, subpage 1
>> hyphenations > 'Some-thing', realpage 2, userpage 2, subpage 2
>>
>> You could then feed that into a script to produce a lists of all kinds, e.g., each hyphenation only once, sorted by page or by letter, with or without information about pages.
> Forget about specific pagenumbers ... pages can get broken in the middle
>     of a paragraph and it's not worth the trouble of keeping track of / analyzing that detail (as usual all can be done but these are features that hardly anyone uses, get forgotten and probably no one can find where it's documents because no one expects it).
> 
> Anyway, a pre-weekend distaction:
> 
> \starttext
> 
> \enabletrackers[hyphenation.applied]
> \enabletrackers[hyphenation.applied.console]
> 
> \dontcomplain \dostepwiserecurse{0}{15}{1}{
>       {\advance\hsize -#1cm \input tufte \par} }
> 
> \stoptext
> 
> This will give you in the log file (first and second); the test run has narrow lines where one word crossed several lines, which actually is the main complication here:
> 
> hyhenated       > start hyphenated words
> 
> hyhenated       >    2 : ab-stract,
> hyhenated       >    1 : ag-gre-gate,
> hyhenated       >    3 : ag-gregate,
> hyhenated       >    3 : aggre-gate,
> hyhenated       >    1 : ap-prox-i-mate,
> hyhenated       >    2 : ap-proximate,
> hyhenated       >    3 : approx-imate,
> hyhenated       >    1 : approxi-mate,
> hyhenated       >    1 : av-er-age,
> hyhenated       >    1 : av-erage,
> hyhenated       >    3 : be-cause
> hyhenated       >    1 : ca-pac-ity
> hyhenated       >    4 : ca-pacity
> hyhenated       >    1 : capac-ity
> hyhenated       >    1 : cat-a-log,
> hyhenated       >    2 : cat-alog,
> hyhenated       >    1 : cat-e-go-rize,
> hyhenated       >    3 : cat-egorize,
> hyhenated       >    1 : cata-log,
> hyhenated       >    1 : cate-gorize,
> hyhenated       >    2 : catego-rize,
> hyhenated       >    3 : clas-sify,
> hyhenated       >    3 : clus-ter,
> hyhenated       >    3 : con-dense,
> hyhenated       >    1 : dis-crim-i-nate,
> hyhenated       >    1 : dis-crimi-nate,
> hyhenated       >    1 : dis-criminate,
> hyhenated       >    1 : dis-tin-guish,
> hyhenated       >    3 : dis-tinguish,
> hyhenated       >    2 : discrim-inate,
> hyhenated       >    2 : distin-guish,
> hyhenated       >    1 : enu-mer-ate,
> hyhenated       >    2 : enu-merate,
> hyhenated       >    1 : enumer-ate,
> hyhenated       >    5 : every-day
> hyhenated       >    3 : fil-ter,
> hyhenated       >    2 : fo-cus,
> hyhenated       >    1 : har-mo-nize,
> hyhenated       >    1 : har-monize,
> hyhenated       >    2 : harmo-nize,
> hyhenated       >    6 : high-light,
> hyhenated       >    1 : ide-al-ize,
> hyhenated       >    2 : ide-alize,
> hyhenated       >    1 : ideal-ize,
> hyhenated       >    1 : in-for-ma-tion
> hyhenated       >    1 : in-forma-tionthick
> hyhenated       >    1 : in-formationthick
> hyhenated       >    3 : in-spect,
> hyhenated       >    1 : in-te-grate,
> hyhenated       >    2 : in-tegrate,
> hyhenated       >    1 : informa-tionthick
> hyhenated       >    3 : inte-grate,
> hyhenated       >    5 : iso-late,
> hyhenated       >    4 : item-ize,
> hyhenated       >    3 : mar-velous
> hyhenated       >    1 : or-ga-nize,
> hyhenated       >    1 : orga-nize,
> hyhenated       >    5 : out-line,
> hyhenated       >    1 : pi-geon-hole,
> hyhenated       >    6 : pigeon-hole,
> hyhenated       >    3 : re-duce,
> hyhenated       >    4 : re-fine,
> hyhenated       >    2 : re-view,
> hyhenated       >    4 : se-lect,
> hyhenated       >    1 : sep-a-rate
> hyhenated       >    1 : sepa-rate
> hyhenated       >    3 : sin-gle
> hyhenated       >    4 : struc-ture,
> hyhenated       >    1 : sum-ma-rize,
> hyhenated       >    4 : sum-marize,
> hyhenated       >    1 : summa-rize,
> hyhenated       >    1 : syn-op-size,
> hyhenated       >    2 : syn-opsize,
> hyhenated       >    1 : syn-the-size,
> hyhenated       >    3 : syn-thesize,
> hyhenated       >    1 : synop-size,
> hyhenated       >    2 : synthe-size,
> hyhenated       >    5 : win-now
> 
> hyhenated       > stop hyphenated words
> 
> and on the console and log (second tracker) lots of
> 
> pages           > flushing realpage 2, userpage 2, subpage 2, total
> 0.482, page 0.011, average 0.482
> open source     > level 2, order 13, name 'tufte.tex'
> close source    > level 2, order 13, name 'tufte.tex'
> hyhenated       > around line 244 in file hup.tex: every-day
> hyhenated       > around line 244 in file hup.tex: cat-egorize,
> hyhenated       > around line 244 in file hup.tex: dis-criminate,
> hyhenated       > around line 244 in file hup.tex: pigeon-hole,
> hyhenated       > around line 244 in file hup.tex: aggre-gate,
> hyhenated       > around line 244 in file hup.tex: re-view,
> hyhenated       > around line 244 in file hup.tex: enu-merate,
> open source     > level 2, order 14, name 'tufte.tex'
> close source    > level 2, order 14, name 'tufte.tex'
> hyhenated       > around line 244 in file hup.tex: every-day
> hyhenated       > around line 244 in file hup.tex: synthe-size,
> hyhenated       > around line 244 in file hup.tex: re-duce,
> hyhenated       > around line 244 in file hup.tex: catego-rize,
> hyhenated       > around line 244 in file hup.tex: inte-grate,
> hyhenated       > around line 244 in file hup.tex: ap-proximate,
> hyhenated       > around line 244 in file hup.tex: out-line,
> hyhenated       > around line 244 in file hup.tex: win-now
> 
> if you want more you need to analyze the log file
> 
> (you have to wait till i update, which will not happen today)
> 
> Hans
> 
> 
> -----------------------------------------------------------------
>                                             Hans Hagen | PRAGMA ADE
>                 Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
>          tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl
> -----------------------------------------------------------------
> 


-- 

-----------------------------------------------------------------
                                           Hans Hagen | PRAGMA ADE
               Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
        tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl
-----------------------------------------------------------------


More information about the ntg-context mailing list