Re: [xsl] Bibliography author repetition handling

Subject: Re: [xsl] Bibliography author repetition handling
From: "Wendell Piez wapiez@xxxxxxxxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>
Date: Mon, 9 Oct 2017 21:21:47 -0000
Eliot,

Chuck may indeed want a function to normalize his associating value
(identifying the lead author), from which is-a-repetition logic is
derived. However, your suggestion takes us down a different pathway
altogether:

<xsl:function name="is-a-repetition" as="xs:boolean">
  <xsl:param name="who" as="element(bibl)"/>
  <xsl:sequence
   select="local:cited-author(.) =
preceding-sibling::bibl[1]/local:cited-author(.)"/>
</xsl:function>

Now we don't need the grouping any more at the XSLT level, we just
match it the way you have.

I dare say it may not be as efficient under the hood as the grouping
approach (Syd what do you think?) but over Dr Muller's runtime x data
set, that might not matter much.

Cheers, Wendell




On Mon, Oct 9, 2017 at 2:05 PM, Eliot Kimber ekimber@xxxxxxxxxxxx
<xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:
> Yes, in this case different modes might be overkill for just this one
difference. Itbs just that passing a parameter to do then do a choose inside
a template just doesnbt feel right. Therebs an argument to be made for
consistency of approachbin my experience these things tend to not be just
one, but the first of many to comeb&
>
> If the check was one that could be easily put into a function, which this
one does not seem to be, then I might use that function in the match
expression rather than modes:
>
> <xsl:template match=bbibl[local:is-a-repetition(.)]b>
>
> Cheers,
>
> Eliot
> --
> Eliot Kimber
> http://contrext.com
>
>
>
> On 10/9/17, 10:35 AM, "Wendell Piez wapiez@xxxxxxxxxxxxxxx"
<xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:
>
>     Eliot didn't I say "you might prefer a mode"? :-)
>
>     Just to remark I agree with your proposal to prefer modes over
>     parameters for this kind of thing -- most of the time.
>
>     This is a case where the only difference between the two pathways over
>     bibl contents are identical except for a single element ... since
>     bibl/* includes a lot, that means a fair amount of 'mode overhead' to
>     handle the single difference.
>
>     For that reason I might be willing to put up with an xsl:choose here,
>     or functional equivalent ...
>
>     <xsl:template match="bibl/author[1]">
>       <xsl:param name=brepeatb tunnel=byesb select=bfalse()b/>
>       <xsl:apply-templates mode="repeating" select=".[$repeat]"/>
>       <xsl:if test="not($repeat)">
>          ...
>       </xsl:if>
>     </xsl:template>
>
>     (Okay, I might leave a comment to help.)
>
>     Cheers, Wendell
>
>     On Mon, Oct 9, 2017 at 11:13 AM, Eliot Kimber ekimber@xxxxxxxxxxxx
>     <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:
>     > No argument with Wendellbs general solution, just one small coding
style twiddle:
>     >
>     > Instead of this:
>     >
>     >        <xsl:apply-templates select="current-group() except .">
>     >           <xsl:with-param tunnel="yes" name="repeat"
select="true()"/>
>     >        </xsl:apply-templates>
>     >
>     > I would probably use a distinct mode for the repeated items:
>     >
>     >        <xsl:apply-templates select="current-group() except ."
mode=brepeated-bib-itemsb/>
>     >
>     > Ibve come to accept that preferring templates and modes over passing
state is the best XSLT style.
>     >
>     > All the items that have no difference in behavior in the two modes can
simply match in both modes:
>     >
>     > <xsl:template match=bbiblb mode=b#default
repeated-bib-itemsb>
>     >   <xsl:apply-templates mode=b#current-modeb/>
>     > </xsl:template>
>     >
>     > Otherwise therebs going to have to be a choose instruction to check
the brepeatb parameter:
>     >
>     > <xsl:template b&>
>     >   <xsl:param name=brepeatb tunnel=byesb
select=bfalse()b/>
>     >  <xsl:choose>
>     >    <xsl:when test=b$repeatb>
>     >      b&
>     >   </xsl:when>
>     >   <xsl:otherwise>
>     >     b&
>     >   </xsl:otherwise>
>     > </choose>
>     > </xsl:template>
>     >
>     > Better to put the choice in the mode and/or match expression.
>     >
>     > Cheers,
>     >
>     > E.
>     > --
>     > Eliot Kimber
>     > http://contrext.com
>     >
>     >
>     >
>     > On 10/9/17, 9:58 AM, "Wendell Piez wapiez@xxxxxxxxxxxxxxx"
<xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:
>     >
>     >     Hi,
>     >
>     >     I think Chuck wants group-adjacent.
>     >
>     >     Chuck, what Martin is hinting is that rather than testing each
bibl
>     >     (each as you match it) to see whether it is a "repeat" or not,
you
>     >     first group all the bibls by their (analytic) author, then every
>     >     member of the group after the first is a repeat.
>     >
>     >     (Define "repeat" as "bibl that should get '----' instead of a
lead
>     >     author's name by virtue of its having the same lead author as its
>     >     immediate predecessor".)
>     >
>     >     This means that most of your groups will have only one member,
but
>     >     once in a while (when authors are repeated in runs) you will have
a
>     >     group of two or even more members, only the first of which is not
a
>     >     repeat.
>     >
>     >     Due to a design quirk in XSLT (semantics of "." inside groups),
you
>     >     can code this fairly elegantly:
>     >
>     >     <xsl:for-each-group select="bibl"
group-adjacent="gimme-author()">
>     >        <!-- apply templates to the first member -->
>     >        <xsl:apply-templates select=".">
>     >        <!-- apply templates to the rest w/ a param -->
>     >        <xsl:apply-templates select="current-group() except .">
>     >           <xsl:with-param tunnel="yes" name="repeat"
select="true()"/>
>     >        </xsl:apply-templates>
>     >     </xsl:for-each-group>
>     >
>     >     Notes:
>     >
>     >     * No element is created for the group since the grouping is only
to
>     >     sort them into categories.
>     >
>     >     * gimme-author() will be whatever XPath (path or other
expression)
>     >     indicates the value to be grouped on (i.e. the leading author
name
>     >     however normalized).
>     >
>     >     * inside the group, "." refers only to the first member of the
group,
>     >     so "current-group() except ." returns all the others.
>     >
>     >     * Leaving it to you to figure out what to do at the receiving end
with
>     >     the Boolean parameter $repeat ... although I've indicated
>     >     "tunnel='yes'" to leave another clue as to how it might be done.
Or
>     >     instead of a parameter you could also use a mode here to
distinguish
>     >     between handling the normal cases and the repeat cases.
(Presumably
>     >     they are the same except for how they treat that author.)
>     >
>     >     * group-by will have a similar effect to group-adjacent but only
if
>     >     the nodes are already in their final sorted order; if there is
some
>     >     anomaly in the listing, group-adjacent is more robust (and will
>     >     reflect that anomaly not reorder around it).
>     >
>     >     It's been a few days, so you may already have tackled this. Still
it's
>     >     a fun little problem.
>     >
>     >     Cheers, Wendell
>     >
>     >
>     >
>     >     On Wed, Oct 4, 2017 at 4:18 AM, Martin Honnen
martin.honnen@xxxxxx
>     >     <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:
>     >     > Am 04.10.2017 um 08:25 schrieb Charles Muller
acmuller@xxxxxxxxxxxxxxx:
>     >     >>
>     >     >> In a TEI <biblStruct> bibliography that uses
biblStruct/*/author/surname
>     >     >> (etc), I have managed to write the code that checks if the
author name in
>     >     >> the prior <biblStruct> is the same, in which case the output is
---. rather
>     >     >> than the author's name. It works fine, but I'm running into a
problem where
>     >     >> there the first author name is the same as that of the prior
<biblStruct>
>     >     >> entry, but there are also multiple authors, in which case the
full name
>     >     >> should be printed out.
>     >     >>
>     >     >> I thought that one way I might resolve this is to test if the
number of
>     >     >> authors in the prior entry is the same as this one, but I can't
get it to
>     >     >> work. I also don't know if this is even the best way of
handling this. The
>     >     >> XML is like this:  (present output here:
>     >     >> http://www.acmuller.net/articles-shisou.html)
>     >     >
>     >     >
>     >     > To me this sounds like a grouping problem where you want to
group your
>     >     > biblStruct elements with a composite key mad up of
analytic/author, I am not
>     >     > sure whether you want group-adjacent or group-by.
>     >     >
>     >     >
>     >
>     >
>     >
>     >     --
>     >     Wendell Piez | http://www.wendellpiez.com
>     >     XML | XSLT | electronic publishing
>     >     Eat Your Vegetables
>     >     _____oo_________o_o___ooooo____ooooooo_^
>     >
>     >
>     >
>     >
>
>
>
>     --
>     Wendell Piez | http://www.wendellpiez.com
>     XML | XSLT | electronic publishing
>     Eat Your Vegetables
>     _____oo_________o_o___ooooo____ooooooo_^
>
>
>
>



--
Wendell Piez | http://www.wendellpiez.com
XML | XSLT | electronic publishing
Eat Your Vegetables
_____oo_________o_o___ooooo____ooooooo_^

Current Thread