How to achieve a simple <p> element with LyXHTML

[Lorenzo Bertini]

Il 11/07/22 00:11, Steve Litt ha scritto:
> tush via lyx-users said on Sat, 09 Jul 2022 18:59:53 +0000
> 
>> I am trying to remove the html attributes assigned by LyXHTML to the
>> Standard text inserted in the editor.
>>
>> LyX's default is to render Standard as
>>
>> <div class="standard" id='magicparlabel-15'>This is text inserted in
>> Standard</div>
>>
>> together with
>>
>> p.standard {
>>
>> text-align: left;
>>
>> }
>>
>> appearing in the header.
>>
>> What I want is to get the simple html code
>>
>> <p>This is text inserted in Standard</p>
>> In order to do that I am writing my Style in a layout file. What I
>> have at the moment is the following:
>>
>> (I copied the first rows from stdclass.inc)
>>
>> Style Standard
>> 	Category              MainText
>> 	Margin                Static
>> 	LatexType             Paragraph
>> 	LatexName             dummy
>> 	ParIndent             MM
>> 	ParSkip               0.4
>> 	Align                 Block
>> 	AlignPossible         Block, Left, Right, Center
>> 	LabelType             No_Label
>>
>> HTMLAttr " "
>>
>> End

I think this exchange from lyx-devel will provide some explanation:
> On 9/6/21 6:36 PM, Lorenzo Bertini wrote:
>> Hello list!
>>
>> LyXHTML outputs standard paragraphs with  a <div>...</div> (as is the
>> standard for many insets), but i would much prefer <p>...</p>.
>>
>> I know i can change the tag of other insets by going in the local
>> layout section of Document->Setting, and write something like
>>> InsetLayout Foot
>>>   HTMLTag p
>>> End
>>
>> but i cant find anything about the standard paragraph. I don't mean
>> the "sectioning" paragraph, that is mapped to "HTMLTag h5", but the
>> piece of text you get after pressing RET in LyX.
> 
> It's a paragraph layout: Standard. See stdclass.inc.
> 
> But yes, at the moment, that is hardcoded. The reason is that the tag
> turns up inside other tags and nested <p>'s are invalid. But it probably
> should be made customizable.
> 
> RikiRemoval of this hardcoded stuff would require a bit of lyxhtml output 
rewrite, which should be under way right now. Unfortunately, I think 
it's just one dev working on it (Thibaut).

>>
>> After a lot of trial and error, using the last line I eliminated the
>> "div class="standard"" from the p element but still LyX assigns to it
>> an id="some number" which I want to eliminate too.
>>
>> Any idea about it would be very welcome.

If you're referring to the "magicparlabel", it's hardcoded. I was 
thinking about making a patch to remove it, but LyX relies on this 
"magic labels" to do a lot of the referencing (a quick glance at how the 
TOC is done will give you an idea). I am not that knowlodgeable about 
HTML referencing, so this will take some time.

> 
> Yeahhhhhh, this is one of the reasons I don't use LyX for ePub. Last
> time I looked, LyX exports used the deprecated and now with HTML5
> eliminated <a name="whatever"/> inside a div specially made to contain
> one <p></p>. What could *possibly* go wrong?
> 
> If I had to eliminate these redundant <div></div> things, I'd do it
> with a postprocessor that corrects LyX' complexifications. It's not
> easy, but you can use Python with the "import xml.etree.ElementTree as
> ET" library. You'd analyze every DOM element, and for every <div/>
> whose only purpose is to encase one <p/>, get rid of the div after
> reading metadata from that <div/> and any <a name=> silliness, applying
> the information to attributes of the <p/> itself.
> 
> Seriously, LyX has failed to export common sense XMLized HTML for 14
> years now, and I doubt they ever will. Best you can hope for is to
> either switch to another authoring program, or repair LyX' mistakes
> post-export. I think trying to repair this stuff from within your LyX
> document is walking the highway to heartache.
> 
> Also, the good thing about the XML-parsing Python post-processor
> approach is that, as time goes on, you can fix any problems newly
> discovered and/or newly added by LyX: You're in control as long as they
> continue to export well-formed XML HTML, which of course isn't a
> for-sure thing.
> 
> SteveT
> 
> Steve Litt
> Summer 2022 featured book: Making Mental Models: Advanced Edition
> http://www.troubleshooters.com/mmm

A solution of this kind could work well, but remove the thing that makes 
LyX so convenient: one click compile and show. We really need a robust 
LyXHTML output, and I don't think it's that far away.

Right now the devs lack the manpower to tackle LyXHTML problems fully, 
but I believe it's also a matter of demand. It's nice to see there are 
people on lyx-users asking about it.

Lorenzo