How to indent a block when exporting to reStructuredText

How to indent a block when exporting to reStructuredText

[Denis Bitouzé]

Hi,

(Disclaimer: I'm a pandoc beginner.)

I have to convert DokuWiki files containing some unusual (meaning from
plugins) tags that are exported verbatim to reStructuredText. For
instance, one of these tags is:

  ┌────
  │ <note>
  │ Foo
  │ 
  │ Bar
  │ </note>
  └────

and I'd like to export it to reStructuredText as:

  ┌────
  │ .. note::
  │    Foo
  │    
  │    Bar
  └────

that is, the tag's content has to be indented with 2 consecutive spaces.

I guess this can be done with a lua filter but I must admit I have no
idea how to:

- identify such a tag (its beginning and its end),
- express the content has to be indented with 2 spaces.

Any help would be welcome.

Thanks.
-- 
Denis

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/875z5n7rsg.fsf%40example.com.

Re: How to indent a block when exporting to reStructuredText

[John MacFarlane]

Denis Bitouzé <denis.bitouze-PToOLB6qFnLddoiwg0DPyg@public.gmane.org> writes:

> I guess this can be done with a lua filter but I must admit I have no
> idea how to:
>
> - identify such a tag (its beginning and its end),

> - express the content has to be indented with 2 spaces.

This second part is easy. Pandoc will convert

```
Div ("",["note"],[]) [...blocks...]
```

to a `.. note::` directive in rst.  So you just need to get
the blocks between the note tags and put them in a Div with
class "note"; you don't need to worry about rendering raw rst.

As for the first part:

Use `pandoc -t native` to tell you how pandoc is parsing the
dokuwiki:

```
% pandoc -f dokuwiki -t native
<note>

a

b

</note>
^D
[Para [Str "<note>"]
,Para [Str "a"]
,Para [Str "b"]
,Para [Str "</note>"]]
```

So, if the note directives are all going to occur at the
outer level (not nested in a list item or such), then
you just need to iterate through the blocks till you find
this starting Para, and collect blocks til you find the
ending one.  Hope that's clear, but feel free to ask
follow ups.

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/m2o8jfof2r.fsf%40MacBook-Pro.hsd1.ca.comcast.net.

Re: How to indent a block when exporting to reStructuredText

[Denis Bitouzé]

Le 29/11/20 à 16h37, John MacFarlane a écrit :

> Denis Bitouzé <denis.bitouze-PToOLB6qFnLddoiwg0DPyg@public.gmane.org> writes:
>
>> I guess this can be done with a lua filter but I must admit I have no
>> idea how to:
>>
>> - identify such a tag (its beginning and its end),
>
>> - express the content has to be indented with 2 spaces.
>
> This second part is easy. Pandoc will convert
>
> ```
> Div ("",["note"],[]) [...blocks...]
> ```
>
> to a `.. note::` directive in rst.  So you just need to get
> the blocks between the note tags and put them in a Div with
> class "note"; you don't need to worry about rendering raw rst.

I'm not sure to understand but "you don't need to worry" is a good
news! :)

> As for the first part:
>
> Use `pandoc -t native` to tell you how pandoc is parsing the
> dokuwiki:
>
> ```
> % pandoc -f dokuwiki -t native

I wasn't aware: nice!

> <note>
>
> a
>
> b
>
> </note>
> ^D
> [Para [Str "<note>"]
> ,Para [Str "a"]
> ,Para [Str "b"]
> ,Para [Str "</note>"]]
> ```
>
> So, if the note directives are all going to occur at the
> outer level (not nested in a list item or such), then
> you just need to iterate through the blocks till you find
> this starting Para, and collect blocks til you find the
> ending one.  Hope that's clear, but feel free to ask
> follow ups.

I'm really sorry: even if I understand the principle, I've absolutely no
idea how to implement it (and all the examples I read either Pandoc Lua
Filters examples or elsewhere didn't help me). A little help would be
welcome!

Thanks!
-- 
Denis

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/87eekbjl2a.fsf%40example.com.

Re: How to indent a block when exporting to reStructuredText

[Denis Bitouzé]

Le 30/11/20 à 09h38, Denis Bitouzé a écrit :

> Le 29/11/20 à 16h37, John MacFarlane a écrit :
>
>> Denis Bitouzé <denis.bitouze-PToOLB6qFnLddoiwg0DPyg@public.gmane.org> writes:
>>
>>> I guess this can be done with a lua filter but I must admit I have no
>>> idea how to:
>>>
>>> - identify such a tag (its beginning and its end),
>>
>>> - express the content has to be indented with 2 spaces.
>>
>> This second part is easy. Pandoc will convert
>>
>> ```
>> Div ("",["note"],[]) [...blocks...]
>> ```
>>
>> to a `.. note::` directive in rst.  So you just need to get
>> the blocks between the note tags and put them in a Div with
>> class "note"; you don't need to worry about rendering raw rst.
>
> I'm not sure to understand but "you don't need to worry" is a good
> news! :)
>
>> As for the first part:
>>
>> Use `pandoc -t native` to tell you how pandoc is parsing the
>> dokuwiki:
>>
>> ```
>> % pandoc -f dokuwiki -t native
>
> I wasn't aware: nice!
>
>> <note>
>>
>> a
>>
>> b
>>
>> </note>
>> ^D
>> [Para [Str "<note>"]
>> ,Para [Str "a"]
>> ,Para [Str "b"]
>> ,Para [Str "</note>"]]
>> ```
>>
>> So, if the note directives are all going to occur at the
>> outer level (not nested in a list item or such), then
>> you just need to iterate through the blocks till you find
>> this starting Para, and collect blocks til you find the
>> ending one.  Hope that's clear, but feel free to ask
>> follow ups.
>
> I'm really sorry: even if I understand the principle, I've absolutely no
> idea how to implement it (and all the examples I read either Pandoc Lua
> Filters examples or elsewhere didn't help me). A little help would be
> welcome!

I'm sorry to send again this message but I'm afraid it will get lost in
the flow of this mailing list :)

Would somebody kindly give me a starting point? I carefully read the
pandoc filters docs and had a look at many examples out there but I've
absolutely no idea how to:

- find the starting Para,
- collect blocks til I find the ending Para,
- put these collected blocks (when I'll succeed to collect them) in
  a Div with class "note".

That is how I'm lost :$

Thanks again!
-- 
Denis

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/877dq2iq4d.fsf%40example.com.

Re: How to indent a block when exporting to reStructuredText

[John MacFarlane]

Denis Bitouzé <denis.bitouze-PToOLB6qFnLddoiwg0DPyg@public.gmane.org> writes:

> Would somebody kindly give me a starting point? I carefully read the
> pandoc filters docs and had a look at many examples out there but I've
> absolutely no idea how to:
>
> - find the starting Para,
> - collect blocks til I find the ending Para,
> - put these collected blocks (when I'll succeed to collect them) in
>   a Div with class "note".

Something like this (untested and surely buggy, you'll have to debug and fix):

function isNoteStart(block)
  return block.t == 'Para' and block.c == ??? <- fill in
end

function isNoteEnd(block)
  return block.t == 'Para' and block.c == ??? <- fill in
end

function Pandoc(el)

  local inNote = false
  local contents = {}
  local note = {}
  for _,block in el.blocks:
    if isNoteEnd(block) then
      inNote = false
      contents:insert(pandoc.Div(note, pandoc.Attr("",{"note"})))
    elseif isNoteStart(block) then
      inNote = true
      note = {}
    elseif inNote then
      note:insert(block)      
    else
      contents:insert(block)
    end
    el.blocks = contents
    return Pandoc(el)

end 

 

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/m2czztxsnf.fsf%40MacBook-Pro.hsd1.ca.comcast.net.

Re: How to indent a block when exporting to reStructuredText

[Denis Bitouzé]

Le 01/12/20 à 10h55, John MacFarlane a écrit :

> Denis Bitouzé <denis.bitouze-PToOLB6qFnLddoiwg0DPyg@public.gmane.org> writes:
>
>> Would somebody kindly give me a starting point? I carefully read the
>> pandoc filters docs and had a look at many examples out there but I've
>> absolutely no idea how to:
>>
>> - find the starting Para,
>> - collect blocks til I find the ending Para,
>> - put these collected blocks (when I'll succeed to collect them) in
>>   a Div with class "note".
>
> Something like this (untested and surely buggy, you'll have to debug
> and fix):

Many thanks!

> function isNoteStart(block)
>   return block.t == 'Para' and block.c == ??? <- fill in
> end
>
> function isNoteEnd(block)
>   return block.t == 'Para' and block.c == ??? <- fill in
> end
>
> function Pandoc(el)
>
>   local inNote = false
>   local contents = {}
>   local note = {}
>   for _,block in el.blocks:
>     if isNoteEnd(block) then
>       inNote = false
>       contents:insert(pandoc.Div(note, pandoc.Attr("",{"note"})))
>     elseif isNoteStart(block) then
>       inNote = true
>       note = {}
>     elseif inNote then
>       note:insert(block)      
>     else
>       contents:insert(block)
>     end
>     el.blocks = contents
>     return Pandoc(el)
>
> end 

Indeed, it's buggy and, unfortunately, I'm unable to debug it
(impossible to get `moddebug` working on my Linux box: `luasocket`
issues).

AFAICT, there are two points I suppose I succeeded to debug:

1. `for _,block in el.blocks:` should be `for _,block in el.blocks do`
2. there was a missing `end` at the end of the script.

Also, I filled in the `block.c=???` with (not sure about this):

- `block.c == 'Str "<note>"'` for `isNoteStart`
- `block.c == 'Str "</note>"'` for `isNoteEnd`

So now, I have:

- as `test.lua`:

--8<---------------cut here---------------start------------->8---
function isNoteStart(block)
   return block.t == 'Para' and block.c == 'Str "<note>"'
end

function isNoteEnd(block)
   return block.t == 'Para' and block.c == 'Str "</note>"'
end

function Pandoc(el)

   local inNote = false
   local contents = {}
   local note = {}
   for _,block in pairs(el.blocks) do
   if isNoteEnd(block) then
      inNote = false
      contents:insert(pandoc.Div(note, pandoc.Attr("",{"note"})))
   elseif isNoteStart(block) then
      inNote = true
      note = {}
   elseif inNote then
      note:insert(block)
   else
      contents:insert(block)
   end
   el.blocks = contents
   return Pandoc(el)

   end
end
--8<---------------cut here---------------end--------------->8---

- as `test.txt`

--8<---------------cut here---------------start------------->8---
<note>
Foo

Bar
</note>
--8<---------------cut here---------------end--------------->8---

but, when I run:

--8<---------------cut here---------------start------------->8---
pandoc -f dokuwiki test.txt -o test.rst --lua-filter=test.lua
--8<---------------cut here---------------end--------------->8---

I get the error message:

  ┌────
  │ Error running filter test.lua:
  │ test.lua:24: attempt to call a nil value (method 'insert')
  │ stack traceback:
  │         test.lua:24: in function 'Pandoc'
  └────

Any help would be very welcome.

Thanks!
-- 
Denis

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/871rg8fqo3.fsf%40example.com.

Re: How to indent a block when exporting to reStructuredText

[EBkysko]

[-- Attachment #1.1: Type: text/plain, Size: 2054 bytes --]

A few comments, if I may:

- for your present error message, it's because note and contents are just 
'plain' Lua tables, so you should use Lua's `table.insert` to insert items 
in them; if you want to use the notation `contents:insert` and 
`note:insert`, you must initialize them as `List`: `contents = 
pandoc.List()`, `note = pandoc.List()`. See module-pandoc.list 
<https://pandoc.org/lua-filters.html#module-pandoc.list>

- use `ipairs` rather than `pairs` to be sure you get the list of blocks in 
the right order. See Lua documentation 
<https://www.lua.org/manual/5.3/manual.html#6.1>.

- for the block checks, the content of `Para` is a list of `Inlines` (see 
`Para` <https://pandoc.org/lua-filters.html#type-para>), so `block.c` is a 
table, not a string, and the comparison must be done with the first 
`Inline`, or rather what it contains.

- in your test example, `<note>` is immediately followed by `Foo`; notice 
how jm's example was a bit different, with a blank line between the two. 
Check the native output of your example and see it is not a simple `Para` 
with only one `Inline`. If in your documents there is no blank line after 
the `<note>` (and before the `</note>`), you'll have to handle those 
situations a bit more carefully.

- return the constructor `pandoc.Pandoc` in the end; it takes two 
arguments, the `blocks` and the `meta`; if the `meta` is not given, it 
assumes it gets the `blocks` only. See Pandoc constructor 
<https://pandoc.org/lua-filters.html#pandoc>.

- move you `el.blocks = contents` and `return pandoc.Pandoc...` out of the 
`for` loop.

I left details open for you to study and complete.

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/0d1533ca-8cf8-4a50-8318-a770eaaf2d31n%40googlegroups.com.

[-- Attachment #1.2: Type: text/html, Size: 2475 bytes --]

Re: How to indent a block when exporting to reStructuredText

[EBkysko]

[-- Attachment #1.1: Type: text/plain, Size: 573 bytes --]

a detail/correction:
for the block check for `</note>`, comparison would be for the last 
`Inline` (which is same as first only if `Para` has only one `Inline`).

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/060f6435-b931-4361-969e-82708b81036cn%40googlegroups.com.

[-- Attachment #1.2: Type: text/html, Size: 888 bytes --]

Re: How to indent a block when exporting to reStructuredText

[Denis Bitouzé]

Le 02/12/20 à 07h16, EBkysko a écrit :

> A few comments, if I may:

Of course, many thanks for your advices!

> - for your present error message, it's because note and contents are just 
> 'plain' Lua tables, so you should use Lua's `table.insert` to insert items 
> in them; if you want to use the notation `contents:insert` and 
> `note:insert`, you must initialize them as `List`: `contents = 
> pandoc.List()`, `note = pandoc.List()`. See module-pandoc.list 
> <https://pandoc.org/lua-filters.html#module-pandoc.list>

Seems to fail (see at the end).

> - use `ipairs` rather than `pairs` to be sure you get the list of blocks in 
> the right order. See Lua documentation 
> <https://www.lua.org/manual/5.3/manual.html#6.1>.

OK.

> - for the block checks, the content of `Para` is a list of `Inlines` (see 
> `Para` <https://pandoc.org/lua-filters.html#type-para>), so `block.c` is a 
> table, not a string, and the comparison must be done with the first 
> `Inline`, or rather what it contains.

If I understand well, it should then be:

  ┌────
  │ block.c[1] == 'Str "<note>"'
  └────

and

  ┌────
  │ block.c[1] == 'Str "</note>"'
  └────

> - in your test example, `<note>` is immediately followed by `Foo`; notice 
> how jm's example was a bit different, with a blank line between the two. 
> Check the native output of your example and see it is not a simple `Para` 
> with only one `Inline`. If in your documents there is no blank line after 
> the `<note>` (and before the `</note>`),

Both cases could occur.

> you'll have to handle those situations a bit more carefully.

Unfortunately, I don't see what to do (though I noticed no blank line
gives `SoftBreak` in the native output gives).

> - return the constructor `pandoc.Pandoc` in the end; it takes two 
> arguments, the `blocks` and the `meta`; if the `meta` is not given, it 
> assumes it gets the `blocks` only. See Pandoc constructor 
> <https://pandoc.org/lua-filters.html#pandoc>.

Not sure to understand :$

> - move you `el.blocks = contents` and `return pandoc.Pandoc...` out of the 
> `for` loop.

OK, done.

> I left details open for you to study and complete.

Thanks :)

> for the block check for `</note>`, comparison would be for the last
> `Inline` (which is same as first only if `Para` has only one
> `Inline`).

I guess it is the same since `</note>` is always alone on a line.

--------------------------------------------------

Here is my home work ;)

If I understood well, the following filter should take  into account
your advices:

--8<---------------cut here---------------start------------->8---
contents = pandoc.List()
note = pandoc.List()

function isNoteStart(block)
   return block.t == 'Para' and block.c[1] == 'Str "<note>"'
end

function isNoteEnd(block)
   return block.t == 'Para' and block.c[1] == 'Str "</note>"'
end

function Pandoc(el)

   local inNote = false
   local contents = {}
   local note = {}
   for _,block in ipairs(el.blocks) do
      if isNoteEnd(block) then
         inNote = false
         contents:insert(pandoc.Div(note, pandoc.Attr("",{"note"})))
      elseif isNoteStart(block) then
         inNote = true
         note = {}
      elseif inNote then
         note:insert(block)
      else
         contents:insert(block)
      end
   end
   el.blocks = contents
   return pandoc.Pandoc(el)
end
--8<---------------cut here---------------end--------------->8---

But, unfortunately, it is not perfect since:

--8<---------------cut here---------------start------------->8---
pandoc -f dokuwiki test.txt -o test.rst --lua-filter=test.lua 
--8<---------------cut here---------------end--------------->8---

still returns the same error message:

  ┌────
  │ Error running filter test.lua:
  │ test.lua:27: attempt to call a nil value (method 'insert')
  │ stack traceback:
  │         test.lua:27: in function 'Pandoc'
  └────

Many thanks again!
-- 
Denis

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/87360oduiq.fsf%40example.com.

Re: How to indent a block when exporting to reStructuredText

[John MacFarlane]

Denis Bitouzé <denis.bitouze-PToOLB6qFnLddoiwg0DPyg@public.gmane.org> writes:


> If I understand well, it should then be:
>
>   ┌────
>   │ block.c[1] == 'Str "<note>"'
>   └────
>
> and
>
>   ┌────
>   │ block.c[1] == 'Str "</note>"'
>   └────

Certainly neither.  This would work if block.[1] were a string,
but it's not.  What you want is probably

    block.c[1].t == 'Str' and block.c[1].text == "</note>"

.

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/m2mtywoz0f.fsf%40MacBook-Pro.hsd1.ca.comcast.net.

Re: How to indent a block when exporting to reStructuredText

[Denis Bitouzé]

Le 02/12/20 à 10h15, John MacFarlane a écrit :

> Denis Bitouzé <denis.bitouze-PToOLB6qFnLddoiwg0DPyg@public.gmane.org> writes:
>
>
>> If I understand well, it should then be:
>>
>>   ┌────
>>   │ block.c[1] == 'Str "<note>"'
>>   └────
>>
>> and
>>
>>   ┌────
>>   │ block.c[1] == 'Str "</note>"'
>>   └────
>
> Certainly neither.  This would work if block.[1] were a string,
> but it's not.  What you want is probably
>
>     block.c[1].t == 'Str' and block.c[1].text == "</note>"

Ah, OK, I understand a little better: thanks!
-- 
Denis

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/87o8jbckqi.fsf%40example.com.

Re: How to indent a block when exporting to reStructuredText

[EBkysko]

[-- Attachment #1.1: Type: text/plain, Size: 2336 bytes --]

> still returns the same error message

because you left the `= {}` later in the code.
for now, perhaps it might be better for you to use Lua's usual 
`table.insert(note,...)` notation etc. rather than using `List` and 
`:insert` notation... one step at a time...

> If I understand well, it should then be...

A good beginning. Actually, `block.c[1]` will be the first `Inline`, which 
is a `pandoc` element... It's not literally a string, it is an element of 
type `Str`... see the nuance?
See jm's answer.

As for `</note>`, as I said in my correction, it would be the last item, 
i.e. `block.c[#block.c]`, which is more adequate if `Para` has more than 
one `Inline` in its list.

> Unfortunately, I don't see what to do

If you look at the code, it discards the `block` when '<note>' is found for 
example. So you'd lose all that paragraph.

For a more complex situation, like when the native output gives `Para [Str 
"<note>", SoftBreak, Str "Foo"]`, you need to discard '<note>' and 
'SoftBreak' from the list, but keep 'Foo'.

In general, you would *remove* the first item from the table `block.c`; 
after this, you can check if the *new* first item is really 'SoftBreak', 
and *remove* it also if true.
You would then be left with a table containing only the rest of the para 
items, i.e. the `Para` without the '<note>' (and possibly without the 
'SoftBreak').
This modified `Para` block should then be inserted in the new `note` table. 
Do similarly in the end for '</note>', but removing the last item(s) of the 
`Para` list of `Inline`s.

I'll let you translate this into code!

> Not sure to understand (about pandoc constructor)

In the end, you want to create a new pandoc document with the new modified 
content, and for this you use `pandoc.Pandoc` constructor/creator.
You will have filled the `content` table with a list of pandoc `Block` 
elements, which is what `pandoc.Pandoc` takes as first argument.

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/c210b412-3d47-45d5-a7e7-c55b1086d587n%40googlegroups.com.

[-- Attachment #1.2: Type: text/html, Size: 2802 bytes --]

Re: How to indent a block when exporting to reStructuredText

[Denis Bitouzé]

Le 02/12/20 à 11h15, EBkysko a écrit :

>> still returns the same error message
>
> because you left the `= {}` later in the code.
> for now, perhaps it might be better for you to use Lua's usual
> `table.insert(note,...)` notation etc. rather than using `List` and
> `:insert` notation... one step at a time...

OK. I still locally initialize `note` and `contents` with
`pandoc.List()`.

>> If I understand well, it should then be...
>
> A good beginning. Actually, `block.c[1]` will be the first `Inline`, which
> is a `pandoc` element... It's not literally a string, it is an element of
> type `Str`... see the nuance?
> See jm's answer.

Seen, and useful :)

> As for `</note>`, as I said in my correction, it would be the last item,
> i.e. `block.c[#block.c]`, which is more adequate if `Para` has more than
> one `Inline` in its list.

OK, done.

>> Unfortunately, I don't see what to do
>
> If you look at the code, it discards the `block` when '<note>' is found for
> example.

Because of the line `note = {}` (that I replaced by `note = pandoc.List()`)

> So you'd lose all that paragraph.

Indeed, the resulting `test.rst` is empty, even with the line `note =
pandoc.List()` is commented :$

> For a more complex situation, like when the native output gives `Para [Str
> "<note>", SoftBreak, Str "Foo"]`, you need to discard '<note>' and
> 'SoftBreak' from the list, but keep 'Foo'.

OK.

> In general, you would *remove* the first item from the table `block.c`;
> after this, you can check if the *new* first item is really 'SoftBreak',
> and *remove* it also if true.

OK, I tried to generalize this in case there are other `SoftBreak`.

> You would then be left with a table containing only the rest of the para
> items, i.e. the `Para` without the '<note>' (and possibly without the
> 'SoftBreak').
> This modified `Para` block should then be inserted in the new `note` table.
> Do similarly in the end for '</note>', but removing the last item(s) of the
> `Para` list of `Inline`s.
>
> I'll let you translate this into code!

Here is my attempt to do that but, unfortunately, it still gives an
empty resulting `test.rst` file:

--8<---------------cut here---------------start------------->8---
function isNoteStart(block)
   return block.t == 'Para' and block.c[1].t == 'Str' and block.c[1].text == "<note>"
end

function isNoteEnd(block)
   return block.t == 'Para' and block.c[#block.c].t == 'Str' and block.c[#block.c].text == "</note>"
end

function Pandoc(el)

   local inNote = false
   local contents = pandoc.List()
   local note = pandoc.List()
   for _,block in ipairs(el.blocks) do
      if isNoteEnd(block) then
         table.remove(block.c, #block.c)
         inNote = false
         -- contents:insert(pandoc.Div(note, pandoc.Attr("",{"note"})))
         table.insert(contents,pandoc.Div(note, pandoc.Attr("",{"note"})))
      elseif isNoteStart(block) then
         inNote = true
         -- note = pandoc.List()
         table.remove(block.c, 1)
         for i=1,#block.c-1 do
            if block.c[i].t == 'SoftBreak' then
               table.remove(block.c, 1)
            end
         end
      elseif inNote then
         -- note:insert(block)
         table.insert(note,block)
      else
         -- contents:insert(block)
         table.insert(contents,block)
      end
   end
   el.blocks = contents
   return pandoc.Pandoc(el)
end
--8<---------------cut here---------------end--------------->8---

>> Not sure to understand (about pandoc constructor)
>
> In the end, you want to create a new pandoc document with the new modified
> content, and for this you use `pandoc.Pandoc` constructor/creator.
> You will have filled the `content` table with a list of pandoc `Block`
> elements, which is what `pandoc.Pandoc` takes as first argument.

I understand better now :)

Many thanks again!
-- 
Denis

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/87k0tzci6s.fsf%40example.com.

Re: How to indent a block when exporting to reStructuredText

[EBkysko]

[-- Attachment #1.1: Type: text/plain, Size: 1659 bytes --]

Almost there!

- most importantly, after removing the 'note' tags from the block, you 
forget to put the modified block in the `note` table, since it contains 
what's left of the paragraphs once the tag is removed;

- also important, in `elseif isNoteStart` chunk, you should really 
uncomment the creation of a new `note` table, this is where the collection 
will begin for the current note;

- in the loop to remove the SoftBreak's, I can guess you wanted to remove 
any SoftBreak at the start, but the loop really isn't doing that; I'll let 
you rethink about it.
  In fact, for now let's forget about the SoftBreak's and concentrate on 
obtaining a working `Div`; the SoftBreak's won't do any arm;

- finally, as I said the pandoc.Pandoc constructor takes the new list of 
blocks as first (or only) argument, but you put back only the 'el', which 
is a table containing el.blocks and el.meta. So you either return 
`pandoc.Pandoc(contents)`, or `pandoc.Pandoc(contents, el.meta)`  (if you 
put `el.blocks=contents`, you can use `el.blocks` rather than `contents`)

- minor comment: `table.remove(block.c, #block.c)` works, but there's a 
simpler way to write it, I'll let you think about it.

I think with these you should have a working code.

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/db1f8570-4afc-45ee-ac1f-0ae4435cf833n%40googlegroups.com.

[-- Attachment #1.2: Type: text/html, Size: 2021 bytes --]

Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[Denis Bitouzé]

Le 03/12/20 à 07h49, EBkysko a écrit :

> Almost there!

Finally! :)

> - most importantly, after removing the 'note' tags from the block, you 
> forget to put the modified block in the `note` table, since it contains 
> what's left of the paragraphs once the tag is removed;

Done.

> - also important, in `elseif isNoteStart` chunk, you should really 
> uncomment the creation of a new `note` table, this is where the collection 
> will begin for the current note;

Done.

> - in the loop to remove the SoftBreak's, I can guess you wanted to remove 
> any SoftBreak at the start, but the loop really isn't doing that;

I guess it should have been:

--8<---------------cut here---------------start------------->8---
    if block.c[i].t == 'SoftBreak' then
      table.remove(block.c[i], 1)
--8<---------------cut here---------------end--------------->8---

instead of

--8<---------------cut here---------------start------------->8---
    if block.c[i].t == 'SoftBreak' then
      table.remove(block.c, 1)
--8<---------------cut here---------------end--------------->8---

> I'll let you rethink about it.

You're a true hard but fair teacher! ;)

>   In fact, for now let's forget about the SoftBreak's and concentrate on 
> obtaining a working `Div`; the SoftBreak's won't do any arm;

OK.

> - finally, as I said the pandoc.Pandoc constructor takes the new list of 
> blocks as first (or only) argument, but you put back only the 'el', which 
> is a table containing el.blocks and el.meta. So you either return 
> `pandoc.Pandoc(contents)`, or `pandoc.Pandoc(contents, el.meta)`  (if you 
> put `el.blocks=contents`, you can use `el.blocks` rather than `contents`)

OK.

> - minor comment: `table.remove(block.c, #block.c)` works, but there's a 
> simpler way to write it, I'll let you think about it.

You're a true hard but fair teacher! ;)

I hoped Lua worked as Python and that:

  ┌────
  │ table.remove(block.c, -1)
  └────

would work, but that's not the case. So, sorry, I don't see, even after
a look at:

  ┌────
  │ https://www.lua.org/manual/5.3/manual.html#6.6
  └────

> I think with these you should have a working code.

Almost! :)

With the following `test.lua` file:

--8<---------------cut here---------------start------------->8---
function isNoteStart(block)
   return block.t == 'Para' and block.c[1].t == 'Str' and block.c[1].text == "<note>"
end

function isNoteEnd(block)
   return block.t == 'Para' and block.c[#block.c].t == 'Str' and block.c[#block.c].text == "</note>"
end

function Pandoc(el)

   local inNote = false
   local contents = pandoc.List()
   local note = pandoc.List()
   for _,block in ipairs(el.blocks) do
      if isNoteEnd(block) then
         table.remove(block.c, #block.c)
         inNote = false
         -- contents:insert(pandoc.Div(note, pandoc.Attr("",{"note"})))
         table.insert(contents,pandoc.Div(note, pandoc.Attr("",{"note"})))
      elseif isNoteStart(block) then
         inNote = true
         note = pandoc.List()
         table.remove(block.c, 1)
         table.insert(note,block)
         -- for i=1,#block.c-1 do
         --    if block.c[i].t == 'SoftBreak' then
         --       table.remove(block.c[i], 1)
         --    end
         -- end
      elseif inNote then
         -- note:insert(block)
         table.insert(note,block)
      else
         -- contents:insert(block)
         table.insert(contents,block)
      end
   end
   return pandoc.Pandoc(contents, el.meta)
end
--8<---------------cut here---------------end--------------->8---

the following `test.txt` file:

--8<---------------cut here---------------start------------->8---
<note>
Foo bar baz 1
Foo bar baz 2

Foo bar baz 3

<code latex>
Foo bar baz 4
Foo bar baz 5
</code>
</note>

<note>
Foo bar baz 6
Foo bar baz 7

Foo bar baz 8

Foo bar baz 9
Foo bar baz 10
</note>
--8<---------------cut here---------------end--------------->8---

and the command:

--8<---------------cut here---------------start------------->8---
pandoc -f dokuwiki test.txt -o test.rst --lua-filter=test.lua
--8<---------------cut here---------------end--------------->8---

the resulting `test.rst` file is missing the last two lines of the
second note:

--8<---------------cut here---------------start------------->8---
.. note::

   Foo bar baz 1 Foo bar baz 2

   Foo bar baz 3

   .. code:: latex

      Foo bar baz 4
      Foo bar baz 5

.. note::

   Foo bar baz 6 Foo bar baz 7

   Foo bar baz 8
--8<---------------cut here---------------end--------------->8---

I first thought the culprit was:

--8<---------------cut here---------------start------------->8---
table.remove(block.c, #block.c)
--8<---------------cut here---------------end--------------->8---

but AFAICS, it hasn't any effect.

Thanks again!
-- 
Denis

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/874kl2dca6.fsf%40example.com.

Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[BPJ]

[-- Attachment #1: Type: text/plain, Size: 7232 bytes --]

If I may show a correct way to check what kind of element you got:

``````lua
local function is_first_str (el, wanted)
  if 'Para' == el.tag then
    if #el.content > 0 then
      if 'Str' == el.content[1].tag then
        if wanted == el.content[1].text then
          return true
        end
      end
    end
  end
  return false
end
``````

armed with this you can do something like this:

``````lua
function Blocks (old)
  local new = {}
  local i = 1
  while i <= #old do
    if is_first_str(old[i], '<note>') then
      note = {}
      i = i+1
      while not is_first_str(old[i], '</note>') do
        note[#note+1] = old[i]
        i = i + 1
      end
      new[#new+1] = pandoc. Div(note, {class = 'note'})
    else
      new[#new+1] = old[i]
      i = i + 1
    end
  end
  return new
end
``````

Put these two functions in a file and you have a filter which should do
what you
want.

HTH,

/bpj


-- 
Better --help|less than helpless

Den tors 3 dec. 2020 18:35Denis Bitouzé <denis.bitouze@univ-littoral.fr>
skrev:

> Le 03/12/20 à 07h49, EBkysko a écrit :
>
> > Almost there!
>
> Finally! :)
>
> > - most importantly, after removing the 'note' tags from the block, you
> > forget to put the modified block in the `note` table, since it contains
> > what's left of the paragraphs once the tag is removed;
>
> Done.
>
> > - also important, in `elseif isNoteStart` chunk, you should really
> > uncomment the creation of a new `note` table, this is where the
> collection
> > will begin for the current note;
>
> Done.
>
> > - in the loop to remove the SoftBreak's, I can guess you wanted to
> remove
> > any SoftBreak at the start, but the loop really isn't doing that;
>
> I guess it should have been:
>
> --8<---------------cut here---------------start------------->8---
>     if block.c[i].t == 'SoftBreak' then
>       table.remove(block.c[i], 1)
> --8<---------------cut here---------------end--------------->8---
>
> instead of
>
> --8<---------------cut here---------------start------------->8---
>     if block.c[i].t == 'SoftBreak' then
>       table.remove(block.c, 1)
> --8<---------------cut here---------------end--------------->8---
>
> > I'll let you rethink about it.
>
> You're a true hard but fair teacher! ;)
>
> >   In fact, for now let's forget about the SoftBreak's and concentrate on
> > obtaining a working `Div`; the SoftBreak's won't do any arm;
>
> OK.
>
> > - finally, as I said the pandoc.Pandoc constructor takes the new list of
> > blocks as first (or only) argument, but you put back only the 'el',
> which
> > is a table containing el.blocks and el.meta. So you either return
> > `pandoc.Pandoc(contents)`, or `pandoc.Pandoc(contents, el.meta)`  (if
> you
> > put `el.blocks=contents`, you can use `el.blocks` rather than `contents`)
>
> OK.
>
> > - minor comment: `table.remove(block.c, #block.c)` works, but there's a
> > simpler way to write it, I'll let you think about it.
>
> You're a true hard but fair teacher! ;)
>
> I hoped Lua worked as Python and that:
>
>   ┌────
>   │ table.remove(block.c, -1)
>   └────
>
> would work, but that's not the case. So, sorry, I don't see, even after
> a look at:
>
>   ┌────
>   │ https://www.lua.org/manual/5.3/manual.html#6.6
>   └────
>
> > I think with these you should have a working code.
>
> Almost! :)
>
> With the following `test.lua` file:
>
> --8<---------------cut here---------------start------------->8---
> function isNoteStart(block)
>    return block.t == 'Para' and block.c[1].t == 'Str' and block.c[1].text
> == "<note>"
> end
>
> function isNoteEnd(block)
>    return block.t == 'Para' and block.c[#block.c].t == 'Str' and
> block.c[#block.c].text == "</note>"
> end
>
> function Pandoc(el)
>
>    local inNote = false
>    local contents = pandoc.List()
>    local note = pandoc.List()
>    for _,block in ipairs(el.blocks) do
>       if isNoteEnd(block) then
>          table.remove(block.c, #block.c)
>          inNote = false
>          -- contents:insert(pandoc.Div(note, pandoc.Attr("",{"note"})))
>          table.insert(contents,pandoc.Div(note, pandoc.Attr("",{"note"})))
>       elseif isNoteStart(block) then
>          inNote = true
>          note = pandoc.List()
>          table.remove(block.c, 1)
>          table.insert(note,block)
>          -- for i=1,#block.c-1 do
>          --    if block.c[i].t == 'SoftBreak' then
>          --       table.remove(block.c[i], 1)
>          --    end
>          -- end
>       elseif inNote then
>          -- note:insert(block)
>          table.insert(note,block)
>       else
>          -- contents:insert(block)
>          table.insert(contents,block)
>       end
>    end
>    return pandoc.Pandoc(contents, el.meta)
> end
> --8<---------------cut here---------------end--------------->8---
>
> the following `test.txt` file:
>
> --8<---------------cut here---------------start------------->8---
> <note>
> Foo bar baz 1
> Foo bar baz 2
>
> Foo bar baz 3
>
> <code latex>
> Foo bar baz 4
> Foo bar baz 5
> </code>
> </note>
>
> <note>
> Foo bar baz 6
> Foo bar baz 7
>
> Foo bar baz 8
>
> Foo bar baz 9
> Foo bar baz 10
> </note>
> --8<---------------cut here---------------end--------------->8---
>
> and the command:
>
> --8<---------------cut here---------------start------------->8---
> pandoc -f dokuwiki test.txt -o test.rst --lua-filter=test.lua
> --8<---------------cut here---------------end--------------->8---
>
> the resulting `test.rst` file is missing the last two lines of the
> second note:
>
> --8<---------------cut here---------------start------------->8---
> .. note::
>
>    Foo bar baz 1 Foo bar baz 2
>
>    Foo bar baz 3
>
>    .. code:: latex
>
>       Foo bar baz 4
>       Foo bar baz 5
>
> .. note::
>
>    Foo bar baz 6 Foo bar baz 7
>
>    Foo bar baz 8
> --8<---------------cut here---------------end--------------->8---
>
> I first thought the culprit was:
>
> --8<---------------cut here---------------start------------->8---
> table.remove(block.c, #block.c)
> --8<---------------cut here---------------end--------------->8---
>
> but AFAICS, it hasn't any effect.
>
> Thanks again!
> --
> Denis
>
> --
> You received this message because you are subscribed to the Google Groups
> "pandoc-discuss" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/pandoc-discuss/874kl2dca6.fsf%40example.com
> .
>

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/CADAJKhANzMvwhURnVUh49xZzRQFKUnpyCyeZtp%2BnQ7Nc82ZmVQ%40mail.gmail.com.

[-- Attachment #2: Type: text/html, Size: 10267 bytes --]

Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[EBkysko]

[-- Attachment #1.1: Type: text/plain, Size: 1035 bytes --]

@BP

Using `Blocks` would indeed be a good way for nested notes, or notes inside 
other blocks, a more general solution.

A few notes :

- After the inner `while` is done, the index i still caries its value to 
the outer `while`, and can produce an extra unwanted `</note>` paragraph. 
Should increase `i` again before the `else`.

- This will miss an example like:

 ```
<note>
Foo

Bar
</note>
 ```
where there's no blank line(s) after/before the tags; you must remove the 
tags and insert what's left in the `note` table.

- Because of the above, you should rather have an `is_last_str` function 
for the inner `while`.


-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/7b398756-0b85-490b-b05e-8fd4fcaaecd5n%40googlegroups.com.

[-- Attachment #1.2: Type: text/html, Size: 1590 bytes --]

Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[EBkysko]

[-- Attachment #1.1: Type: text/plain, Size: 657 bytes --]

Correcting myself: no, *not* for nested notes.

On Thursday, December 3, 2020 at 4:03:09 PM UTC-5 EBkysko wrote:

> @BP
>
> Using `Blocks` would indeed be a good way for nested notes, or notes 
> inside other blocks, a more general solution.
>
>

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/6221eae9-b70a-4f53-b512-7655c2ac6858n%40googlegroups.com.

[-- Attachment #1.2: Type: text/html, Size: 1203 bytes --]

Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[BPJ]

[-- Attachment #1: Type: text/plain, Size: 3042 bytes --]

It works for nested notes if you have the loop function call itself when it
sees
`<note>`, passing the next index number and an inside-note state flag as
(extra)
second and third arguments, and returning the list of collected blocks and
the
next index number when it sees `</note>`.
The below (untested) code should be able to handle nested notes to any
depth.
Depending on whether `<note>` and `</note>` are always the only thing in a
Para and/or Str you may want to do `table.remove(el.content[i], pos)` or
`el.content[i][pos].text:gsub('%<%/?note%>',"")`, remove anchors on
the patterns in the `is_string_match` calls and push the matching
Para onto `new`.

``````lua
local function is_str_match (el, pat, pos)
  if 'Para' == el.tag then
    if #el.content > 0 then
      if 'Str' == el.content[1].tag then
        pos = pos or #el.content
        if el.content[pos].text:match(pat) then
          return true
        end
      end
    end
  end
  return false
end

local function collect_notes (old, i, inside_note)
  local new = {}
  local i = i or 1
  while i < #old do
    if is_str_match(old[i], '^%<%/note%>$') then
      assert inside_note, "Saw unbalanced end-of-note"
      return new, i+1
    elseif is_str_match(old[i], '^%<note%>$', 1) then
      i = i+1
      local note
      note, i = collect_notes(old, i+1, true)
      if #note > 0 then
        new[#new+1] = pandoc. Div(note, {class = 'note'})
      end
    else
      new[#new+1] = old[i]
      i = i + 1
    end
  end
  assert not(inside_note), "Saw unbalanced start-of-note"
  return new
end

function Blocks (old)
  local new = collect_notes(old, 1, false)
  return new
end
``````

-- 
Better --help|less than helpless

Den tors 3 dec. 2020 23:00EBkysko <ebkysko-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> skrev:

> Correcting myself: no, *not* for nested notes.
>
> On Thursday, December 3, 2020 at 4:03:09 PM UTC-5 EBkysko wrote:
>
>> @BP
>>
>> Using `Blocks` would indeed be a good way for nested notes, or notes
>> inside other blocks, a more general solution.
>>
>> --
> You received this message because you are subscribed to the Google Groups
> "pandoc-discuss" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/pandoc-discuss/6221eae9-b70a-4f53-b512-7655c2ac6858n%40googlegroups.com
> <https://groups.google.com/d/msgid/pandoc-discuss/6221eae9-b70a-4f53-b512-7655c2ac6858n%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/CADAJKhDptqz5OCCDNxXL6p6qDBk0kKkpr-y%2B8%2BikZmspxC0ULA%40mail.gmail.com.

[-- Attachment #2: Type: text/html, Size: 5508 bytes --]

Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[BPJ]

[-- Attachment #1: Type: text/plain, Size: 3420 bytes --]

Oops! Remove the `i = i+1` just after the `elseif`, since `i` is
incremented in the recursive call just after!


-- 
Better --help|less than helpless

Den fre 4 dec. 2020 10:28BPJ <melroch-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> skrev:

> It works for nested notes if you have the loop function call itself when
> it sees
> `<note>`, passing the next index number and an inside-note state flag as
> (extra)
> second and third arguments, and returning the list of collected blocks and
> the
> next index number when it sees `</note>`.
> The below (untested) code should be able to handle nested notes to any
> depth.
> Depending on whether `<note>` and `</note>` are always the only thing in a
> Para and/or Str you may want to do `table.remove(el.content[i], pos)` or
> `el.content[i][pos].text:gsub('%<%/?note%>',"")`, remove anchors on
> the patterns in the `is_string_match` calls and push the matching
> Para onto `new`.
>
> ``````lua
> local function is_str_match (el, pat, pos)
>   if 'Para' == el.tag then
>     if #el.content > 0 then
>       if 'Str' == el.content[1].tag then
>         pos = pos or #el.content
>         if el.content[pos].text:match(pat) then
>           return true
>         end
>       end
>     end
>   end
>   return false
> end
>
> local function collect_notes (old, i, inside_note)
>   local new = {}
>   local i = i or 1
>   while i < #old do
>     if is_str_match(old[i], '^%<%/note%>$') then
>       assert inside_note, "Saw unbalanced end-of-note"
>       return new, i+1
>     elseif is_str_match(old[i], '^%<note%>$', 1) then
>       i = i+1
>       local note
>       note, i = collect_notes(old, i+1, true)
>       if #note > 0 then
>         new[#new+1] = pandoc. Div(note, {class = 'note'})
>       end
>     else
>       new[#new+1] = old[i]
>       i = i + 1
>     end
>   end
>   assert not(inside_note), "Saw unbalanced start-of-note"
>   return new
> end
>
> function Blocks (old)
>   local new = collect_notes(old, 1, false)
>   return new
> end
> ``````
>
> --
> Better --help|less than helpless
>
> Den tors 3 dec. 2020 23:00EBkysko <ebkysko-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> skrev:
>
>> Correcting myself: no, *not* for nested notes.
>>
>> On Thursday, December 3, 2020 at 4:03:09 PM UTC-5 EBkysko wrote:
>>
>>> @BP
>>>
>>> Using `Blocks` would indeed be a good way for nested notes, or notes
>>> inside other blocks, a more general solution.
>>>
>>> --
>> You received this message because you are subscribed to the Google Groups
>> "pandoc-discuss" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/pandoc-discuss/6221eae9-b70a-4f53-b512-7655c2ac6858n%40googlegroups.com
>> <https://groups.google.com/d/msgid/pandoc-discuss/6221eae9-b70a-4f53-b512-7655c2ac6858n%40googlegroups.com?utm_medium=email&utm_source=footer>
>> .
>>
>

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/CADAJKhAHahcvktOyqyY%2BEB87DJSqo3aaVT7eeBtpDvCcfka5Vg%40mail.gmail.com.

[-- Attachment #2: Type: text/html, Size: 6202 bytes --]

Re: [SOCIAL]Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[Denis Bitouzé]

Le 04/12/20 à 10h37, BPJ a écrit :

> Oops! Remove the `i = i+1` just after the `elseif`, since `i` is
> incremented in the recursive call just after!

Many thanks for your help!

I have taken into account the corrections given by @EBkysko (otherwise
I had incomprehensible compile errors), but:

--8<---------------cut here---------------start------------->8---
pandoc -f dokuwiki test.txt -o test.rst --lua-filter=test.lua
--8<---------------cut here---------------end--------------->8---

returns:

  ┌────
  │ Error running filter test.lua:
  │ test.lua:20: syntax error near 'inside_note
  └────

and the culprit is:

  ┌────
  │ assert inside_note, "Saw unbalanced end-of-note"
  └────

I tested with:

  ┌────
  │ assert(inside_note, "Saw unbalanced end-of-note")
  └────

no compile error now :) but, with the following input file `test.txt`:

--8<---------------cut here---------------start------------->8---
<note>
Foo bar baz 1
Foo bar baz 2

Foo bar baz 3

<code latex>
Foo bar baz 4
Foo bar baz 5
</code>
</note>

<note>
Foo bar baz 6
Foo bar baz 7

Foo bar baz 8

Foo bar baz 9
Foo bar baz 10
</note>
--8<---------------cut here---------------end--------------->8---

the resulting `test.rst` is missing some lines (the first two of the
first note, and the last two of the last note):

--8<---------------cut here---------------start------------->8---
.. note::

   Foo bar baz 3

   .. code:: latex

      Foo bar baz 4
      Foo bar baz 5

.. note::

   Foo bar baz 8
--8<---------------cut here---------------end--------------->8---

For the moment, I have to look in detail your code in order to extend it
to the new cases I didn't consider first (as already told to @EBkysko):
- the filter didn't work when there weren't blank lines between `<note>`
and`</note>` since, in such a case, these strings were both in the same
`Para`,
- in my real documents, there were other "notes":
  - `<note tip>`
  - `<note warning>`
  - `<note important>`.

Many thanks again!
-- 
Denis

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/878sadxpkk.fsf%40example.com.

Re: [SOCIAL]Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[BPJ]

[-- Attachment #1: Type: text/plain, Size: 3430 bytes --]

It turned out to be harder than I thought but I'm working on it. I may not
have time tomorrow but I'll probably fix it by Monday.

-- 
Better --help|less than helpless

Den fre 4 dec. 2020 21:52Denis Bitouzé <denis.bitouze-PToOLB6qFnLddoiwg0DPyg@public.gmane.org>
skrev:

> Le 04/12/20 à 10h37, BPJ a écrit :
>
> > Oops! Remove the `i = i+1` just after the `elseif`, since `i` is
> > incremented in the recursive call just after!
>
> Many thanks for your help!
>
> I have taken into account the corrections given by @EBkysko (otherwise
> I had incomprehensible compile errors), but:
>
> --8<---------------cut here---------------start------------->8---
> pandoc -f dokuwiki test.txt -o test.rst --lua-filter=test.lua
> --8<---------------cut here---------------end--------------->8---
>
> returns:
>
>   ┌────
>   │ Error running filter test.lua:
>   │ test.lua:20: syntax error near 'inside_note
>   └────
>
> and the culprit is:
>
>   ┌────
>   │ assert inside_note, "Saw unbalanced end-of-note"
>   └────
>
> I tested with:
>
>   ┌────
>   │ assert(inside_note, "Saw unbalanced end-of-note")
>   └────
>
> no compile error now :) but, with the following input file `test.txt`:
>
> --8<---------------cut here---------------start------------->8---
> <note>
> Foo bar baz 1
> Foo bar baz 2
>
> Foo bar baz 3
>
> <code latex>
> Foo bar baz 4
> Foo bar baz 5
> </code>
> </note>
>
> <note>
> Foo bar baz 6
> Foo bar baz 7
>
> Foo bar baz 8
>
> Foo bar baz 9
> Foo bar baz 10
> </note>
> --8<---------------cut here---------------end--------------->8---
>
> the resulting `test.rst` is missing some lines (the first two of the
> first note, and the last two of the last note):
>
> --8<---------------cut here---------------start------------->8---
> .. note::
>
>    Foo bar baz 3
>
>    .. code:: latex
>
>       Foo bar baz 4
>       Foo bar baz 5
>
> .. note::
>
>    Foo bar baz 8
> --8<---------------cut here---------------end--------------->8---
>
> For the moment, I have to look in detail your code in order to extend it
> to the new cases I didn't consider first (as already told to @EBkysko):
> - the filter didn't work when there weren't blank lines between `<note>`
> and`</note>` since, in such a case, these strings were both in the same
> `Para`,
> - in my real documents, there were other "notes":
>   - `<note tip>`
>   - `<note warning>`
>   - `<note important>`.
>
> Many thanks again!
> --
> Denis
>
> --
> You received this message because you are subscribed to the Google Groups
> "pandoc-discuss" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/pandoc-discuss/878sadxpkk.fsf%40example.com
> .
>

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/CADAJKhCGc5Xd%3DeqchzoJ8ydj%3DvDjumXktN17KiAVWrTppuKCoQ%40mail.gmail.com.

[-- Attachment #2: Type: text/html, Size: 4668 bytes --]

Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[EBkysko]

[-- Attachment #1.1: Type: text/plain, Size: 836 bytes --]

@BP: works with a few corrections:

- in is_str_match: `el.content[1]` should be `el.content[pos]`
- in is_str_match: `pos = pos or #el.content` should be moved before the 
`if 'Str'...`
- in collect_notes: `while i < #old` should be `while i <= #old` ('</note>` 
could be in last para)
- assert args should be within parenthesis


On Friday, December 4, 2020 at 4:29:03 AM UTC-5 BP wrote:

> It works for nested notes....
>

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/ac0cbc5c-8107-456e-8994-dc89619078d2n%40googlegroups.com.

[-- Attachment #1.2: Type: text/html, Size: 1515 bytes --]

Re: [SOCIAL]Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[Denis Bitouzé]

Le 03/12/20 à 13h03, EBkysko a écrit :

> Using `Blocks` would indeed be a good way for nested notes, or notes inside 
> other blocks, a more general solution.

Isn't `Blocks` an arbitrary name of a function?
-- 
Denis

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/87im9iarlq.fsf%40example.com.

Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[BPJ]

[-- Attachment #1: Type: text/plain, Size: 8457 bytes --]

Sorry I forgot to localize some variables. Corrected below.

``````lua
local function is_first_str (el, wanted)
  if 'Para' == el.tag then
    if #el.content > 0 then
      if 'Str' == el.content[1].tag then
        if wanted == el.content[1].text then
          return true
        end
      end
    end
  end
  return false
end

function Blocks (old)
  local new = {}
  local i = 1
  while i < #old do
    if is_first_str(old[i], '<note>') then
      local note = {}
      i = i+1
      while not is_first_str(old[i], '</note>') do
        note[#note+1] = old[i]
        i = i + 1
      end
      new[#new+1] = pandoc. Div(note, {class = 'note'})
    else
      new[#new+1] = old[i]
      i = i + 1
    end
  end
  return new
end
``````


-- 
Better --help|less than helpless

Den tors 3 dec. 2020 21:17BPJ <melroch-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> skrev:

> If I may show a correct way to check what kind of element you got:
>
> ``````lua
> local function is_first_str (el, wanted)
>   if 'Para' == el.tag then
>     if #el.content > 0 then
>       if 'Str' == el.content[1].tag then
>         if wanted == el.content[1].text then
>           return true
>         end
>       end
>     end
>   end
>   return false
> end
> ``````
>
> armed with this you can do something like this:
>
> ``````lua
> function Blocks (old)
>   local new = {}
>   local i = 1
>   while i <= #old do
>     if is_first_str(old[i], '<note>') then
>       note = {}
>       i = i+1
>       while not is_first_str(old[i], '</note>') do
>         note[#note+1] = old[i]
>         i = i + 1
>       end
>       new[#new+1] = pandoc. Div(note, {class = 'note'})
>     else
>       new[#new+1] = old[i]
>       i = i + 1
>     end
>   end
>   return new
> end
> ``````
>
> Put these two functions in a file and you have a filter which should do
> what you
> want.
>
> HTH,
>
> /bpj
>
>
> --
> Better --help|less than helpless
>
> Den tors 3 dec. 2020 18:35Denis Bitouzé <denis.bitouze@univ-littoral.fr>
> skrev:
>
>> Le 03/12/20 à 07h49, EBkysko a écrit :
>>
>> > Almost there!
>>
>> Finally! :)
>>
>> > - most importantly, after removing the 'note' tags from the block, you
>> > forget to put the modified block in the `note` table, since it contains
>> > what's left of the paragraphs once the tag is removed;
>>
>> Done.
>>
>> > - also important, in `elseif isNoteStart` chunk, you should really
>> > uncomment the creation of a new `note` table, this is where the
>> collection
>> > will begin for the current note;
>>
>> Done.
>>
>> > - in the loop to remove the SoftBreak's, I can guess you wanted to
>> remove
>> > any SoftBreak at the start, but the loop really isn't doing that;
>>
>> I guess it should have been:
>>
>> --8<---------------cut here---------------start------------->8---
>>     if block.c[i].t == 'SoftBreak' then
>>       table.remove(block.c[i], 1)
>> --8<---------------cut here---------------end--------------->8---
>>
>> instead of
>>
>> --8<---------------cut here---------------start------------->8---
>>     if block.c[i].t == 'SoftBreak' then
>>       table.remove(block.c, 1)
>> --8<---------------cut here---------------end--------------->8---
>>
>> > I'll let you rethink about it.
>>
>> You're a true hard but fair teacher! ;)
>>
>> >   In fact, for now let's forget about the SoftBreak's and concentrate
>> on
>> > obtaining a working `Div`; the SoftBreak's won't do any arm;
>>
>> OK.
>>
>> > - finally, as I said the pandoc.Pandoc constructor takes the new list
>> of
>> > blocks as first (or only) argument, but you put back only the 'el',
>> which
>> > is a table containing el.blocks and el.meta. So you either return
>> > `pandoc.Pandoc(contents)`, or `pandoc.Pandoc(contents, el.meta)`  (if
>> you
>> > put `el.blocks=contents`, you can use `el.blocks` rather than
>> `contents`)
>>
>> OK.
>>
>> > - minor comment: `table.remove(block.c, #block.c)` works, but there's a
>> > simpler way to write it, I'll let you think about it.
>>
>> You're a true hard but fair teacher! ;)
>>
>> I hoped Lua worked as Python and that:
>>
>>   ┌────
>>   │ table.remove(block.c, -1)
>>   └────
>>
>> would work, but that's not the case. So, sorry, I don't see, even after
>> a look at:
>>
>>   ┌────
>>   │ https://www.lua.org/manual/5.3/manual.html#6.6
>>   └────
>>
>> > I think with these you should have a working code.
>>
>> Almost! :)
>>
>> With the following `test.lua` file:
>>
>> --8<---------------cut here---------------start------------->8---
>> function isNoteStart(block)
>>    return block.t == 'Para' and block.c[1].t == 'Str' and block.c[1].text
>> == "<note>"
>> end
>>
>> function isNoteEnd(block)
>>    return block.t == 'Para' and block.c[#block.c].t == 'Str' and
>> block.c[#block.c].text == "</note>"
>> end
>>
>> function Pandoc(el)
>>
>>    local inNote = false
>>    local contents = pandoc.List()
>>    local note = pandoc.List()
>>    for _,block in ipairs(el.blocks) do
>>       if isNoteEnd(block) then
>>          table.remove(block.c, #block.c)
>>          inNote = false
>>          -- contents:insert(pandoc.Div(note, pandoc.Attr("",{"note"})))
>>          table.insert(contents,pandoc.Div(note, pandoc.Attr("",{"note"})))
>>       elseif isNoteStart(block) then
>>          inNote = true
>>          note = pandoc.List()
>>          table.remove(block.c, 1)
>>          table.insert(note,block)
>>          -- for i=1,#block.c-1 do
>>          --    if block.c[i].t == 'SoftBreak' then
>>          --       table.remove(block.c[i], 1)
>>          --    end
>>          -- end
>>       elseif inNote then
>>          -- note:insert(block)
>>          table.insert(note,block)
>>       else
>>          -- contents:insert(block)
>>          table.insert(contents,block)
>>       end
>>    end
>>    return pandoc.Pandoc(contents, el.meta)
>> end
>> --8<---------------cut here---------------end--------------->8---
>>
>> the following `test.txt` file:
>>
>> --8<---------------cut here---------------start------------->8---
>> <note>
>> Foo bar baz 1
>> Foo bar baz 2
>>
>> Foo bar baz 3
>>
>> <code latex>
>> Foo bar baz 4
>> Foo bar baz 5
>> </code>
>> </note>
>>
>> <note>
>> Foo bar baz 6
>> Foo bar baz 7
>>
>> Foo bar baz 8
>>
>> Foo bar baz 9
>> Foo bar baz 10
>> </note>
>> --8<---------------cut here---------------end--------------->8---
>>
>> and the command:
>>
>> --8<---------------cut here---------------start------------->8---
>> pandoc -f dokuwiki test.txt -o test.rst --lua-filter=test.lua
>> --8<---------------cut here---------------end--------------->8---
>>
>> the resulting `test.rst` file is missing the last two lines of the
>> second note:
>>
>> --8<---------------cut here---------------start------------->8---
>> .. note::
>>
>>    Foo bar baz 1 Foo bar baz 2
>>
>>    Foo bar baz 3
>>
>>    .. code:: latex
>>
>>       Foo bar baz 4
>>       Foo bar baz 5
>>
>> .. note::
>>
>>    Foo bar baz 6 Foo bar baz 7
>>
>>    Foo bar baz 8
>> --8<---------------cut here---------------end--------------->8---
>>
>> I first thought the culprit was:
>>
>> --8<---------------cut here---------------start------------->8---
>> table.remove(block.c, #block.c)
>> --8<---------------cut here---------------end--------------->8---
>>
>> but AFAICS, it hasn't any effect.
>>
>> Thanks again!
>> --
>> Denis
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "pandoc-discuss" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/pandoc-discuss/874kl2dca6.fsf%40example.com
>> .
>>
>

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/CADAJKhAM88Cm%2B-G9oVGM0eJUpPpUUpCkPVQrabNSqKXJOY7r-w%40mail.gmail.com.

[-- Attachment #2: Type: text/html, Size: 12452 bytes --]

Re: [SOCIAL]Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[Denis Bitouzé]

Le 03/12/20 à 22h06, BPJ a écrit :

> Sorry I forgot to localize some variables. Corrected below.

Ma,y thanks for your help!

> ``````lua
> local function is_first_str (el, wanted)
>   if 'Para' == el.tag then
>     if #el.content > 0 then
>       if 'Str' == el.content[1].tag then
>         if wanted == el.content[1].text then
>           return true
>         end
>       end
>     end
>   end
>   return false
> end
>
> function Blocks (old)
>   local new = {}
>   local i = 1
>   while i < #old do
>     if is_first_str(old[i], '<note>') then
>       local note = {}
>       i = i+1
>       while not is_first_str(old[i], '</note>') do
>         note[#note+1] = old[i]
>         i = i + 1
>       end
>       new[#new+1] = pandoc. Div(note, {class = 'note'})

I guess the space in `pandoc. Div` has to be removed, or is it harmless?

>     else
>       new[#new+1] = old[i]
>       i = i + 1
>     end
>   end
>   return new
> end
> ``````

Running:

--8<---------------cut here---------------start------------->8---
pandoc -f dokuwiki test.txt -o test.rst --lua-filter=test-bis.lua
--8<---------------cut here---------------end--------------->8---

with `test-bis.lua` containing your code and the following `test.txt`
file:

--8<---------------cut here---------------start------------->8---
<note>
Foo bar baz 1
Foo bar baz 2

Foo bar baz 3

<code latex>
Foo bar baz 4
Foo bar baz 5
</code>
</note>

<note>
Foo bar baz 6
Foo bar baz 7

Foo bar baz 8

Foo bar baz 9
Foo bar baz 10
</note>
--8<---------------cut here---------------end--------------->8---

gave me the error:

  ┌────
  │ Error running filter test-bis.lua:
  │ test-bis.lua:2: attempt to index a nil value (local 'el')
  │ stack traceback:
  │         test-bis.lua:2: in upvalue 'is_first_str'
  │         test-bis.lua:21: in function 'Blocks
  └────

So I tried to circumvent by replacing:

  ┌────
  │ if 'Para' == el.tag then
  └────

by:

  ┌────
  │ if not el and 'Para' == el.tag then
  └────

No more error but not the expected result:

  ┌────
  │ <note> Foo bar baz 1 Foo bar baz 2
  │ 
  │ Foo bar baz 3
  │ 
  │ .. code:: latex
  │ 
  │    Foo bar baz 4
  │    Foo bar baz 5
  │ 
  │ </note>
  │ 
  │ <note> Foo bar baz 6 Foo bar baz 7
  │ 
  │ Foo bar baz 8
  └────

Thanks again!
-- 
Denis

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/87mtyuarnw.fsf%40example.com.

Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[EBkysko]

[-- Attachment #1.1: Type: text/plain, Size: 1502 bytes --]

> the resulting `test.rst` file is missing the last two lines of the second 
note 

you forgot to insert the modified `block` into the `note` table in the `if 
isNoteEnd(block)` chunk. Insert it, and you're done!

> I guess it should have been... (about removing SoftBreaks)

well, no... take an example, a pen and paper, and do (symbolically) what 
your loop is doing...
Or if you're used to python, imagine you have a list `["a", "a", "a", ... , 
"a", "b", "c", "d", "e"]' with an indefinite number of "a"'s in the front, 
and write a program to remove all the "a"'s. The logic you found could help 
you for `block.c`.

> I hoped Lua worked as Python... (about table.remove)

Unfortunately, no, although the '-1' index trick works in Lua for some 
string functions.
Check `table.remove` in the Lua documentation for tables 
<https://www.lua.org/manual/5.3/manual.html#6.6>.

If you're more comfortable with Python, have you thought about python 
filters <https://pandoc.org/filters.html#but-i-dont-want-to-learn-haskell>? 
You could use panflute <http://scorreia.com/software/panflute/>.


-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/0a5a4d99-e862-43c4-8664-4b1b60f36e00n%40googlegroups.com.

[-- Attachment #1.2: Type: text/html, Size: 2062 bytes --]

Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[EBkysko]

[-- Attachment #1.1: Type: text/plain, Size: 893 bytes --]

... and anyways, there will be only one `SoftBreak` *at most* after/before 
the tag. So forget about the python exercise I proposed :)

Question about the format: could you have this format for a note:

```
<note>a simple note</note>
```

In that case there's no soft brake to eliminate.
So check if there's one, eliminate if so. 

On Thursday, December 3, 2020 at 3:20:28 PM UTC-5 EBkysko wrote:

> > I guess it should have been... (about removing SoftBreaks)
>
> well, no... 
>
>

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/ef8524a3-956b-4e42-adeb-761197c42d08n%40googlegroups.com.

[-- Attachment #1.2: Type: text/html, Size: 1560 bytes --]

Re: [SOCIAL]Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[Denis Bitouzé]

Le 03/12/20 à 14h44, EBkysko a écrit :

> ... and anyways, there will be only one `SoftBreak` *at most* after/before 
> the tag. So forget about the python exercise I proposed :)

I didn't see any proposed exercise :)

> Question about the format: could you have this format for a note:
>
> ```
> <note>a simple note</note>
> ```
>
> In that case there's no soft brake to eliminate.

No: `<note>` and `</note>` are always alone on their lines.

> So check if there's one, eliminate if so. 

Why is it so important to remove these soft breaks?
-- 
Denis

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/87eek6arh1.fsf%40example.com.

Re: [SOCIAL]Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[EBkysko]

[-- Attachment #1.1: Type: text/plain, Size: 1086 bytes --]

> you told me "you need to discard [...] 'SoftBreak' from the list"

that was in the *initial* example, with *that* particular list: the soft 
break between the tag and succeeding string (or preceding, for end tag)

wasn't meant to be general, so don't remove "all".

> besides, you told me "the SoftBreak's won't do any arm": I'm a bit lost

it won't... but it's weird having it there; the writer probably discards it

> Isn't `Blocks` an arbitrary name of a function?

see  Filters on element sequence 
<https://pandoc.org/lua-filters.html#filters-on-element-sequences>

> `<note>` and `</note>` are always alone on their lines

good: won't have to check another variation

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/8c84d892-5412-4fd4-9718-70866142d19dn%40googlegroups.com.

[-- Attachment #1.2: Type: text/html, Size: 1480 bytes --]

Re: [SOCIAL]Re: [SOCIAL]Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[Denis Bitouzé]

Le 04/12/20 à 06h31, EBkysko a écrit :

>> you told me "you need to discard [...] 'SoftBreak' from the list"
>
> that was in the *initial* example, with *that* particular list: the soft 
> break between the tag and succeeding string (or preceding, for end tag)
>
> wasn't meant to be general, so don't remove "all".

OK.

>> besides, you told me "the SoftBreak's won't do any arm": I'm a bit lost
>
> it won't... but it's weird having it there; the writer probably
> discards it

Indeed, AFAICS.

>> Isn't `Blocks` an arbitrary name of a function?
>
> see  Filters on element sequence 
> <https://pandoc.org/lua-filters.html#filters-on-element-sequences>

Thanks for the pointer. Currently, I don't see the added value compared
to `Pandoc`: maybe later, when I'll more experienced :)

>> `<note>` and `</note>` are always alone on their lines
>
> good: won't have to check another variation

In fact, I encountered other troubles:

- the filter didn't work when there weren't blank lines between `<note>`
  and`</note>` since, in such a case, these strings were both in the same
  `Para`,
- in my real documents, there were other "notes":
  - `<note tip>`
  - `<note warning>`
  - `<note important>`.

For the record, here is a Lua filter which seems to work (though
probably not as efficient as it could be):

--8<---------------cut here---------------start------------->8---
function isNoteStart(block)
   -- return block.t == 'Para' and block.c[1].t == 'Str' and block.c[1].text == "<note>"
   return block.t == 'Para' and block.c[1].t == 'Str' and string.match(block.c[1].text, "<note")
end

function isNoteEnd(block)
   return block.t == 'Para' and block.c[#block.c].t == 'Str' and block.c[#block.c].text == "</note>"
end

function isNoteStartEnd(block)
   return
      block.t == 'Para'
      and
      block.c[1].t == 'Str'
      and
      string.match(block.c[1].text, "<note")
      and
      block.c[#block.c].text == "</note>"
end

function isSpecialAdmonition (block)
   if block.c[3].text == "warning>" then
      isWarning = true
      table.remove(block.c, 3)
   elseif block.c[3].text == "tip>" then
      isTip = true
      table.remove(block.c, 3)
   elseif block.c[3].text == "important>" then
      isImportant = true
      table.remove(block.c, 3)
   elseif block.c[2].tag == 'Space' then
      table.remove(block.c, 3)
      table.remove(block.c, 2)
   end
end

function insertPandocDiv (contents,note)
   if isWarning then
      table.insert(contents,pandoc.Div(note, pandoc.Attr("",{"warning"})))
      isWarning = false
   elseif isTip then
      table.insert(contents,pandoc.Div(note, pandoc.Attr("",{"tip"})))
      isTip = false
   elseif isImportant then
      table.insert(contents,pandoc.Div(note, pandoc.Attr("",{"important"})))
      isImportant = false
   else
      table.insert(contents,pandoc.Div(note, pandoc.Attr("",{"note"})))
   end
end

isWarning = false
isTip = false
isImportant = false

function Pandoc(el)

   local inNote = false
   local contents = pandoc.List()
   local note = pandoc.List()
   for _,block in ipairs(el.blocks) do
      if isNoteStartEnd(block) then
         isSpecialAdmonition (block)
         note = pandoc.List()
         table.remove(block.c, 1)
         table.remove(block.c)
         table.insert(note,block)
         insertPandocDiv (contents,note)
      elseif isNoteEnd(block) then
         table.remove(block.c, #block.c)
         table.insert(note,block)
         inNote = false
         insertPandocDiv (contents,note)
      elseif isNoteStart(block) then
         inNote = true
         isSpecialAdmonition (block)
         note = pandoc.List()
         table.remove(block.c, 1)
         table.insert(note,block)
         -- for i=1,#block.c-1 do
         --    if block.c[i].t == 'SoftBreak' then
         --       table.remove(block.c[i], 1)
         --    end
         -- end
      elseif inNote then
         table.insert(note,block)
      else
         table.insert(contents,block)
      end
   end
   return pandoc.Pandoc(contents, el.meta)
end
--8<---------------cut here---------------end--------------->8---

Thanks everybody for the great help!
-- 
Denis

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/87czzpxqmg.fsf%40example.com.

Re: [SOCIAL]Re: [SOCIAL]Re: How to indent a block when exporting to reStructuredText

[Denis Bitouzé]

Le 03/12/20 à 12h20, EBkysko a écrit :

>> the resulting `test.rst` file is missing the last two lines of the second 
>> note 
>
> you forgot to insert the modified `block` into the `note` table in the `if 
> isNoteEnd(block)` chunk. Insert it, and you're done!

Indeed! Finally... :)

>> I guess it should have been... (about removing SoftBreaks)
>
> well, no... take an example, a pen and paper, and do (symbolically) what 
> your loop is doing...
> Or if you're used to python, imagine you have a list `["a", "a", "a", ... , 
> "a", "b", "c", "d", "e"]' with an indefinite number of "a"'s in the front, 
> and write a program to remove all the "a"'s. The logic you found could help 
> you for `block.c`.

In fact:

- my aim wasn't to remove any `SoftBreak` /at the start/ but any in the
  `Para` block's list (and I didn't consider the case where multiple
  consecutive occurrences of it could arise),
- I guess `SoftBreak` stands for line breaks and I don't understand why:
  - you told me "you need to discard [...] 'SoftBreak' from the
    list" (besides, you told me "the SoftBreak's won't do any arm": I'm
    a bit lost ;),
  - they are discarded from the output:
  ┌────
  │ Foo bar baz 1
  │ Foo bar baz 2
  └────
    is exported as:
  ┌────
  │ Foo bar baz 1 Foo bar baz 2
  └────
    even when the code supposed to remove them is commented.

>> I hoped Lua worked as Python... (about table.remove)
>
> Unfortunately, no, although the '-1' index trick works in Lua for some 
> string functions.
> Check `table.remove` in the Lua documentation for tables 
> <https://www.lua.org/manual/5.3/manual.html#6.6>.

Indeed, I didn't look carefully enough (sorry):

  ┌────
  │ table.remove(block.c) 
  └────

does the trick.

> If you're more comfortable with Python, have you thought about python 
> filters <https://pandoc.org/filters.html#but-i-dont-want-to-learn-haskell>? 
> You could use panflute <http://scorreia.com/software/panflute/>.

Despite the beginner's mistakes I made, I guess my problem is more with
the logic of Pandoc parser (e.g. what is a table, what is not) than the
Lua language (though I'm missing some of its syntax tricks).

Many thanks again!
-- 
Denis

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/87r1o6asg2.fsf%40example.com.