Creation of a filter that accepts an option

Creation of a filter that accepts an option

[kiko.fernandez-1zs4UD6AkMk]

[-- Attachment #1.1: Type: text/plain, Size: 2251 bytes --]

Hi,

I am trying to create a filter that shows/hides notes. I have the notation 
that if there is a fenced code block with attribute "note" then the fenced 
block is really a note.
With this filter, I will remove all "note" blocks from the final output and 
my supervisor will only see the good stuff. However, I would like to 
re-include the notes for myself, in case I read the PDF or HTML version. Is 
there a way to create a filter that, based on the first character or option 
that I pass, decides to show or hide these notes?

My code so far looks like this:

```
module Main where

import Text.Pandoc.JSON
import Debug.Trace

removeFenceNotes :: Block -> Block
removeFenceNotes cb@(CodeBlock (id, classes, namevals) contents) =
  case lookup "note" namevals of
       Just f     -> Div ("", [], []) []
       Nothing    -> cb
removeFenceNotes x = x

main :: IO ()
main = do c <- getChar
          if c == 'e' then toJSONFilter removeFenceNotes
          else toJSONFilter removeFenceNotes -- Should show notes removing 
the code block and reinterpreting the text
```

When I decide to show notes, I would like to remove the code block and 
reinterpret the markdown in the standard way.

I got it to work if there is no case to consider, i.e., no 'e' (for 
enabled) or such stuff.
Is this possible to do?

Thanks!

P.S. I do not mind to do it using footnotes, since there is no 
reinterpretation of the string received. However I thought that it would be 
a good exercise to see if this is doable.
If I get this to work, I would like to write a brief article on 
opensource.com in line with recent articles that I have been putting there. 
I hope that this is ok.

Thanks.

-- 
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 post to this group, send email to pandoc-discuss-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/643e81d5-c3a2-4f68-927c-0500bb6ffd46%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[-- Attachment #1.2: Type: text/html, Size: 3059 bytes --]

Re: Creation of a filter that accepts an option

[Robert Zenz]

You could use environment variables. Something like this:

    $ KEEP_NOTES=true pandoc ...

And in the Lua filter:

    if os.getenv("KEEP_NOTES") == nil then
        -- Hide the notes here.
    end

On 02.10.2018 15:55, kiko.fernandez-1zs4UD6AkMk@public.gmane.org wrote:
> Hi,
> 
> I am trying to create a filter that shows/hides notes. I have the notation 
> that if there is a fenced code block with attribute "note" then the fenced 
> block is really a note.
> With this filter, I will remove all "note" blocks from the final output and 
> my supervisor will only see the good stuff. However, I would like to 
> re-include the notes for myself, in case I read the PDF or HTML version. Is 
> there a way to create a filter that, based on the first character or option 
> that I pass, decides to show or hide these notes?
> 
> My code so far looks like this:
> 
> ```
> module Main where
> 
> import Text.Pandoc.JSON
> import Debug.Trace
> 
> removeFenceNotes :: Block -> Block
> removeFenceNotes cb@(CodeBlock (id, classes, namevals) contents) =
>   case lookup "note" namevals of
>        Just f     -> Div ("", [], []) []
>        Nothing    -> cb
> removeFenceNotes x = x
> 
> main :: IO ()
> main = do c <- getChar
>           if c == 'e' then toJSONFilter removeFenceNotes
>           else toJSONFilter removeFenceNotes -- Should show notes removing 
> the code block and reinterpreting the text
> ```
> 
> When I decide to show notes, I would like to remove the code block and 
> reinterpret the markdown in the standard way.
> 
> I got it to work if there is no case to consider, i.e., no 'e' (for 
> enabled) or such stuff.
> Is this possible to do?
> 
> Thanks!
> 
> P.S. I do not mind to do it using footnotes, since there is no 
> reinterpretation of the string received. However I thought that it would be 
> a good exercise to see if this is doable.
> If I get this to work, I would like to write a brief article on 
> opensource.com in line with recent articles that I have been putting there. 
> I hope that this is ok.
> 
> Thanks.
> 

Re: Creation of a filter that accepts an option

[Robert Zenz]

Well, just realized you don't use Lua...but I guess the mechanic is the same, or
at least close.

Also you'd only provide the environment variable if you wanted to keep the
notes, obviously.

On 02.10.2018 16:01, Robert Zenz wrote:
> You could use environment variables. Something like this:
> 
>     $ KEEP_NOTES=true pandoc ...
> 
> And in the Lua filter:
> 
>     if os.getenv("KEEP_NOTES") == nil then
>         -- Hide the notes here.
>     end
> 
> On 02.10.2018 15:55, kiko.fernandez-1zs4UD6AkMk@public.gmane.org wrote:
>> Hi,
>>
>> I am trying to create a filter that shows/hides notes. I have the notation 
>> that if there is a fenced code block with attribute "note" then the fenced 
>> block is really a note.
>> With this filter, I will remove all "note" blocks from the final output and 
>> my supervisor will only see the good stuff. However, I would like to 
>> re-include the notes for myself, in case I read the PDF or HTML version. Is 
>> there a way to create a filter that, based on the first character or option 
>> that I pass, decides to show or hide these notes?
>>
>> My code so far looks like this:
>>
>> ```
>> module Main where
>>
>> import Text.Pandoc.JSON
>> import Debug.Trace
>>
>> removeFenceNotes :: Block -> Block
>> removeFenceNotes cb@(CodeBlock (id, classes, namevals) contents) =
>>   case lookup "note" namevals of
>>        Just f     -> Div ("", [], []) []
>>        Nothing    -> cb
>> removeFenceNotes x = x
>>
>> main :: IO ()
>> main = do c <- getChar
>>           if c == 'e' then toJSONFilter removeFenceNotes
>>           else toJSONFilter removeFenceNotes -- Should show notes removing 
>> the code block and reinterpreting the text
>> ```
>>
>> When I decide to show notes, I would like to remove the code block and 
>> reinterpret the markdown in the standard way.
>>
>> I got it to work if there is no case to consider, i.e., no 'e' (for 
>> enabled) or such stuff.
>> Is this possible to do?
>>
>> Thanks!
>>
>> P.S. I do not mind to do it using footnotes, since there is no 
>> reinterpretation of the string received. However I thought that it would be 
>> a good exercise to see if this is doable.
>> If I get this to work, I would like to write a brief article on 
>> opensource.com in line with recent articles that I have been putting there. 
>> I hope that this is ok.
>>
>> Thanks.
>>
> 

Re: Creation of a filter that accepts an option

[Kolen Cheung]

[-- Attachment #1.1: Type: text/plain, Size: 731 bytes --]

c.f. the thread 
in https://groups.google.com/d/msg/pandoc-discuss/LIAfgkZKUiE/_KfI3VI6EgAJ about 
filter arg. In short env. var. is the way to go right now. 

-- 
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 post to this group, send email to pandoc-discuss-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/6d5f7166-099c-4fa7-92c9-b60146ea97a5%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[-- Attachment #1.2: Type: text/html, Size: 1146 bytes --]

Re: Creation of a filter that accepts an option

[kiko.fernandez-1zs4UD6AkMk]

[-- Attachment #1.1: Type: text/plain, Size: 4223 bytes --]

Thanks everyone, the env vars worked perfectly well.

I got it to almost work doing reinterpretation of the AST.
There is just one tiny thing that doesn't seem to work, and that is adding 
a citation to a Listing.

I am going to copy-paste the filter and the example. Hopefully, someone who 
knows more Pandoc
internals can help me understand what is going on, and possibly, how to fix 
it.

Thanks


```Haskell filter
module Main where

import Text.Pandoc.JSON
import System.IO.Error (catchIOError)
import System.Environment (getEnv)
import Text.Pandoc
import Data.Text (Text, pack)
import qualified Data.Map as M

writeMd :: Pandoc -> PandocIO Text
writeMd p = writeMarkdown def p

readMd :: Text -> PandocIO Pandoc
readMd t = readMarkdown def t

removeFenceNotes :: Block -> Block
removeFenceNotes cb@(CodeBlock (_, _, namevals) _) =
  case lookup "note" namevals of
       Just _     -> Div ("", [], []) []
       Nothing    -> cb
removeFenceNotes x = x

reinterpretation :: Block -> IO [Block]
reinterpretation cb@(CodeBlock (_, _, namevals) text) =
  case lookup "note" namevals of
    Just _ -> do
      p <- runIOorExplode $ readMd (pack text)
      textMd <- runIOorExplode $ writeMd p
      p' <- runIOorExplode $ readMd textMd
      case p' of
        Pandoc _ b -> return b
    Nothing -> return [cb]
reinterpretation x = return [x]

main = do
  v <- catchIOError (getEnv "FILTERING_MODE") (\_ -> return "")
  case v of
    "WITH_COMMENTS" -> toJSONFilter reinterpretation
    "NO_COMMENTS"  -> toJSONFilter removeFenceNotes -- disable comments
    _ -> return ()
```

I call Pandoc with the following command:

```
FILTERING_MODE=WITH_COMMENTS pandoc -s -f markdown -F filters -F 
pandoc-crossref meta.yaml -o test.pdf TEST.md
```

and the TEST.md file looks as follows:

```
# Capabilities status

```{#lst:maybeLinearSupport .encore
    caption="Support for `Maybe`'s do not fully work with `linear` types."}

subord class Two
  var x: int
  def init(): unit
    this.x = 5
  end
end

linear class One: Id
  var sub: Two
  def init(): unit
    this.sub = new Two()
  end
  def x(): int
    this.sub.x
  end
  def setX(x: int): unit
    this.sub.x = x
  end
end

-- this function does not compile. I cannot get it to
-- compile in any possible way

fun moo(var m: Maybe[One]): One
  match consume m with
    case Nothing => consume new One()
    case Just(z) => consume(z)
  end
end

```


In @lst:maybeLinearSupport, if `One` is `read`, then the function `moo` who 
receives
a `Maybe[One]` works. The distinction between a `read Maybe[One]` and a 
`linear Maybe[One]`
prevents us from reusing this function, i.e., the annotations of `var` and 
`val`
are different and the function cannot be polymorphic on its *mode*.

~~~{note=""}
linearity seems to be more related to a property of the object rather than
a mode by itself. this means, @lst:maybeLinearSupport can declare a 
function that
we know does not use any linear and we could declare another function that 
may
be `read`, `local` or any other mode that has to be treated *linearly*.
what are the implications of this?
~~~

```

The end result contains the notes without the code block scope and also 
contain the
all markdown items well-formatted, i.e., bold, italics, etc. The only thing 
that didn't work
was reading the reference `@lst:maybeLinearSupport`, which is printed as it 
looks, instead
of referencing the figure.

Thanks.



On Tuesday, October 2, 2018 at 11:39:41 PM UTC+2, Kolen Cheung wrote:
>
> c.f. the thread in 
> https://groups.google.com/d/msg/pandoc-discuss/LIAfgkZKUiE/_KfI3VI6EgAJ about 
> filter arg. In short env. var. is the way to go right now. 
>

-- 
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 post to this group, send email to pandoc-discuss-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/53d16307-c278-4f3a-8934-f44d9e40862c%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[-- Attachment #1.2: Type: text/html, Size: 6103 bytes --]

Re: Creation of a filter that accepts an option

[kiko.fernandez-1zs4UD6AkMk]

[-- Attachment #1.1: Type: text/plain, Size: 2972 bytes --]

I fixed it walking a bit more. For some reason, the citation @XXXX is read 
as a (Str x)
value instead of a Cite value. Walking the tree and pattern matching solved 
it.

Thanks.

I will hopefully publish this in opensource.com in a month or so, maybe 
earlier.
I think the "walk" function is awesome and more people should know that 
building
a filter is not that difficult and it is fun!

I will post the link to the article once it is published, so that the 
community can
read how I built it and how it works.

Thanks everyone

On Tuesday, October 2, 2018 at 3:55:10 PM UTC+2, kiko.fe...-1zs4UD6AkMk@public.gmane.org wrote:
>
> Hi,
>
> I am trying to create a filter that shows/hides notes. I have the notation 
> that if there is a fenced code block with attribute "note" then the fenced 
> block is really a note.
> With this filter, I will remove all "note" blocks from the final output 
> and my supervisor will only see the good stuff. However, I would like to 
> re-include the notes for myself, in case I read the PDF or HTML version. Is 
> there a way to create a filter that, based on the first character or option 
> that I pass, decides to show or hide these notes?
>
> My code so far looks like this:
>
> ```
> module Main where
>
> import Text.Pandoc.JSON
> import Debug.Trace
>
> removeFenceNotes :: Block -> Block
> removeFenceNotes cb@(CodeBlock (id, classes, namevals) contents) =
>   case lookup "note" namevals of
>        Just f     -> Div ("", [], []) []
>        Nothing    -> cb
> removeFenceNotes x = x
>
> main :: IO ()
> main = do c <- getChar
>           if c == 'e' then toJSONFilter removeFenceNotes
>           else toJSONFilter removeFenceNotes -- Should show notes removing 
> the code block and reinterpreting the text
> ```
>
> When I decide to show notes, I would like to remove the code block and 
> reinterpret the markdown in the standard way.
>
> I got it to work if there is no case to consider, i.e., no 'e' (for 
> enabled) or such stuff.
> Is this possible to do?
>
> Thanks!
>
> P.S. I do not mind to do it using footnotes, since there is no 
> reinterpretation of the string received. However I thought that it would be 
> a good exercise to see if this is doable.
> If I get this to work, I would like to write a brief article on 
> opensource.com in line with recent articles that I have been putting 
> there. I hope that this is ok.
>
> Thanks.
>

-- 
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 post to this group, send email to pandoc-discuss-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/e28b2c92-2f45-4c2a-9bb6-8084c2f95ad3%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[-- Attachment #1.2: Type: text/html, Size: 4408 bytes --]

Re: Creation of a filter that accepts an option

[John MacFarlane]

kiko.fernandez-1zs4UD6AkMk@public.gmane.org writes:

> I fixed it walking a bit more. For some reason, the citation @XXXX is read 
> as a (Str x)
> value instead of a Cite value. Walking the tree and pattern matching solved 
> it.

Perhaps because you're running your filter after
pandoc-crossref, which would convert it to a Str?

Yes, filters are definitely fun and not too hard.
Good to publicize that!

Re: Creation of a filter that accepts an option

[kiko.fernandez-1zs4UD6AkMk]

[-- Attachment #1.1: Type: text/plain, Size: 1480 bytes --]

Hi John,

thanks for your advise. However, wouldn't this apply the my filter, named 
filters, first and then the pandoc-cross-ref?

FILTERING_MODE=WITH_COMMENTS pandoc -s -f markdown -F filters -F 
pandoc-crossref meta.yaml -o test.pdf TEST.md

In any case, I tried changing the order and didn't get it to work without 
pattern matching on the Str data value.
I found that weird.

On Thursday, October 4, 2018 at 6:23:05 PM UTC+2, John MacFarlane wrote:
>
> kiko.fe...-1zs4UD6AkMk@public.gmane.org <javascript:> writes: 
>
> > I fixed it walking a bit more. For some reason, the citation @XXXX is 
> read 
> > as a (Str x) 
> > value instead of a Cite value. Walking the tree and pattern matching 
> solved 
> > it. 
>
> Perhaps because you're running your filter after 
> pandoc-crossref, which would convert it to a Str? 
>
> Yes, filters are definitely fun and not too hard. 
> Good to publicize that! 
>

-- 
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 post to this group, send email to pandoc-discuss-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/f91cf9f6-8d38-43ac-b3a5-b21070c66e14%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[-- Attachment #1.2: Type: text/html, Size: 2368 bytes --]