I’m writing blogs in org-mode and..
Not to forget how to write down in org-mode or literate haskell, I keep writing my blog in .lhs or .org and I realized that when I need to post in dev.to. I found myself being a coverter. It is not very hard but annoying process. And I tried pandoc as a commandline tool.
A programmer is a person who is willing to spend 2 hours to make an automatic system, to save 2 minutes for now. 😱
So I was looking for file format converter. which is actually something I always use when I use hakyll – pandoc!
Install pandoc
you can install via pacman on arch linux!
sh> sudo pacman -S pandoc
resolving dependencies...
looking for conflicting packages...
Packages (129) haskell-aeson-1.5.6.0-84 haskell-aeson-pretty-0.8.9-35
haskell-ansi-terminal-0.11-218 haskell-appar-0.1.8-13
haskell-asn1-encoding-0.9.6-154 haskell-asn1-parse-0.9.5-154
haskell-asn1-types-0.3.4-133 haskell-assoc-1.0.2-124
haskell-async-2.2.4-14 haskell-attoparsec-0.14.4-3
.. snip ..
haskell-x509-store-1.6.9-5 haskell-x509-system-1.6.7-8
haskell-x509-validation-1.6.12-10 haskell-xml-1.3.14-30
haskell-xml-conduit-1.9.1.1-85 haskell-xml-types-0.3.8-8
haskell-zip-archive-0.4.1-161 haskell-zlib-0.6.2.3-62
pandoc-2.14.2-42
Total Download Size: 44.20 MiB
Total Installed Size: 297.95 MiB
:: Proceed with installation? [Y/n] nOh… Kay… abort mission. Because I’m using stack and stack has installed many haskell libraries already.
sh> stack install pandoc
.. compiling ....TBH, if you are compiling for the first time, it would take a lot of time.
ghc is not blazing fast software. It is up to you, but if you don’t use other haskell
software, I recommends you try to install it via your package system first.
pandoc -t markdown
sh> pandoc -to markdown some-article.org > out.markdownand that’s it.
Fix metadata
However, you might need to fix your metadata, if you repost in markdown format within some sort of static stie builder like hakyll.
After previous command pandoc -t markdwon, metadata unusable because it looks like below:
--- title: Builtin Benchmark with Go description: writing the benchmark
code in go langauge. keywords: go, golang, benchmark, org-mode, hakyll
author: Myoungjin Jeon ---
You can edit by yourself on your favourite editor or below awk programme can take out the metadata part.
save below as hakyll-pandoc-preproc.awk
#!/usr/bin/awk -f
BEGIN {
numof_printed = -1;
}
{
if ( $0 ~ /^--*$/ ) {
numof_printed++;
if ( numof_printed > 0 ) {
print $0;
exit;
}
}
if ( numof_printed >= 0 ) {
print $0;
numof_printed++;
}
}and extact like below:
sh> chmod u+x hakyll-pandoc-preproc.awk
sh> ./hakyll-pandoc-preproc.awk "your-document.org"
---
title: Builtin Benchmark with Go
description: writing the benchmark code in go langauge.
keywords: go, golang, benchmark, org-mode, hakyll
author: Myoungjin Jeon
---Or.. maybe it is better for other script language to handle it. I tried with fish shell because yesterday I talked about fish shell is good for scripting as well!
#!/usr/bin/env fish
# filename: pandoc-any-to-markdown.fish
# version: 0.1
set -l input_filepath $argv[1]
if not test -f $input_filepath
echo \
"Usage: <document file path> > <output file path>
will print out converted text to stdout.
"
exit 1;
end
set -l numof_printed -1
# ^ indicating phase -1: meta not found (>=0): meta found
set -l pandoc_from (string split "." $input_filepath | tail -n1)
# ^ pandoc reader relys on extension
begin
while read -l line
if string match -q --regex '^---' -- "$line"
set numof_printed (math "$numof_printed + 1")
if test $numof_printed -gt 0
echo $line
break;
end
end
if test $numof_printed -ge 0
echo $line
set numof_printed (math "$numof_printed + 1")
end
end
pandoc -f $pandoc_from -t markdown -
end < $input_filepathYou can find the recent version at here.
And make it executable and run! (now I found myself beging an installer..)
sh> chmod u+x pandoc-any-to-markdown.fish
sh> ./pandoc-any-to-markdown.fish test-pandoc.orgI think I can post more articles to dev.to even if I wrote down in org-mode or literate haskell. Oki… That’s it all for today!