Publish VimWiki to Blog

, ,

What is VimWiki?

I use VimWiki as my personal wiki and note taking tool. Sometimes what I write in my wiki I want to publish to my blog. This script aims to do that.

Requirements

In essence this is just a copy from my ~/vimwiki folder into my ~/Projects/blogs/source/_posts/ folder. Hexo (my site generator) wants some meta data added to the top of the file, so I’ll need to have that included. I’ll most likely be in vim, editing the wiki, when I decide to publish it so I can create a vim command to initialize the copy, but I don’t want to write the script in vimscript; that would be too much reliance on vim and I want to be able to publish any markdown file.

  • Copy first parameter to ~/Projects/blogs/source/_posts/ make configurable if it’s easy.
  • Add the hexo meta data to the top of the file. There is a template for this in ~/Projects/blogs/scafolds/
  • Add tags as switches to the command, as --tags "Programming, Markdown" or as --tag Programming --tag Markdown

The fun thing about writing requirements for even the smallest of scripts is you get a first shot of thinking about your interface. While looking up the location of hexo’s scaffolds folder and thought to my self “I can just use hexo to create the file and then append the vimwiki file to the end”. Will that work? Probably, it couples this tool tightly to hexo, but so does using hexo’s template format. Uncoupling from the template format is more configuration than I want to do for this small “get some automation in place script”.

I guess bash?

Originally I wanted to write this program in JavaScript and run it in with npm. There are loads of command line argument parsers for node.js and file manipulation is easy. The requirements tell me that two commands should do what I want…

Never mind… The hexo command line tools don’t have the ability to add tags from the invocation of hexo new page ....

I could use sed to parse the header and write the tags in. I’m not great at sed and I would be just copy something off stack overflow to solve the issue. I don’t want to do that.

I could write a hexo plugin but that’s whole different rabbit hole.

Maybe Perl?

A fun maxim is: “if bash scripts can get 75% of the way there Perl can finish it”[1]. You can quote a bash command in backticks (or px()) and Perl will run it and return the result. This is not a bad choice considering the short comings I describe in the previous chapter. I don’t want to “tool up” for this project so I hope the version of per that ships with MacOS is good enough, v5.30.0[2]

The Script

Process command line arguments
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
use strict;
use warnings;
use File::Basename;
use Getopt::Long qw(GetOptions);
 
my ($path) = shift;
die "Filename Required" unless $path;

my @tags;
my $taglist = "";
GetOptions("tag:s@" => \@tags, "tags:s" => \$taglist);
$taglist = $taglist . ', ' if $taglist;
$taglist = $taglist . join(', ', @tags);

my ($title,$dir,$ext) = fileparse($path, qr/\.[^.]*/);
(#save publish_from_vimwiki.pl)

This allows commands like publish_from_vimwiki.pl ~/vimwiki/somefile.md --tags "SomeTag, Two, Three" the GetOpt() call will look for individual tags in --tag or/and a comma separated list of tags in --tags. Everything ends up in the $tagslist scalar.

Create blog post
1
2
3
4
5
6
my @result = qx(hexo --cwd="~/Projects/blog" new post "$title");
my $blogfile = $result[-1];
$blogfile =~ s/INFO  Created:\s+([^\r\n]+)[\r\n]/$1/;
$blogfile = glob($blogfile);

rename($blogfile, $blogfile.'.bak') or die $!;
(#save publish_from_vimwiki.pl)

Here we crate the new post, calling hexo new post directly. Then munch the results of the hexo command to get the name of the file it created. We make a back of the new post; need to remove that after we are done:

1
unlink $blogfile.'.bak';
(#save publish_from_vimwiki.pl|cleanup)

Now we can copy the backup file back into the blog post file modifying the tags along the way.

Write the header of the blog w/ tags
1
2
3
4
5
6
7
8
open(BLOGIN, '<'.$blogfile.'.bak') or die $!;
open(OUT, '>'.$blogfile) or die $!;

while(my $line = <BLOGIN>) {
    $line =~ s/tags:/tags: [$taglist]/g;
    print OUT $line;
}
close(BLOGIN);
(#save publish_from_vimwiki.pl)

Finally we are ready to append with VimWiki file contents to the blog post.

1
2
3
4
5
6
7
8
open WIKIIN, '<', $path or die $!;

while (my $line = <WIKIIN>) {
    print OUT $line;
}

close(WIKIIN);
close(OUT);
(#save publish_from_vimwiki.pl)
  1. 1.

    This is never true.

    ↩
  2. 2.

    It is fine.

    ↩