Sep 07 2015 0

Improving the Tag Cloud script

The script used to update the Tag Cloud has a couple of issues.

  1. Tags are not counted correctly when the Markdown file has been saved using DOS/Windows style line endings
  2. All Markdown files are included even if a file contains a scheduled post

The following sections describe the changes to the script The final version of the script can be downloaded here.

(1) Ensure proper (Unix style) line-endings

In case the Markdown file has been saved using DOS/Windows style line-endings

<li><a href="">font-awesome</a> <span class="badge">1</span></li>
<li><a href="
</a> <span class="badge">1</span></li>

While this would be the desired result

<li><a href="">font-awesome</a> <span class="badge">2</span></li>

The Unix command tr can be used to translate or delete characters. In this case we want to remove all carriage returns (CR) from the files via the -d option 1

# Ensure correct line-endings
for filename in *.md; do
    if grep -rq $'\r' "$filename"; then
        content=$(tr -d '\015' < $filename)
        echo "$content" > $filename

(2) Proper handling of scheduled posts

The original line of code to extract the tags from all of the posts was

tags=$(cat *.md | grep "^Tags:" | sed 's/Tags:[ ]*//' | awk -F ', ' '{for(i=1;i<=NF;i++){print $i}}' | sort -f | uniq -c)

To include the filtering of files based on the Date: meta data within the post it has been changed to the following 1

# Get tags from only the non-scheduled posts
for filename in *.md; do
    postdate=$(grep "^Date:" $filename | sed 's/Date:[ ]*//' | tr -d ' :-')'000000000000'

    if [[ "$postdate" -le "$today" ]]; then
        tags="$tags"$'\n'$(cat $filename | grep "^Tags:" | sed 's/Tags:[ ]*//' | awk -F ', ' '{for(i=1;i<=NF;i++){print $i}}')

# Sort and count unique tags
tags=$(echo "$tags" | sed '/^$/d' | sort -f | uniq -c)

Previous post
Social Sharing Buttons Recently I updated the layout for my blog to include some buttons at the end of each post so readers can easily share it to some
Next post
Changing the look-n-feel of the Archives page I was not completely satisfied with the way the Archives looked within the design template I was using. Default layout Desired
This blog is powered by Blot