Group Markdown Links by Their Domains

group-markdown-links.sh

#!/usr/bin/env bash

# Limit the number of links, under which they go to the miscellaneous section
limit=3

# Define newline character
nl=$'\n'

# Storage of miscellaneous links - this is the section containing
# too few links to be included in a separate domain group
misc="";

# First get the lines in question [1]
original=$(echo "$1")
# Then sort all links by their names
sorted=$(echo "$original" | sort | uniq)
# Then sort all links by their domains [2]
master=$(cat $sorted | sed 's/\(.*:\/\/\(www\.\)*\)/\1\	/' | sort -t$'\t' -k2 | sed 's/	//')

# Main section
# with links grouped by their domains
echo "### Links by Domain"
echo ""

while IFS= read -r line; do
	# Get links as text to reuse it later
	links_txt=$(echo "$master" | grep "$line")
	# Get number of links belonging to one domain
	links_num=$(echo "$links_txt" | wc -l)
	# Parse lines
	if [ -n "$line" ]; then
		# Add lines to the miscellaneous section, if it contains few links
		if [ "$links_num" -le "$limit" ]; then
			misc="$misc${nl}$links_txt"
		# Write lines out directly, if there are much enough links
		else
			echo "#### $line"
			echo ""
			echo "$links_txt"
			echo ""
		fi
	fi
done < <(echo "$master" |\
# Remove everything until the URL
sed 's/\(.*](http[s]*:\/\/\)\(.*\)/\2/' |\
# Remove subfolder(s) from the URL
sed 's/\([^\/]*\)\(\/\)\(.*\)/\1/' |\
# Remove subdomain
sed 's/\([^.]\{1,\}\)\([.]\{1\}\)\([^.]\{1,\}\)\(.\)\([.]\{1\}\)/\3\4\5/' |\
# Sort result
sort |\
# Keep unique lines only
uniq)

# Miscellaneous section
# with assorted links without grouping
echo "### Miscellaneous links"
echo "$misc"

# --------------------------------------------------
# COMMENTS
# --------------------------------------------------
#
# [1] For using this script in a text editor (i.e. not as
#     a bash text filter) change $1 to EDITOR_INPUT_VARIABLE.
#     (E.g. in case of TextMate it is $TM_SELECTED_TEXT.)
#
# [2] As the syntax \t for a tab character in sed is not standard:
#     OS X sed (...) doesn't support \t for tab and instead treats
#     \t as meaning backslash followed by t.
#     Therefore the easiest way to circumvent this headache is
#     to use a literal tab character.
#     See https://goo.gl/11Zg1s