#!/bin/sh

# any *.md will be processed (ie. have _header, date and _footer added)
# a *.md file won't be processed if it starts with an _
# any other file will be copied to dst

export src="$(readlink -f blog_raw)"
export dst="$(readlink -f blog)"
export SITENAME="yoannlr's blog"
export SITEURL='https://yoannlr.github.io/blog'
export HEADER=$(cat "$src/_header.html")
export FOOTER=$(cat "$src/_footer.html")

export RSSFILE='rss.xml'
export RSSHEAD="$src/_rss.xml"
export RSSURL="$SITEURL/$RSSFILE"
export RSSTITLE="yoannlr's blog"
export RSSLANG="en-us"
export RSSDESC="Latest news from yoannlr"

main() {
	cd "$src"

	# replicate folder structure to dst
	# 'cpio -pdvm' for verbose and to keep mtime
	find . -type d | cpio -pd "$dst"

	# clear rss file and _list before processing articles
	sed "s#RSSTITLE#$RSSTITLE#g;s#RSSDESC#$RSSDESC#g;s#RSSLANG#$RSSLANG#g;s#RSSURL#$RSSURL#g" $RSSHEAD > $RSSFILE
	echo '' > "$src/_list"

	# render articles as html files
	find . -type f -name '*.md' ! -name '*_index.md' ! -name '_*' -printf '%T@\t%p\n' | sort -r | cut -d$'\t' -f2 | colrm 1 2 | render_articles isart
	echo -e '</channel>\n</rss>' >> $RSSFILE

	echo "Articles processed"

	# generate index
	cat "$src/_index.md" "$src/_list" > "$src/index.md"
	echo 'index.md' | render_articles
	rm "$src/index.md" 

	# copy other files (such as images) to dst
	find . -type f ! -name '*.md' ! -name '_*' -exec cp -- "{}" "$dst/{}" \;
}

render_articles() {
	while read art
	do
		artdate=$(date -r "$art" '+%d %b %Y')
		rssdate=$(date -R -r "$art")
		lowdown < "$art" | awk -v isart="$1" -v date="$artdate" -v rssdate="$rssdate" -v article="${art%\.md}.html" -v artlist="_list" '
		BEGIN {
			title = ENVIRON["SITENAME"]
			t = ""
		}
		{
			body = body "\n" $0
			if (isart == "isart" && match($0, /<[[:space:]]*[Hh]1(>|[[:space:]][^>]*>)/)) {
				body = body "\n<p class=\"date\">" date "</p>\n"
			}
		}
		END {
			body = substr(body, 2)
			if (body ~ /<[Hh][Tt][Mm][Ll]/) {
				print body
				exit
			}
			if (match(body, /<[[:space:]]*[Hh]1(>|[[:space:]][^>]*>)/)) {
				t = substr(body, RSTART + RLENGTH)
				sub("<[[:space:]]*/[[:space:]]*[Hh]1.*", "", t)
				gsub(/^[[:space:]]*|[[:space:]]$/, "", t)
				if (isart == "isart") printf "* %s &mdash; [%s](%s)\n", date, t, article >> artlist
				title = t " &mdash; " title
			}
			n = split(ENVIRON["HEADER"], header, /\n/)
			for (i = 1; i <= n; i++) {
				if (match(tolower(header[i]), "<title></title>")) {
					head = substr(header[i], 1, RSTART - 1)
					tail = substr(header[i], RSTART + RLENGTH)
					print head "<title>" title "</title>" tail
				} else print header[i]
			}
			if (isart == "isart") {
				printf "<item>\n<title>%s</title>\n<guid>%s/%s</guid>\n<pubDate>%s</pubDate>\n<description>\n<![CDATA[\n%s\n]]>\n</description>\n</item>\n", t, ENVIRON["SITEURL"], article, rssdate, body >> ENVIRON["RSSFILE"]
				close(ENVIRON["RSSFILE"])
			}
			print body
			print ENVIRON["FOOTER"]
		}' > "$dst/${art%\.md}.html"
	done
}

main