feat: explicitly listed and orderable categories

3 files changed, 87 insertions, 30 deletions

diff --git a/README.md b/README.md
index 03ab255..d5e4db5 100644
--- a/README.md
+++ b/README.md
@@ -17,7 +17,7 @@ chmod +x /usr/bin/ssg-blog
 ssg-blog src dst 'https://my-website.net'
 ```
 
-ssg-blog checks for subdirectories in the src directory, each of which becomes a category heading (non-recursively). It will generate a sorted list of links to any .md file in those directory (including nested directories).
+Blog categories are hard-coded in the `CATEGORIES` variable at the top of the script. Each one becomes a category heading in the order listed. Blog posts should be in a directory with the same name.
 
 ```
 ssg-blog/
@@ -60,14 +60,6 @@ Note a few things about the output:
 2. Files named index.md will get have a link generated without <filename>.html at the end.
 3. Files outside subdirectories (e.g. about.md) will be rendered but won't show up on the index.
 
-## Change how categories are sorted
-
-The category headings are sorted alphabetically. If you want to change this, you have to edit the `sort` on this line:
-
-```
-find "$src" -mindepth 1 -maxdepth 1 -type d ! -path "*/.*" | sort |
-```
-
 # Differences from ssg
 
 The title argument from ssg is unnecessary so I've removed it. `render_html_file` used it to modify the first `<h1>` tag and auto-populate the `<title>` if none is specified, which is functionality I've never needed. Instead I check the frontmatter of each blog post and update the `<title>` tag with the `title:` field.
diff --git a/demo/example.webp b/demo/example.webp
deleted file mode 100644
index d9153d9..0000000
--- a/demo/example.webp
+++ /dev/null
diff --git a/ssg-blog b/ssg-blog
index 398aa77..44997e0 100755
--- a/ssg-blog
+++ b/ssg-blog
@@ -16,6 +16,13 @@
 # OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 #
 
+# Blog categories: top-level source directories that make up the blog,
+# in the order they appear on the index page. .html files inside these
+# directories get the site header and footer; .html files anywhere else
+# are copied verbatim (e.g. standalone pages like JS tools). If empty,
+# every top-level directory is a category, sorted alphabetically.
+CATEGORIES="games lists money technology tutorials"
+
 main() {
   test -n "$1" || usage
   test -n "$2" || usage
@@ -73,9 +80,13 @@ main() {
     fi
 
     echo "$files" | grep '\.html$' |
+      wrap_html |
       render_html_files "$src" "$dst"
 
-    echo "$files" | grep -Ev '\.md$|\.html$' |
+    {
+      echo "$files" | grep '\.html$' | verbatim_html
+      echo "$files" | grep -Ev '\.md$|\.html$'
+    } |
       (cd "$src" && cpio -pu "$dst")
   fi
 
@@ -192,6 +203,45 @@ render_html_file() {
   }
 }
 
+in_category_dir() {
+  f="${1#./}"
+  dir="${f%%/*}"
+  if [ "$dir" = "$f" ]; then
+    return 1
+  fi
+  for category in $CATEGORIES; do
+    if [ "$dir" = "$category" ]; then
+      return 0
+    fi
+  done
+  return 1
+}
+
+wrap_html() {
+  if [ -z "$CATEGORIES" ]; then
+    cat
+    return 0
+  fi
+  while read -r f; do
+    if in_category_dir "$f"; then
+      echo "$f"
+    fi
+  done
+  return 0
+}
+
+verbatim_html() {
+  if [ -z "$CATEGORIES" ]; then
+    return 0
+  fi
+  while read -r f; do
+    if ! in_category_dir "$f"; then
+      echo "$f"
+    fi
+  done
+  return 0
+}
+
 list_pages() {
   e="\\( -name '*.html' -o -name '*.md' \\)"
   cd "$1" && eval "find . -type f ! -path '*/.*' ! -path '*/_*' $IGNORE $e" |
@@ -206,25 +256,40 @@ list_pages() {
     sed 's#^./##;s#.md$#.html#;s#/index.html$#/#'
 }
 
+list_category_articles() {
+  # $1 = src, $2 = category, $3 = category order
+  find "$1/$2" -type f -name "*.md" ! -path "*/.*" |
+  while read -r article; do
+    test "$(get_frontmatter_field "$article" "publish")" != "false" || continue
+    date=$(get_frontmatter_field "$article" "date")
+    test -n "$date" || continue
+    title=$(get_frontmatter_field "$article" "title")
+    url="${article#"$1"}"
+    url="${url%.md}.html"
+    if [ "${url##*/}" = "index.html" ]; then
+      url="${url%/index.html}/"
+    fi
+    url="${url#/}"
+    echo "$3|$2|$date|$title|$url"
+  done
+}
+
 list_articles() {
-  find "$1" -mindepth 1 -maxdepth 1 -type d ! -path "*/.*" | sort |
-  while read -r category_dir; do
-    category=$(basename "$category_dir")
-    find "$category_dir" -type f \( -name "*.md" -o -name "index.md" \) ! -path "*/.*" |
-    while read -r article; do
-      test "$(get_frontmatter_field "$article" "publish")" != "false" || continue
-      date=$(get_frontmatter_field "$article" "date")
-      test -n "$date" || continue
-      title=$(get_frontmatter_field "$article" "title")
-      url="${article#"$1"}"
-      url="${url%.md}.html"
-      if [ "${url##*/}" = "index.html" ]; then
-        url="${url%/index.html}/"
+  if [ -n "$CATEGORIES" ]; then
+    n=0
+    for category in $CATEGORIES; do
+      if [ -d "$1/$category" ]; then
+        n=$((n+1))
+        list_category_articles "$1" "$category" "$n"
       fi
-      url="${url#/}"
-      echo "$category|$date|$title|$url"
     done
-  done
+  else
+    find "$1" -mindepth 1 -maxdepth 1 -type d ! -path "*/.*" | sort |
+    while read -r category_dir; do
+      n=$((n+1))
+      list_category_articles "$1" "$(basename "$category_dir")" "$n"
+    done
+  fi
 }
 
 render_index() {
@@ -232,8 +297,8 @@ render_index() {
   echo "$HEADER" > "$output"
 
   last_category=""
-  list_articles "$src" | sort -t'|' -k1,1 -k2,2r |
-  while IFS='|' read -r category date title url; do
+  list_articles "$src" | sort -t'|' -k1,1n -k3,3r |
+  while IFS='|' read -r order category date title url; do
     if [ "$category" != "$last_category" ]; then
       display_category=$(echo "$category" | sed 's/_/ /g' | awk '{for(i=1;i<=NF;i++){s=$i;$i=toupper(substr(s,1,1)) substr(s,2)}; print}')
       echo "<h1>$display_category</h1>" >> "$output"
@@ -258,7 +323,7 @@ rfc822_date() {
 }
 
 render_rss() {
-  list_articles "$src" | sort -t'|' -k2,2r |
+  list_articles "$src" | sort -t'|' -k3,3r |
   {
     echo '<?xml version="1.0" encoding="UTF-8"?>'
     echo '<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">'
@@ -268,7 +333,7 @@ render_rss() {
     echo "<description>$base_url</description>"
     echo "<atom:link href=\"$base_url/rss.xml\" rel=\"self\" type=\"application/rss+xml\"/>"
 
-    while IFS='|' read -r category date title url; do
+    while IFS='|' read -r order category date title url; do
       escaped_title=$(echo "$title" | xml_escape)
       escaped_url=$(echo "$url" | xml_escape)
       pub_date=$(rfc822_date "$date")