Add RuneCount to Page

Fixes #1266

docs/content/templates/variables.md

@@ -47,6 +47,7 @@ matter, content or derived from file location.
 **.NextInSection** Pointer to the following content within the same section (based on pub date)<br>
 **.FuzzyWordCount** The approximate number of words in the content.<br>
 **.WordCount** The number of words in the content.<br>
+**.RuneCount** The number of [runes](http://blog.golang.org/strings) in the content, excluding any whitespace. This may be a good alternative to `.WordCount`  for Japanese and other CJK languages where a word-split by spaces makes no sense.
 **.ReadingTime** The estimated time it takes to read the content in minutes.<br>
 **.Weight** Assigned weight (in the front matter) to this content, used in sorting.<br>
 **.RawContent** Raw Markdown content without the metadata header. Useful with [remarkjs.com](http://remarkjs.com)<br>

helpers/general.go

@@ -228,6 +228,11 @@ func Md5String(f string) string {
 	return hex.EncodeToString(h.Sum([]byte{}))
 }
 
+// IsWhitespace determines if the given rune is whitespace.
+func IsWhitespace(r rune) bool {
+	return r == ' ' || r == '\t' || r == '\n' || r == '\r'
+}
+
 // Seq creates a sequence of integers.
 // It's named and used as GNU's seq.
 // Examples:

hugolib/page.go

@@ -66,6 +66,7 @@ type Page struct {
 	contentShortCodes   map[string]string
 	plain               string // TODO should be []byte
 	plainWords          []string
+	plainRuneCount      int
 	plainInit           sync.Once
 	renderingConfig     *helpers.Blackfriday
 	renderingConfigInit sync.Once
@@ -108,10 +109,24 @@ func (p *Page) PlainWords() []string {
 	return p.plainWords
 }
 
+// RuneCount returns the rune count, excluding any whitespace, of the plain content.
+func (p *Page) RuneCount() int {
+	p.initPlain()
+	return p.plainRuneCount
+}
+
 func (p *Page) initPlain() {
 	p.plainInit.Do(func() {
 		p.plain = helpers.StripHTML(string(p.Content))
 		p.plainWords = strings.Fields(p.plain)
+		runeCount := 0
+		for _, r := range p.plain {
+			if !helpers.IsWhitespace(r) {
+				runeCount++
+			}
+		}
+		p.plainRuneCount = runeCount
+		return
 	})
 }
 

hugolib/page_test.go

@@ -138,6 +138,16 @@ title: Simple
 Summary Same Line<!--more-->
 
 Some more text
+`
+
+	SIMPLE_PAGE_WITH_FIVE_MULTIBYTE_UFT8_RUNES = `---
+title: Simple
+---
+
+
+€ € € € €
+
+
 `
 
 	SIMPLE_PAGE_WITH_LONG_CONTENT = `---
@@ -470,6 +480,21 @@ func TestPageWithDate(t *testing.T) {
 	checkPageDate(t, p, d)
 }
 
+func TestRuneCount(t *testing.T) {
+	p, _ := NewPage("simple.md")
+	_, err := p.ReadFrom(strings.NewReader(SIMPLE_PAGE_WITH_FIVE_MULTIBYTE_UFT8_RUNES))
+	p.Convert()
+	p.analyzePage()
+	if err != nil {
+		t.Fatalf("Unable to create a page with frontmatter and body content: %s", err)
+	}
+
+	if p.RuneCount() != 5 {
+		t.Fatalf("incorrect rune count for content '%s'. expected %v, got %v", p.plain, 1, p.RuneCount())
+
+	}
+}
+
 func TestWordCount(t *testing.T) {
 	p, _ := NewPage("simple.md")
 	_, err := p.ReadFrom(strings.NewReader(SIMPLE_PAGE_WITH_LONG_CONTENT))