Add RuneCount to Page

Fixes #1266

diff --git a/docs/content/templates/variables.md b/docs/content/templates/variables.md
index cfe010e36755454b4b1a74aef64a466d846e8bb3..ca30174e2f2f49b3efa1b723dff3e117aafdb944 100644 (file)
--- a/docs/content/templates/variables.md
+++ b/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>

diff --git a/helpers/general.go b/helpers/general.go
index 540fb43328a847e47cc085d33372d67114ef86c0..c6adc855aaac1749ba80349f44e9dbed4ecf62e2 100644 (file)
--- a/helpers/general.go
+++ b/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:

diff --git a/hugolib/page.go b/hugolib/page.go
index d8d8c00e6cfae8f6aa8ead3ae8a77b375532d66a..1fa010abef9d918a88fdacbb1479623ff8377ab5 100644 (file)
--- a/hugolib/page.go
+++ b/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
        })
 }
 

diff --git a/hugolib/page_test.go b/hugolib/page_test.go
index 2bf078f1872348d90f4027d87ddb80e4e272e594..cdb38651f93d6c4ae310bd833dfa85989ac25f8d 100644 (file)
--- a/hugolib/page_test.go
+++ b/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))