fix #3:

- clean mobi page breaks
- reorganize the code a little
- avoid for duplicate sections

pkg/epub/content.go

@@ -8,11 +8,12 @@ import (
 )
 
 var (
-	cleanentitles = regexp.MustCompile(`&[a-z]+;`)
-	empty         = regexp.MustCompile(`(?s)^[\s ]*$`)
-	newlines      = regexp.MustCompile(`[\r\n\s]+`)
-	cleansvg      = regexp.MustCompile(`(<svg.+</svg>|<!\[CDATA\[.+\]\]>)`)
-	cleanmarkup   = regexp.MustCompile(`<[^<>]+>`)
+	cleanenTitles       = regexp.MustCompile(`&[a-z]+;`)
+	isEmpty             = regexp.MustCompile(`(?s)^[\s ]*$`)
+	cleanNewlines       = regexp.MustCompile(`[\r\n\s]+`)
+	cleanSVG            = regexp.MustCompile(`(<svg.+</svg>|<!\[CDATA\[.+\]\]>)`)
+	cleanMarkup         = regexp.MustCompile(`<[^<>]+>`)
+	cleanMobiPageBreaks = regexp.MustCompile(`<mbp:pagebreak/>`)
 )
 
 // Content nav-point content
@@ -25,13 +26,28 @@ type Content struct {
 }
 
 // parse XML, look for title and <p>.*</p> stuff
-func (c *Content) String(content []byte) error {
-	doc, err := xmlquery.Parse(
-		strings.NewReader(
-			cleansvg.ReplaceAllString(
-				cleanentitles.ReplaceAllString(string(content), " "), "")))
+func (c *Content) Extract(content []byte) error {
+	rawXML := cleanSVG.ReplaceAllString(
+		cleanenTitles.ReplaceAllString(string(content), " "), "")
+
+	var doc *xmlquery.Node
+	var err error
+
+	doc, err = xmlquery.Parse(strings.NewReader(rawXML))
 	if err != nil {
-		return err
+		if strings.Contains(err.Error(), `namespace mbp is missing`) {
+			fixedmbp := strings.NewReader(
+				cleanMobiPageBreaks.ReplaceAllString(
+					rawXML, `<span style="page-break-after: always" />`))
+
+			doc, err = xmlquery.Parse(fixedmbp)
+			if err != nil {
+				return err
+			}
+
+		} else {
+			return err
+		}
 	}
 
 	if c.Title == "" {
@@ -47,9 +63,9 @@ func (c *Content) String(content []byte) error {
 	txt := strings.Builder{}
 	var have_p bool
 	for _, item := range xmlquery.Find(doc, "//p") {
-		if !empty.MatchString(item.InnerText()) {
+		if !isEmpty.MatchString(item.InnerText()) {
 			have_p = true
-			txt.WriteString(newlines.ReplaceAllString(item.InnerText(), " ") + "\n\n")
+			txt.WriteString(cleanNewlines.ReplaceAllString(item.InnerText(), " ") + "\n\n")
 		}
 	}
 
@@ -57,9 +73,9 @@ func (c *Content) String(content []byte) error {
 		// try  <div></div>, which some  ebooks use, so get  all divs,
 		// remove markup and paragraphify the parts
 		for _, item := range xmlquery.Find(doc, "//div") {
-			if !empty.MatchString(item.InnerText()) {
-				cleaned := cleanmarkup.ReplaceAllString(item.InnerText(), "")
-				txt.WriteString(newlines.ReplaceAllString(cleaned, " ") + "\n\n")
+			if !isEmpty.MatchString(item.InnerText()) {
+				cleaned := cleanMarkup.ReplaceAllString(item.InnerText(), "")
+				txt.WriteString(cleanNewlines.ReplaceAllString(cleaned, " ") + "\n\n")
 			}
 		}
 	}