fix #3:

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

pkg/epub/open.go

@@ -112,6 +112,8 @@ func (bk *Book) getSections() error {
 
 	// we have ncx points from the TOC, try those
 	if len(bk.Ncx.Points) > 0 {
+		known := map[string]int{}
+
 		for _, block := range bk.Ncx.Points {
 			sect := Section{
 				File:  "OEBPS/" + block.Content.Src,
@@ -128,7 +130,13 @@ func (bk *Book) getSections() error {
 				}
 			}
 
-			sections = append(sections, sect)
+			if _, haveFile := known[sect.File]; !haveFile {
+				// sometimes  epub's have  many sections but  they all
+				// point to the same  file. To avoid duplicate content
+				// we ignore sections (thus files) we have already seen.
+				sections = append(sections, sect)
+				known[sect.File] = 1
+			}
 		}
 
 		if len(sections) < manifestcount {
@@ -189,7 +197,7 @@ func (bk *Book) readSectionContent() error {
 		ct := Content{Src: section.File, Title: section.Title}
 
 		if types.MatchString(section.MediaType) {
-			if err := ct.String(content); err != nil {
+			if err := ct.Extract(content); err != nil {
 				return err
 			}
 		}