add xlsx support

docconv.go

@@ -25,6 +25,8 @@ func MimeTypeByExtension(filename string) string {
 		return "application/msword"
 	case ".docx":
 		return "application/vnd.openxmlformats-officedocument.wordprocessingml.document"
+	case ".xlsx":
+		return "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"
 	case ".odt":
 		return "application/vnd.oasis.opendocument.text"
 	case ".pages":
@@ -69,6 +71,9 @@ func Convert(r io.Reader, mimeType string, readability bool) (*Response, error)
 	case "application/vnd.oasis.opendocument.text":
 		body, meta, err = ConvertODT(r)
 
+	case "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet":
+		body, meta, err = ConvertXLSX(r)
+
 	case "application/vnd.apple.pages", "application/x-iwork-pages-sffpages":
 		body, meta, err = ConvertPages(r)
 

xlsx.go

@@ -0,0 +1,59 @@
+package docconv
+
+import (
+	"fmt"
+	"github.com/tealeg/xlsx"
+	"io"
+)
+
+// Convert MS Excel Spreadsheet
+func ConvertXLSX(r io.Reader) (string, map[string]string, error) {
+	f, err := NewLocalFile(r, "/tmp", "sajari-convert-")
+	if err != nil {
+		return "", nil, fmt.Errorf("error creating local file: %v", err)
+	}
+	defer f.Done()
+
+	fileStat, err := f.Stat()
+	if err != nil {
+		return "", nil, fmt.Errorf("error on getting file stats: %v", err)
+	}
+	xlsFile, err := xlsx.OpenReaderAt(f, fileStat.Size())
+	if err != nil {
+		return "", nil, fmt.Errorf("error on xlsx parsing: %v", err)
+	}
+
+	// Meta data
+	mc := make(chan map[string]string, 1)
+	go func() {
+		meta := make(map[string]string)
+		meta["ModifiedDate"] = fmt.Sprintf("%d", fileStat.ModTime().Unix())
+		mc <- meta
+	}()
+
+	// Document body
+	bc := make(chan string, 1)
+	go func() {
+		var body string
+		for _, sheet := range xlsFile.Sheets {
+			for rowIndex, row := range sheet.Rows {
+				for cellIndex, cell := range row.Cells {
+					text, _ := cell.String()
+					body += text
+					if cellIndex < len(row.Cells)-1 {
+						body += ","
+					}
+				}
+				if rowIndex < len(sheet.Rows)-1 {
+					body += "\n"
+				}
+			}
+		}
+		bc <- body
+	}()
+
+	body := <-bc
+	meta := <-mc
+
+	return body, meta, nil
+}