Add DOCXFileParserToString

DOCX2Text.go

@@ -4,6 +4,8 @@ Copyright:  2018 Kleissner Investments s.r.o.
 Author:     Peter Kleissner
 
 This code is forked from https://github.com/guylaor/goword and extracts text from DOCX files.
+Project from: https://github.com/IntelligenceX/fileconversion
+Modify by: Vytek (9/3/2023)
 */
 
 package opencrucible
@@ -14,7 +16,6 @@ import (
 	"encoding/xml"
 	"fmt"
 	"io"
-	"io/ioutil"
 	"strings"
 )
 
@@ -115,7 +116,7 @@ func openWordFile(file io.ReaderAt, size int64) (string, error) {
 		}
 		defer rc.Close()
 		if f.Name == "word/document.xml" {
-			doc, err := ioutil.ReadAll(rc)
+			doc, err := io.ReadAll(rc)
 			if err != nil {
 				return "", err
 			}

opencrucible.go

@@ -58,11 +58,13 @@ func DOCXFileParseToString(FileToParse string) (string, error) {
 	// extract text from an XLSX file
 	file, err := os.Open(FileToParse)
 	if err != nil {
-		return "", errors.New(fmt.Sprintf("Error opening file: %s\n", err))
+		return "", fmt.Errorf("error opening file: %s", err)
 	}
 
 	defer file.Close()
 	stat, _ := file.Stat()
+	docx, err := DOCX2Text(file, stat.Size())
+	return docx, err
 }
 
 func ODTParseToString(StreamToParse []byte) (string, error) {

opencrucible_test.go

@@ -51,4 +51,28 @@ func TestRTFarse(t *testing.T) {
 	if got != want {
 		t.Errorf("got %q, wanted %q", got, want)
 	}
+}
+
+func TestDOCXParse(t *testing.T) {
+	got, err := DOCXFileParseToString("test_file/test_file_docx.docx")
+	if err != nil {
+		panic(err)
+	}
+	want := "This is a test file to test library\n"
+	t.Logf("Parsed: %s", got)
+	if got != want {
+		t.Errorf("got %q, wanted %q", got, want)
+	}
+}
+
+func TestDOCXMSParse(t *testing.T) {
+	got, err := DOCXFileParseToString("test_file/test_file_docx_ms.docx")
+	if err != nil {
+		panic(err)
+	}
+	want := "This is a test file to test library\n"
+	t.Logf("Parsed: %s", got)
+	if got != want {
+		t.Errorf("got %q, wanted %q", got, want)
+	}
 }