32-transform-to-and-from-xliff-2

pkg/convert/xliff2.go

@@ -0,0 +1,96 @@
+package convert
+
+import (
+	"encoding/xml"
+	"fmt"
+
+	"go.expect.digital/translate/pkg/model"
+	"golang.org/x/text/language"
+)
+
+// XLIFF 2 Specification: https://docs.oasis-open.org/xliff/xliff-core/v2.0/os/xliff-core-v2.0-os.html
+// XLIFF 2 Example: https://localizely.com/xliff-file/?tab=xliff-20
+
+type xliff2 struct {
+	XMLName xml.Name     `xml:"urn:oasis:names:tc:xliff:document:2.0 xliff"`
+	Version string       `xml:"version,attr"`
+	SrcLang language.Tag `xml:"srcLang,attr"`
+	File    file         `xml:"file"`
+}
+type file struct {
+	Units []unit `xml:"unit"`
+}
+
+type unit struct {
+	ID     string  `xml:"id,attr"`
+	Notes  *[]note `xml:"notes>note"` // Set as pointer to avoid empty <notes></notes> when marshalling.
+	Source string  `xml:"segment>source"`
+}
+
+type note struct {
+	Category string `xml:"category,attr"`
+	Content  string `xml:",chardata"`
+}
+
+// FromXliff2 converts serialized data from the XML data in the XLIFF 2 format into a model.Messages struct.
+func FromXliff2(data []byte) (model.Messages, error) {
+	var xlf xliff2
+	if err := xml.Unmarshal(data, &xlf); err != nil {
+		return model.Messages{}, fmt.Errorf("unmarshal XLIFF 2 formatted XML into xliff2 struct: %w", err)
+	}
+
+	messages := model.Messages{Language: xlf.SrcLang, Messages: make([]model.Message, 0, len(xlf.File.Units))}
+
+	findDescription := func(u unit) string {
+		for _, note := range *u.Notes {
+			if note.Category == "description" {
+				return note.Content
+			}
+		}
+
+		return ""
+	}
+
+	for _, unit := range xlf.File.Units {
+		messages.Messages = append(messages.Messages, model.Message{
+			ID:          unit.ID,
+			Message:     unit.Source,
+			Description: findDescription(unit),
+		})
+	}
+
+	return messages, nil
+}
+
+// ToXliff2 converts a model.Messages struct into a byte slice in the XLIFF 2 format.
+func ToXliff2(messages model.Messages) ([]byte, error) {
+	xlf := xliff2{
+		Version: "2.0",
+		SrcLang: messages.Language,
+		File: file{
+			Units: make([]unit, 0, len(messages.Messages)),
+		},
+	}
+
+	for _, msg := range messages.Messages {
+		var notes *[]note
+		if msg.Description != "" {
+			notes = &[]note{{Category: "description", Content: msg.Description}}
+		}
+
+		xlf.File.Units = append(xlf.File.Units, unit{
+			ID:     msg.ID,
+			Source: msg.Message,
+			Notes:  notes,
+		})
+	}
+
+	data, err := xml.Marshal(&xlf)
+	if err != nil {
+		return nil, fmt.Errorf("marshal xliff2 struct to XLIFF 2 formatted XML: %w", err)
+	}
+
+	dataWithHeader := append([]byte(xml.Header), data...) // prepend generic XML header
+
+	return dataWithHeader, nil
+}

pkg/convert/xliff2_test.go

@@ -0,0 +1,203 @@
+package convert
+
+import (
+	"fmt"
+	"regexp"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"go.expect.digital/translate/pkg/model"
+	"golang.org/x/text/language"
+)
+
+func assertEqualXml(t *testing.T, expected, actual []byte) bool {
+	t.Helper()
+
+	// Matches a substring that starts with > and ends with < with zero or more whitespace in between.
+	re := regexp.MustCompile(`>(\s*)<`)
+	expectedTrimmed := re.ReplaceAllString(string(expected), "><")
+	actualTrimmed := re.ReplaceAllString(string(actual), "><")
+
+	return assert.Equal(t, expectedTrimmed, actualTrimmed)
+}
+
+func TestFromXliff2(t *testing.T) {
+	t.Parallel()
+
+	tests := []struct {
+		name    string
+		wantErr error
+		data    []byte
+		want    model.Messages
+	}{
+		{
+			name: "All OK",
+			data: []byte(`<?xml version="1.0" encoding="UTF-8"?>
+<xliff version="2.0" xmlns="urn:oasis:names:tc:xliff:document:2.0" srcLang="en" trgLang="fr">
+  <file id="ngi18n" original="ng.template">
+    <unit id="common.welcome">
+      <notes>
+        <note category="location">src/app/app.component.html:16</note>
+      </notes>
+      <segment>
+        <source>Welcome!</source>
+        <target>Bienvenue!</target>
+      </segment>
+    </unit>
+    <unit id="common.app.title">
+      <notes>
+        <note category="location">src/app/app.component.html:4</note>
+        <note category="description">App title</note>
+      </notes>
+      <segment>
+        <source>Diary</source>
+        <target>Agenda</target>
+      </segment>
+    </unit>
+  </file>
+</xliff>`),
+			want: model.Messages{
+				Language: language.English,
+				Messages: []model.Message{
+					{
+						ID:      "common.welcome",
+						Message: "Welcome!",
+					},
+					{
+						ID:          "common.app.title",
+						Message:     "Diary",
+						Description: "App title",
+					},
+				},
+			},
+			wantErr: nil,
+		},
+		{
+			name: "Malformed language tag",
+			data: []byte(`<?xml version="1.0" encoding="UTF-8"?>
+<xliff version="2.0" xmlns="urn:oasis:names:tc:xliff:document:2.0" srcLang="xyz-ZY-Latn" trgLang="fr">
+  <file id="ngi18n" original="ng.template">
+    <unit id="common.welcome">
+      <notes>
+        <note category="location">src/app/app.component.html:16</note>
+      </notes>
+      <segment>
+        <source>Welcome!</source>
+        <target>Bienvenue!</target>
+      </segment>
+    </unit>
+    <unit id="common.app.title">
+      <notes>
+        <note category="location">src/app/app.component.html:4</note>
+        <note category="description">App title</note>
+      </notes>
+      <segment>
+        <source>Diary</source>
+        <target>Agenda</target>
+      </segment>
+    </unit>
+  </file>
+</xliff>`),
+			wantErr: fmt.Errorf("language: subtag \"xyz\" is well-formed but unknown"),
+		},
+	}
+	for _, tt := range tests {
+		tt := tt
+		t.Run(tt.name, func(t *testing.T) {
+			t.Parallel()
+
+			result, err := FromXliff2(tt.data)
+			if tt.wantErr != nil {
+				assert.ErrorContains(t, err, tt.wantErr.Error())
+				return
+			}
+
+			if !assert.NoError(t, err) {
+				return
+			}
+
+			assert.Equal(t, tt.want.Language, result.Language)
+			assert.ElementsMatch(t, tt.want.Messages, result.Messages)
+		})
+	}
+}
+
+func Test_ToXliff2(t *testing.T) {
+	t.Parallel()
+
+	tests := []struct {
+		name     string
+		want     []byte
+		wantErr  error
+		messages model.Messages
+	}{
+		{
+			name: "All OK",
+			want: []byte(`<?xml version="1.0" encoding="UTF-8"?>
+<xliff xmlns="urn:oasis:names:tc:xliff:document:2.0" version="2.0" srcLang="en">
+  <file>
+    <unit id="Welcome">
+      <notes>
+        <note category="description">To welcome a new visitor</note>
+      </notes>
+      <segment>
+        <source>Welcome to our website!</source>
+      </segment>
+    </unit>
+    <unit id="Error">
+      <notes>
+        <note category="description">To inform the user of an error</note>
+      </notes>
+      <segment>
+        <source>Something went wrong. Please try again later.</source>
+      </segment>
+    </unit>
+    <unit id="Feedback">
+      <segment>
+        <source>We appreciate your feedback. Thank you for using our service.</source>
+      </segment>
+    </unit>
+  </file>
+</xliff>`),
+			wantErr: nil,
+			messages: model.Messages{
+				Language: language.English,
+				Messages: []model.Message{
+					{
+						ID:          "Welcome",
+						Message:     "Welcome to our website!",
+						Description: "To welcome a new visitor",
+					},
+					{
+						ID:          "Error",
+						Message:     "Something went wrong. Please try again later.",
+						Description: "To inform the user of an error",
+					},
+					{
+						ID:      "Feedback",
+						Message: "We appreciate your feedback. Thank you for using our service.",
+					},
+				},
+			},
+		},
+	}
+	for _, tt := range tests {
+		tt := tt
+		t.Run(tt.name, func(t *testing.T) {
+			t.Parallel()
+
+			result, err := ToXliff2(tt.messages)
+
+			if tt.wantErr != nil {
+				assert.ErrorContains(t, err, tt.wantErr.Error())
+				return
+			}
+
+			if !assert.NoError(t, err) {
+				return
+			}
+
+			assertEqualXml(t, tt.want, result)
+		})
+	}
+}