fix(dingtalk): parse rich text image messages

internal/im/dingtalk/adapter.go

@@ -176,23 +176,14 @@ func parseCallbackMessage(msg *callbackMessage) *im.IncomingMessage {
 		userID = msg.SenderID
 	}
 
-	content := ""
+	text := ""
 	if msg.Text != nil {
-		content = strings.TrimSpace(msg.Text.Content)
+		text = msg.Text.Content
 	}
-	messageType := im.MessageTypeText
-	fileKey := ""
+	messageType, content, fileKey := parseDingTalkStreamPayload(msg.Msgtype, msg.Content, text)
 	fileName := ""
-	if strings.EqualFold(msg.Msgtype, "picture") {
-		messageType = im.MessageTypeImage
+	if messageType == im.MessageTypeImage {
 		fileName = msg.MsgID + ".png"
-		content = ""
-		var picture struct {
-			DownloadCode string `json:"downloadCode"`
-		}
-		if json.Unmarshal(msg.Content, &picture) == nil {
-			fileKey = strings.TrimSpace(picture.DownloadCode)
-		}
 	}
 
 	return &im.IncomingMessage{

internal/im/dingtalk/adapter_test.go

@@ -41,6 +41,28 @@ func TestParseCallbackMessagePicture(t *testing.T) {
 	}
 }
 
+func TestParseCallbackMessageRichTextWithPictureAndText(t *testing.T) {
+	msg := &callbackMessage{
+		MsgID:   "msg-richtext-1",
+		Msgtype: "richText",
+		Content: []byte(`{"richText":[{"text":"  请描述这张图片  "},{"type":"picture","downloadCode":"rich-webhook-code"}]}`),
+	}
+
+	got := parseCallbackMessage(msg)
+	if got.MessageType != im.MessageTypeImage {
+		t.Fatalf("MessageType = %q, want image", got.MessageType)
+	}
+	if got.Content != "请描述这张图片" {
+		t.Fatalf("Content = %q", got.Content)
+	}
+	if got.FileKey != "rich-webhook-code" {
+		t.Fatalf("FileKey = %q", got.FileKey)
+	}
+	if got.FileName != "msg-richtext-1.png" {
+		t.Fatalf("FileName = %q", got.FileName)
+	}
+}
+
 func TestAdapterDownloadFileUsesMessageFilesAPI(t *testing.T) {
 	var sawToken bool
 	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {

internal/im/dingtalk/longconn.go

@@ -102,9 +102,11 @@ func (c *LongConnClient) onChatBotMessage(ctx context.Context, data *chatbot.Bot
 }
 
 func parseDingTalkStreamPayload(msgType string, rawContent interface{}, text string) (im.MessageType, string, string) {
-	if !strings.EqualFold(strings.TrimSpace(msgType), "picture") {
+	normalizedType := strings.ToLower(strings.TrimSpace(msgType))
+	if normalizedType != "picture" && normalizedType != "richtext" {
 		return im.MessageTypeText, strings.TrimSpace(text), ""
 	}
+
 	var payload []byte
 	switch value := rawContent.(type) {
 	case string:
@@ -120,6 +122,41 @@ func parseDingTalkStreamPayload(msgType string, rawContent interface{}, text str
 			return im.MessageTypeImage, "", ""
 		}
 	}
+
+	if normalizedType == "richtext" {
+		var rich struct {
+			Items []struct {
+				Text         string `json:"text"`
+				Type         string `json:"type"`
+				DownloadCode string `json:"downloadCode"`
+			} `json:"richText"`
+		}
+		if json.Unmarshal(payload, &rich) != nil {
+			return im.MessageTypeText, strings.TrimSpace(text), ""
+		}
+
+		var textParts []string
+		fileKey := ""
+		hasPicture := false
+		for _, item := range rich.Items {
+			if itemText := strings.TrimSpace(item.Text); itemText != "" {
+				textParts = append(textParts, itemText)
+			}
+			if strings.EqualFold(strings.TrimSpace(item.Type), "picture") {
+				hasPicture = true
+				if fileKey == "" {
+					fileKey = strings.TrimSpace(item.DownloadCode)
+				}
+			}
+		}
+
+		content := strings.Join(textParts, "\n")
+		if hasPicture {
+			return im.MessageTypeImage, content, fileKey
+		}
+		return im.MessageTypeText, content, ""
+	}
+
 	var picture struct {
 		DownloadCode string `json:"downloadCode"`
 	}

internal/im/dingtalk/longconn_test.go

@@ -35,3 +35,58 @@ func TestParseDingTalkStreamPayloadPictureJSONString(t *testing.T) {
 		t.Fatalf("got type=%q fileKey=%q", messageType, fileKey)
 	}
 }
+
+func TestParseDingTalkStreamPayloadRichTextWithPictureAndText(t *testing.T) {
+	messageType, content, fileKey := parseDingTalkStreamPayload("richText", map[string]interface{}{
+		"richText": []interface{}{
+			map[string]interface{}{"text": "  请回答图片中的问题  "},
+			map[string]interface{}{"type": "picture", "downloadCode": "rich-stream-code"},
+		},
+	}, "")
+
+	if messageType != im.MessageTypeImage {
+		t.Fatalf("MessageType = %q, want image", messageType)
+	}
+	if content != "请回答图片中的问题" {
+		t.Fatalf("content = %q", content)
+	}
+	if fileKey != "rich-stream-code" {
+		t.Fatalf("fileKey = %q", fileKey)
+	}
+}
+
+func TestParseDingTalkStreamPayloadRichTextJoinsTextParts(t *testing.T) {
+	messageType, content, fileKey := parseDingTalkStreamPayload("richText", map[string]interface{}{
+		"richText": []interface{}{
+			map[string]interface{}{"text": "第一段"},
+			map[string]interface{}{"type": "picture", "downloadCode": "rich-code"},
+			map[string]interface{}{"text": "第二段"},
+		},
+	}, "")
+	if messageType != im.MessageTypeImage || content != "第一段\n第二段" || fileKey != "rich-code" {
+		t.Fatalf("got type=%q content=%q fileKey=%q", messageType, content, fileKey)
+	}
+}
+
+func TestParseDingTalkStreamPayloadRichTextWithoutPictureStaysText(t *testing.T) {
+	messageType, content, fileKey := parseDingTalkStreamPayload("richText", map[string]interface{}{
+		"richText": []interface{}{
+			map[string]interface{}{"text": "  只有文字  "},
+		},
+	}, "")
+	if messageType != im.MessageTypeText || content != "只有文字" || fileKey != "" {
+		t.Fatalf("got type=%q content=%q fileKey=%q", messageType, content, fileKey)
+	}
+}
+
+func TestParseDingTalkStreamPayloadRichTextPictureWithoutDownloadCodeStaysImage(t *testing.T) {
+	messageType, content, fileKey := parseDingTalkStreamPayload("richText", map[string]interface{}{
+		"richText": []interface{}{
+			map[string]interface{}{"text": "识别图片"},
+			map[string]interface{}{"type": "picture"},
+		},
+	}, "")
+	if messageType != im.MessageTypeImage || content != "识别图片" || fileKey != "" {
+		t.Fatalf("got type=%q content=%q fileKey=%q", messageType, content, fileKey)
+	}
+}