Skip to content

Open-source LLM knowledge platform: turn raw documents into a queryable RAG, an autonomous reasoning agent, and a self-maintaining Wiki.

fix(notion): 修复页面清空后旧正文残留

yeliheng-010b9be7d0d0d52
diff --git a/internal/datasource/connector/notion/connector.go b/internal/datasource/connector/notion/connector.go
index cda1f055d..268fae564 100644
--- a/internal/datasource/connector/notion/connector.go
+++ b/internal/datasource/connector/notion/connector.go
@@ -347,10 +347,19 @@ func (c *Connector) fetchPage(ctx context.Context, client *notionClient, page *n
resolveFileUploads(ctx, client, blocks)
markdown, attachmentList := BlocksToMarkdown(blocks)
+ // A successfully cleared page still exists. Keep its title as content so
+ // ingestion replaces stale text instead of silently acknowledging the edit.
+ if len(blocks) == 0 {
+ title := strings.TrimSpace(page.Title)
+ if title == "" {
+ title = defaultUntitledName
+ }
+ markdown = "# " + title + "\n"
+ }
var items []types.FetchedItem
- // Only skip truly empty pages (no content at all)
+ // Do not replace unreadable blocks or attachment-only pages with a title.
if strings.TrimSpace(markdown) != "" {
fileName := page.Title + ".md"
if page.Title == "" {
diff --git a/internal/datasource/connector/notion/empty_page_test.go b/internal/datasource/connector/notion/empty_page_test.go
new file mode 100644
index 000000000..fa74265c2
--- /dev/null
+++ b/internal/datasource/connector/notion/empty_page_test.go
@@ -0,0 +1,108 @@
+package notion
+
+import (
+ "context"
+ "encoding/json"
+ "fmt"
+ "net/http"
+ "net/http/httptest"
+ "testing"
+ "time"
+
+ "github.com/Tencent/WeKnora/internal/types"
+ "github.com/stretchr/testify/require"
+)
+
+func emptyPageConfig(t *testing.T, blocks, title string) *types.DataSourceConfig {
+ t.Helper()
+ allowNotionTestServer(t)
+ titleJSON, err := json.Marshal(title)
+ require.NoError(t, err)
+ page := `{"id":"page","object":"page","url":"https://notion.so/page",` +
+ `"last_edited_time":"2026-09-26T10:00:00Z","parent":{"type":"workspace"},` +
+ `"properties":{"title":{"type":"title","title":[{"plain_text":` + string(titleJSON) + `}]}}}`
+ server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+ w.Header().Set("Content-Type", "application/json")
+ switch r.URL.Path {
+ case "/v1/pages/page":
+ _, _ = fmt.Fprint(w, page)
+ case "/v1/search":
+ _, _ = fmt.Fprint(w, `{"results":[`+page+`],"has_more":false}`)
+ case "/v1/blocks/page/children":
+ if blocks == "later_failure" && r.URL.Query().Get("start_cursor") == "" {
+ _, _ = fmt.Fprint(w, `{"results":[],"has_more":true,"next_cursor":"later"}`)
+ return
+ }
+ if blocks == "unauthorized" || blocks == "later_failure" {
+ w.WriteHeader(http.StatusUnauthorized)
+ return
+ }
+ _, _ = fmt.Fprint(w, `{"results":`+blocks+`,"has_more":false}`)
+ default:
+ http.NotFound(w, r)
+ }
+ }))
+ t.Cleanup(server.Close)
+ return makeNotionConfig(&Config{APIKey: "test-token"}, server.URL, []string{"page"})
+}
+
+func TestClearedPageEmitsReplacement(t *testing.T) {
+ config := emptyPageConfig(t, `[]`, "Retained title")
+ previous := buildCursor(map[string]time.Time{"page": time.Date(2026, 9, 25, 10, 0, 0, 0, time.UTC)})
+ connector := NewConnector()
+ items, next, err := connector.FetchIncremental(context.Background(), config, previous)
+ require.NoError(t, err)
+ require.Len(t, items, 1, "an empty source page must replace previously indexed content")
+ require.Equal(t, "page", items[0].ExternalID)
+ require.Equal(t, "# Retained title\n", string(items[0].Content))
+ require.Equal(t, "Retained title.md", items[0].FileName)
+ require.False(t, items[0].IsDeleted, "the source page still exists")
+ require.Equal(t, "page", items[0].Metadata["object_type"])
+ items, _, err = connector.FetchIncremental(context.Background(), config, next)
+ require.NoError(t, err)
+ require.Empty(t, items, "unchanged empty pages must not be ingested repeatedly")
+}
+
+func TestFullSyncIncludesEmptyPageTitle(t *testing.T) {
+ config := emptyPageConfig(t, `[]`, "Retained title")
+ items, err := NewConnector().FetchAll(context.Background(), config, config.ResourceIDs)
+ require.NoError(t, err)
+ require.Len(t, items, 1)
+ require.Equal(t, "# Retained title\n", string(items[0].Content))
+}
+
+func TestEmptyPageFallbackDoesNotMaskUnreadableContent(t *testing.T) {
+ for _, blocks := range []string{
+ `[{"id":"unknown","type":"unsupported","unsupported":{}}]`, "unauthorized", "later_failure",
+ } {
+ t.Run(blocks, func(t *testing.T) {
+ config := emptyPageConfig(t, blocks, "Retained title")
+ items, _ := NewConnector().FetchAll(context.Background(), config, config.ResourceIDs)
+ require.Empty(t, items, "unreadable content must not be replaced with a title-only document")
+ })
+ }
+}
+
+func TestEmptyPageUsesUntitledFallback(t *testing.T) {
+ for _, title := range []string{"", " "} {
+ config := emptyPageConfig(t, `[]`, title)
+ items, err := NewConnector().FetchAll(context.Background(), config, config.ResourceIDs)
+ require.NoError(t, err)
+ require.Len(t, items, 1)
+ require.Equal(t, "# "+defaultUntitledName+"\n", string(items[0].Content))
+ }
+}
+
+func TestAttachmentPageKeepsRenderedContent(t *testing.T) {
+ blocks := []map[string]interface{}{{"id": "file", "type": "file", "file": map[string]interface{}{
+ "type": "external", "name": "reference.pdf", "external": map[string]string{"url": ""},
+ }}}
+ data, err := json.Marshal(blocks)
+ require.NoError(t, err)
+ config := emptyPageConfig(t, string(data), "Retained title")
+ items, err := NewConnector().FetchAll(context.Background(), config, config.ResourceIDs)
+ require.NoError(t, err)
+ require.Len(t, items, 1)
+ require.Contains(t, string(items[0].Content), "[reference.pdf]")
+ require.NotContains(t, string(items[0].Content), "# Retained title")
+}