fix: store digital-human upload text in files, slim jsonb for list

Avoid pulling multi-MB result payloads on uploads?view=summary over remote DB, which caused 30s client timeouts and HTTP 500.

Co-authored-by: Cursor <cursoragent@cursor.com>

internal/resumable/material_test.go

@@ -0,0 +1,75 @@
+package resumable
+
+import (
+	"encoding/json"
+	"os"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/Tencent/WeKnora/internal/types"
+	"github.com/stretchr/testify/require"
+)
+
+func TestMaterialSaveLoadAndCancelRemovesBothFiles(t *testing.T) {
+	s := uploadService(t)
+	ctx := uploadContext(1, "alice")
+	data := []byte("material-body")
+	u := create(t, s, data)
+	_, err := s.Put(ctx, u.ID, 0, int64(len(data)), sum(data), strings.NewReader(string(data)))
+	require.NoError(t, err)
+
+	source := types.DigitalHumanSource{Name: "fixture.txt", Text: strings.Repeat("正文内容-", 200)}
+	require.NoError(t, s.SaveMaterial(u.ID, source))
+	loaded, err := s.LoadMaterial(u.ID)
+	require.NoError(t, err)
+	require.Equal(t, source, loaded)
+	_, err = os.Stat(s.materialPath(u.ID))
+	require.NoError(t, err)
+
+	require.NoError(t, s.Cancel(ctx, u.ID))
+	_, err = os.Stat(s.path(u.ID))
+	require.True(t, os.IsNotExist(err))
+	_, err = os.Stat(s.materialPath(u.ID))
+	require.True(t, os.IsNotExist(err))
+}
+
+func TestListSummaryDoesNotTransferEmbeddedSourceText(t *testing.T) {
+	s := uploadService(t)
+	secret := strings.Repeat("huge-source-payload-", 8000)
+	fat, err := json.Marshal(map[string]any{
+		"selection": "full",
+		"source":    map[string]string{"name": "big.txt", "text": secret},
+	})
+	require.NoError(t, err)
+	require.NoError(t, s.db.Create(&Upload{
+		ID: "fat-1", TenantID: 1, OwnerID: "alice", Kind: "digital-human", TargetID: "dh-1",
+		Name: "big.txt", Size: 10, Offset: 10, IngestOffset: 10, State: "completed",
+		Options: types.JSON(`{}`), Result: fat, ExpiresAt: time.Now().Add(time.Hour),
+	}).Error)
+
+	rows, err := s.List(uploadContext(1, "alice"), "", "", true)
+	require.NoError(t, err)
+	require.Len(t, rows, 1)
+	require.JSONEq(t, `{"selection":"full"}`, string(rows[0].Result))
+	require.NotContains(t, string(rows[0].Result), "huge-source-payload")
+	require.Less(t, len(rows[0].Result), 64)
+
+	full, err := s.List(uploadContext(1, "alice"), "", "", false)
+	require.NoError(t, err)
+	require.Contains(t, string(full[0].Result), secret)
+}
+
+func TestCleanupRemovesMaterialFile(t *testing.T) {
+	s := uploadService(t)
+	ctx := uploadContext(1, "alice")
+	data := []byte("x")
+	u := create(t, s, data)
+	_, err := s.Put(ctx, u.ID, 0, 1, sum(data), strings.NewReader("x"))
+	require.NoError(t, err)
+	require.NoError(t, s.SaveMaterial(u.ID, types.DigitalHumanSource{Name: "x.txt", Text: "hello"}))
+	require.NoError(t, s.db.Model(&Upload{}).Where("id=?", u.ID).Update("expires_at", time.Now().Add(-time.Hour)).Error)
+	require.NoError(t, s.Cleanup(ctx))
+	_, err = os.Stat(s.materialPath(u.ID))
+	require.True(t, os.IsNotExist(err))
+}