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/application/service/resumable_upload.go

@@ -161,18 +161,27 @@ func NewResumableUploadService(db *gorm.DB, bases interfaces.KnowledgeBaseServic
 		}
 		return accesspolicy.WithSnapshot(ctx, snapshot), nil
 	}
+	var uploads *resumable.Service
 	ingest := func(ctx context.Context, row *resumable.Upload, path string) (int64, types.JSON, bool, error) {
 		if row.Kind == "digital-human" {
 			source, selection, e := humans.readResumableMaterial(ctx, row.Name, path, row.Size)
 			if e != nil {
 				return 0, nil, false, e
 			}
-			result, e := json.Marshal(uploadResult{Source: &source, Selection: selection})
+			if e = uploads.SaveMaterial(row.ID, source); e != nil {
+				return 0, nil, false, e
+			}
+			// jsonb keeps only small metadata; full text lives in {id}.material.
+			result, e := json.Marshal(uploadResult{
+				Selection: selection,
+				Source:    &types.DigitalHumanSource{Name: source.Name},
+			})
 			return row.Size, result, true, e
 		}
 		return ingestResumableKnowledge(ctx, knowledge, row, path)
 	}
-	uploads, e := resumable.NewService(db, os.Getenv("RESUMABLE_UPLOAD_DIR"), authorize, workerContext, ingest)
+	var e error
+	uploads, e = resumable.NewService(db, os.Getenv("RESUMABLE_UPLOAD_DIR"), authorize, workerContext, ingest)
 	if e != nil {
 		return nil, e
 	}
@@ -192,19 +201,40 @@ func NewResumableUploadService(db *gorm.DB, bases interfaces.KnowledgeBaseServic
 				return nil, digitalHumanInputError("材料尚未解析完成、已过期或不属于此数字人")
 			}
 			var result uploadResult
-			if json.Unmarshal(row.Result, &result) != nil || result.Source == nil {
+			if json.Unmarshal(row.Result, &result) != nil {
 				return nil, digitalHumanInputError("材料解析结果无效")
 			}
 			if result.Selection != "full" {
 				return nil, digitalHumanInputError("材料“" + row.Name + "”是旧版原文节选，请重新上传以完整蒸馏")
 			}
-			sources = append(sources, *result.Source)
+			source, e := resolveDigitalHumanUploadSource(ctx, humans, uploads, row, result)
+			if e != nil {
+				return nil, e
+			}
+			sources = append(sources, source)
 		}
 		return sources, nil
 	}
 	return uploads, nil
 }
 
+// resolveDigitalHumanUploadSource prefers the file-backed material, then legacy
+// jsonb text, then re-parses the raw upload and backfills .material.
+func resolveDigitalHumanUploadSource(ctx context.Context, humans *DigitalHumanService, uploads *resumable.Service, row *resumable.Upload, result uploadResult) (types.DigitalHumanSource, error) {
+	if source, e := uploads.LoadMaterial(row.ID); e == nil {
+		return source, nil
+	}
+	if result.Source != nil && strings.TrimSpace(result.Source.Text) != "" {
+		return *result.Source, nil
+	}
+	source, _, e := humans.readResumableMaterial(ctx, row.Name, uploads.FilePath(row.ID), row.Size)
+	if e != nil {
+		return types.DigitalHumanSource{}, digitalHumanInputError("材料解析结果无效")
+	}
+	_ = uploads.SaveMaterial(row.ID, source)
+	return source, nil
+}
+
 func textUpload(name string) bool {
 	switch strings.ToLower(filepath.Ext(name)) {
 	case ".txt", ".md", ".markdown", ".srt", ".vtt", ".csv", ".json", ".html", ".htm":

internal/application/service/resumable_upload_material_test.go

@@ -0,0 +1,66 @@
+package service
+
+import (
+	"context"
+	"encoding/json"
+	"os"
+	"path/filepath"
+	"testing"
+	"time"
+
+	"github.com/Tencent/WeKnora/internal/resumable"
+	"github.com/Tencent/WeKnora/internal/types"
+	"github.com/stretchr/testify/require"
+	"gorm.io/driver/sqlite"
+	"gorm.io/gorm"
+	"gorm.io/gorm/logger"
+)
+
+func TestResolveDigitalHumanUploadSourcePrefersMaterialThenLegacyThenReparse(t *testing.T) {
+	db, err := gorm.Open(sqlite.Open(filepath.Join(t.TempDir(), "mat.db")), &gorm.Config{Logger: logger.Default.LogMode(logger.Silent)})
+	require.NoError(t, err)
+	sqlDB, err := db.DB()
+	require.NoError(t, err)
+	t.Cleanup(func() { _ = sqlDB.Close() })
+	require.NoError(t, db.AutoMigrate(&resumable.Upload{}))
+	root := t.TempDir()
+	uploads, err := resumable.NewService(db, root, func(context.Context, string, string) error { return nil },
+		func(ctx context.Context, _ *resumable.Upload) (context.Context, error) { return ctx, nil },
+		func(context.Context, *resumable.Upload, string) (int64, types.JSON, bool, error) {
+			return 0, nil, true, nil
+		})
+	require.NoError(t, err)
+	humans := &DigitalHumanService{}
+
+	row := &resumable.Upload{
+		ID: "u1", Name: "note.txt", Size: 11, Kind: "digital-human",
+		State: "completed", ExpiresAt: time.Now().Add(time.Hour),
+		Result: types.JSON(`{"selection":"full","source":{"name":"note.txt"}}`),
+	}
+	require.NoError(t, os.WriteFile(uploads.FilePath(row.ID), []byte("from-upload"), 0600))
+
+	// 1) material file wins
+	require.NoError(t, uploads.SaveMaterial(row.ID, types.DigitalHumanSource{Name: "note.txt", Text: "from-material"}))
+	source, err := resolveDigitalHumanUploadSource(context.Background(), humans, uploads, row, uploadResult{Selection: "full", Source: &types.DigitalHumanSource{Name: "note.txt"}})
+	require.NoError(t, err)
+	require.Equal(t, "from-material", source.Text)
+
+	// 2) legacy jsonb text when material missing
+	require.NoError(t, os.Remove(filepath.Join(root, row.ID+".material")))
+	legacy := uploadResult{Selection: "full", Source: &types.DigitalHumanSource{Name: "note.txt", Text: "from-jsonb"}}
+	source, err = resolveDigitalHumanUploadSource(context.Background(), humans, uploads, row, legacy)
+	require.NoError(t, err)
+	require.Equal(t, "from-jsonb", source.Text)
+
+	// 3) reparse upload and backfill material
+	source, err = resolveDigitalHumanUploadSource(context.Background(), humans, uploads, row, uploadResult{Selection: "full", Source: &types.DigitalHumanSource{Name: "note.txt"}})
+	require.NoError(t, err)
+	require.Equal(t, "from-upload", source.Text)
+	loaded, err := uploads.LoadMaterial(row.ID)
+	require.NoError(t, err)
+	require.Equal(t, "from-upload", loaded.Text)
+
+	meta, err := json.Marshal(uploadResult{Selection: "full", Source: &types.DigitalHumanSource{Name: "note.txt"}})
+	require.NoError(t, err)
+	require.NotContains(t, string(meta), "from-upload")
+}