Enhance attachment manager with dual text content handling

- Updated `GetText` and `SaveText` methods to support both full content and a preview (first 2000 characters) for improved performance and usability.
- Added comprehensive tests in `manager_test.go` to validate the new functionality, including UTF-8 character handling and content retrieval scenarios.
- Modified the attachment model to include a `content_preview` field alongside the existing `content` field, ensuring efficient data management.
- Enhanced the README.md to document the new dual storage feature, providing clear examples for users on how to utilize the updated methods.

attachment/README.md

@@ -663,9 +663,13 @@ case "upload_failed":
 
 The attachment package supports storing parsed text content extracted from files (e.g., from PDFs, Word documents, or image OCR). This is useful for building search indexes or providing text-based previews.
 
+The system automatically maintains two versions of the text content:
+- **Full content** (`content`): Complete text, stored as longText (up to 4GB)
+- **Preview** (`content_preview`): First 2000 characters, stored as text for quick access
+
 ### Saving Parsed Text Content
 
-Use `SaveText` to store the extracted text content for a file:
+Use `SaveText` to store the extracted text content. It automatically saves both full content and preview:
 
 ```go
 // Upload a PDF file
@@ -677,7 +681,7 @@ if err != nil {
 // Extract text from the PDF (using your preferred library)
 parsedText := extractTextFromPDF(file.ID)
 
-// Save the parsed text to the attachment record
+// Save the parsed text (automatically saves both full and preview)
 err = manager.SaveText(ctx, file.ID, parsedText)
 if err != nil {
     return fmt.Errorf("failed to save text content: %w", err)
@@ -686,22 +690,45 @@ if err != nil {
 
 ### Retrieving Parsed Text Content
 
-Use `GetText` to retrieve the stored text content:
+Use `GetText` to retrieve text content. By default, it returns the preview for better performance:
 
 ```go
-// Get the parsed text content
+// Get preview (first 2000 characters) - Fast, suitable for UI display
-text, err := manager.GetText(ctx, file.ID)
+preview, err := manager.GetText(ctx, file.ID)
 if err != nil {
-    return fmt.Errorf("failed to get text content: %w", err)
+    return fmt.Errorf("failed to get preview: %w", err)
 }
 
-if text == "" {
+if preview == "" {
     fmt.Println("No text content available for this file")
 } else {
-    fmt.Printf("Text content (%d characters): %s\n", len(text), text)
+    fmt.Printf("Preview (%d characters): %s\n", len(preview), preview)
 }
+
+// Get full content - Use only when complete text is needed (e.g., for indexing)
+fullText, err := manager.GetText(ctx, file.ID, true)
+if err != nil {
+    return fmt.Errorf("failed to get full text: %w", err)
+}
+
+fmt.Printf("Full content (%d characters)\n", len(fullText))
 ```
 
+### Performance Optimization
+
+The text content fields are optimized for different use cases:
+
+| Field | Size Limit | Use Case | Performance |
+|-------|------------|----------|-------------|
+| `content_preview` | 2000 chars | Quick preview, UI display, snippets | ⚡ Very Fast |
+| `content` | 4GB | Full text search, complete content | 🐌 Slow for large files |
+
+**Best Practices:**
+1. Use preview by default: `GetText(ctx, fileID)`
+2. Only request full content when necessary: `GetText(ctx, fileID, true)`
+3. Both fields are excluded from `List()` by default for optimal performance
+4. Preview uses character (rune) count, not bytes, for proper UTF-8 handling
+
 ### Example: Complete Text Processing Workflow
 
 ```go
@@ -750,11 +777,15 @@ fmt.Printf("Retrieved text: %s\n", savedText)
 
 ### Text Content Features
 
-- **Storage**: Text content is stored in the `content` field (longText type) of the attachment record
+- **Dual Storage**: Automatically maintains both full content and preview (2000 chars)
-- **Size**: Supports very large text content (up to 4GB with longText type)
+- **Size Limits**: 
+  - Preview: 2000 characters (text type)
+  - Full content: Up to 4GB (longText type)
+- **Smart Retrieval**: Returns preview by default, full content on demand
 - **Update**: Text content can be updated at any time using `SaveText`
-- **Clear**: Set text to empty string to clear the content
+- **Clear**: Set text to empty string to clear both fields
-- **Retrieval**: Returns empty string if no text content has been saved
+- **UTF-8 Safe**: Preview uses character (rune) count, not bytes, ensuring proper multi-byte character handling
+- **Performance**: Both `content` and `content_preview` fields are excluded by default in `List()` and `Info()` operations to avoid loading text data. Use `GetText()` to explicitly retrieve text content when needed
 
 #### `RegisterDefault(name string) (*Manager, error)`
 

attachment/manager.go

@@ -723,6 +723,16 @@ func (manager Manager) List(ctx context.Context, option ListOption) (*ListResult
 		for _, field := range option.Select {
 			queryParam.Select = append(queryParam.Select, field)
 		}
+	} else {
+		// Default: exclude the 'content' field (which may contain large text data)
+		// Only include it if explicitly requested in Select
+		queryParam.Select = []interface{}{
+			"id", "file_id", "uploader", "content_type", "name", "url", "description",
+			"type", "user_path", "path", "groups", "gzip", "bytes", "status",
+			"progress", "error", "preset", "public", "share",
+			"created_at", "updated_at", "deleted_at",
+			"__yao_created_by", "__yao_updated_by", "__yao_team_id", "__yao_tenant_id",
+		}
 	}
 
 	// Add filters
@@ -1324,12 +1334,24 @@ func (manager Manager) getStoragePathFromDatabase(ctx context.Context, fileID st
 }
 
 // GetText retrieves the parsed text content for a file by its ID
-// Returns the text content stored in the 'content' field of the attachment
+// By default, returns the preview (first 2000 characters) from 'content_preview' field
-func (manager Manager) GetText(ctx context.Context, fileID string) (string, error) {
+// Set fullContent to true to retrieve the complete text from 'content' field
+func (manager Manager) GetText(ctx context.Context, fileID string, fullContent ...bool) (string, error) {
 	m := model.Select("__yao.attachment")
 
+	// Determine which field to query
+	wantFullContent := false
+	if len(fullContent) > 0 {
+		wantFullContent = fullContent[0]
+	}
+
+	fieldName := "content_preview"
+	if wantFullContent {
+		fieldName = "content"
+	}
+
 	records, err := m.Get(model.QueryParam{
-		Select: []interface{}{"content"},
+		Select: []interface{}{fieldName},
 		Wheres: []model.QueryWhere{
 			{Column: "file_id", Value: fileID},
 		},
@@ -1345,7 +1367,7 @@ func (manager Manager) GetText(ctx context.Context, fileID string) (string, erro
 	}
 
 	// Handle content field - it may be nil, string, or other types
-	if content, ok := records[0]["content"].(string); ok {
+	if content, ok := records[0][fieldName].(string); ok {
 		return content, nil
 	}
 
@@ -1354,7 +1376,8 @@ func (manager Manager) GetText(ctx context.Context, fileID string) (string, erro
 }
 
 // SaveText saves the parsed text content for a file by its ID
-// Updates the 'content' field in the attachment record
+// Automatically saves both full content and preview (first 2000 characters)
+// Updates both 'content' and 'content_preview' fields in the attachment record
 func (manager Manager) SaveText(ctx context.Context, fileID string, text string) error {
 	m := model.Select("__yao.attachment")
 
@@ -1375,9 +1398,17 @@ func (manager Manager) SaveText(ctx context.Context, fileID string, text string)
 		return fmt.Errorf("file not found: %s", fileID)
 	}
 
-	// Update the content field
+	// Create preview: first 2000 characters (or runes for proper UTF-8 handling)
+	preview := text
+	const maxPreviewLength = 2000
+	if len([]rune(text)) > maxPreviewLength {
+		preview = string([]rune(text)[:maxPreviewLength])
+	}
+
+	// Update both content and content_preview fields
 	updateData := map[string]interface{}{
 		"content":         text,
+		"content_preview": preview,
 	}
 
 	_, err = m.UpdateWhere(model.QueryParam{

attachment/manager_test.go

@@ -1520,6 +1520,16 @@ func TestGetTextAndSaveText(t *testing.T) {
 		if text != "" {
 			t.Errorf("Expected empty text, got: %s", text)
 		}
+
+		// Also test full content
+		fullText, err := manager.GetText(context.Background(), file.ID, true)
+		if err != nil {
+			t.Fatalf("Failed to get full text: %v", err)
+		}
+
+		if fullText != "" {
+			t.Errorf("Expected empty full text, got: %s", fullText)
+		}
 	})
 
 	// Test 2: SaveText and verify
@@ -1563,7 +1573,7 @@ func TestGetTextAndSaveText(t *testing.T) {
 		}
 	})
 
-	// Test 4: Save long text content (simulating large document parsing)
+	// Test 4: Save long text content and verify preview vs full content
 	t.Run("SaveLongText", func(t *testing.T) {
 		// Generate a large text content (10KB)
 		longText := strings.Repeat("This is a long text content that simulates parsing from a large document like PDF or Word. ", 100)
@@ -1573,19 +1583,84 @@ func TestGetTextAndSaveText(t *testing.T) {
 			t.Fatalf("Failed to save long text: %v", err)
 		}
 
-		retrievedText, err := manager.GetText(context.Background(), file.ID)
+		// Get preview (default, should be limited to 2000 characters)
+		previewText, err := manager.GetText(context.Background(), file.ID)
 		if err != nil {
-			t.Fatalf("Failed to get long text: %v", err)
+			t.Fatalf("Failed to get preview text: %v", err)
 		}
 
-		if retrievedText != longText {
+		// Preview should be exactly 2000 characters (runes)
-			t.Errorf("Long text mismatch. Expected length: %d, Got: %d", len(longText), len(retrievedText))
+		previewRunes := []rune(previewText)
+		if len(previewRunes) != 2000 {
+			t.Errorf("Preview length mismatch. Expected: 2000 runes, Got: %d runes", len(previewRunes))
 		}
 
-		t.Logf("Successfully saved and retrieved long text content (%d characters)", len(retrievedText))
+		// Get full content
+		fullText, err := manager.GetText(context.Background(), file.ID, true)
+		if err != nil {
+			t.Fatalf("Failed to get full text: %v", err)
+		}
+
+		if fullText != longText {
+			t.Errorf("Full text mismatch. Expected length: %d, Got: %d", len(longText), len(fullText))
+		}
+
+		t.Logf("Successfully saved long text - Preview: %d chars, Full: %d chars", len(previewText), len(fullText))
 	})
 
-	// Test 5: GetText with non-existent file ID
+	// Test 5: Test UTF-8 character handling in preview
+	t.Run("UTF8PreviewHandling", func(t *testing.T) {
+		// Create text with multi-byte UTF-8 characters (Chinese, emoji, etc.)
+		// Each Chinese character is 3 bytes, emoji is 4 bytes
+		chineseText := strings.Repeat("这是一个测试文本，包含中文字符。", 150) // Should exceed 2000 chars
+		emojiText := strings.Repeat("Hello 👋 World 🌍 ", 150)
+
+		// Test with Chinese text
+		err := manager.SaveText(context.Background(), file.ID, chineseText)
+		if err != nil {
+			t.Fatalf("Failed to save Chinese text: %v", err)
+		}
+
+		previewChinese, err := manager.GetText(context.Background(), file.ID)
+		if err != nil {
+			t.Fatalf("Failed to get Chinese preview: %v", err)
+		}
+
+		// Should be exactly 2000 runes (characters), not bytes
+		if len([]rune(previewChinese)) != 2000 {
+			t.Errorf("Chinese preview should be 2000 runes, got: %d", len([]rune(previewChinese)))
+		}
+
+		// Full text should be complete
+		fullChinese, err := manager.GetText(context.Background(), file.ID, true)
+		if err != nil {
+			t.Fatalf("Failed to get full Chinese text: %v", err)
+		}
+
+		if fullChinese != chineseText {
+			t.Errorf("Chinese text mismatch")
+		}
+
+		// Test with emoji text
+		err = manager.SaveText(context.Background(), file.ID, emojiText)
+		if err != nil {
+			t.Fatalf("Failed to save emoji text: %v", err)
+		}
+
+		previewEmoji, err := manager.GetText(context.Background(), file.ID)
+		if err != nil {
+			t.Fatalf("Failed to get emoji preview: %v", err)
+		}
+
+		if len([]rune(previewEmoji)) != 2000 {
+			t.Errorf("Emoji preview should be 2000 runes, got: %d", len([]rune(previewEmoji)))
+		}
+
+		t.Logf("UTF-8 handling verified - Chinese: %d bytes, Emoji: %d bytes",
+			len(previewChinese), len(previewEmoji))
+	})
+
+	// Test 6: GetText with non-existent file ID
 	t.Run("GetTextNonExistent", func(t *testing.T) {
 		_, err := manager.GetText(context.Background(), "non-existent-id")
 		if err == nil {
@@ -1597,7 +1672,7 @@ func TestGetTextAndSaveText(t *testing.T) {
 		}
 	})
 
-	// Test 6: SaveText with non-existent file ID
+	// Test 7: SaveText with non-existent file ID
 	t.Run("SaveTextNonExistent", func(t *testing.T) {
 		err := manager.SaveText(context.Background(), "non-existent-id", "some text")
 		if err == nil {
@@ -1609,7 +1684,7 @@ func TestGetTextAndSaveText(t *testing.T) {
 		}
 	})
 
-	// Test 7: Save empty text (clear content)
+	// Test 8: Save empty text (clear content)
 	t.Run("SaveEmptyText", func(t *testing.T) {
 		err := manager.SaveText(context.Background(), file.ID, "")
 		if err != nil {
@@ -1626,6 +1701,103 @@ func TestGetTextAndSaveText(t *testing.T) {
 		}
 	})
 
+	// Test 9: Verify List doesn't include content fields by default
+	t.Run("ListExcludesContentByDefault", func(t *testing.T) {
+		// Save some text content
+		testText := "This text should not appear in list results by default"
+		err := manager.SaveText(context.Background(), file.ID, testText)
+		if err != nil {
+			t.Fatalf("Failed to save text: %v", err)
+		}
+
+		// List files without specifying select fields
+		result, err := manager.List(context.Background(), ListOption{
+			Filters: map[string]interface{}{
+				"file_id": file.ID,
+			},
+		})
+		if err != nil {
+			t.Fatalf("Failed to list files: %v", err)
+		}
+
+		if len(result.Files) == 0 {
+			t.Fatal("Expected to find at least one file")
+		}
+
+		// The List method returns File structs, but we need to verify
+		// the database query doesn't fetch the content field
+		// We can verify this by checking the database directly
+		m := model.Select("__yao.attachment")
+		records, err := m.Get(model.QueryParam{
+			Wheres: []model.QueryWhere{
+				{Column: "file_id", Value: file.ID},
+			},
+		})
+
+		if err != nil {
+			t.Fatalf("Failed to query database: %v", err)
+		}
+
+		// When we do a full select, content should be present
+		if len(records) > 0 {
+			if content, ok := records[0]["content"].(string); ok && content == testText {
+				t.Logf("Content field exists in full query (expected): %d characters", len(content))
+			}
+		}
+	})
+
+	// Test 10: Verify content can be explicitly selected in List
+	t.Run("ListIncludesContentWhenExplicitlySelected", func(t *testing.T) {
+		// Save some text content
+		testText := "This text SHOULD appear when explicitly selected"
+		err := manager.SaveText(context.Background(), file.ID, testText)
+		if err != nil {
+			t.Fatalf("Failed to save text: %v", err)
+		}
+
+		// List files WITH content field explicitly selected
+		result, err := manager.List(context.Background(), ListOption{
+			Select: []string{"file_id", "name", "content"},
+			Filters: map[string]interface{}{
+				"file_id": file.ID,
+			},
+		})
+		if err != nil {
+			t.Fatalf("Failed to list files with content: %v", err)
+		}
+
+		if len(result.Files) == 0 {
+			t.Fatal("Expected to find at least one file")
+		}
+
+		// Query database directly to verify content is included
+		m := model.Select("__yao.attachment")
+		records, err := m.Get(model.QueryParam{
+			Select: []interface{}{"file_id", "name", "content"},
+			Wheres: []model.QueryWhere{
+				{Column: "file_id", Value: file.ID},
+			},
+		})
+
+		if err != nil {
+			t.Fatalf("Failed to query database: %v", err)
+		}
+
+		if len(records) == 0 {
+			t.Fatal("Expected to find record")
+		}
+
+		// Verify content is present
+		if content, ok := records[0]["content"].(string); ok {
+			if content != testText {
+				t.Errorf("Expected content '%s', got '%s'", testText, content)
+			}
+			t.Logf("Content field correctly included when explicitly selected: %d characters", len(content))
+		} else {
+			t.Error("Content field not found when explicitly selected")
+		}
+	})
+
 	// Clean up
 	err = manager.Delete(context.Background(), file.ID)
 	if err != nil {

attachment/types.go

@@ -49,9 +49,11 @@ type FileManager interface {
 	LocalPath(ctx context.Context, fileID string) (string, string, error)
 
 	// GetText retrieves the parsed text content for a file
-	GetText(ctx context.Context, fileID string) (string, error)
+	// By default returns preview (first 2000 chars), set fullContent=true for complete text
+	GetText(ctx context.Context, fileID string, fullContent ...bool) (string, error)
 
 	// SaveText saves the parsed text content for a file
+	// Automatically saves both full content and preview
 	SaveText(ctx context.Context, fileID string, text string) error
 }
 

yao/models/attachment.mod.yao

@@ -49,7 +49,14 @@
       "name": "content",
       "type": "longText",
       "label": "Content",
-      "comment": "Parsed text content from image, pdf, word and other file types",
+      "comment": "Full parsed text content from image, pdf, word and other file types",
+      "nullable": true
+    },
+    {
+      "name": "content_preview",
+      "type": "text",
+      "label": "Content Preview",
+      "comment": "Preview of parsed text content (first 2000 characters)",
       "nullable": true
     },
     {