change: remove unused code

internal/version/VERSION

@@ -1 +1 @@
-v5.0.0-beta-2503192228
+v5.0.0-beta-2503192247

uixt/ai/llm.go

@@ -55,5 +55,4 @@ const (
 	ActionTypeCallUser ActionType = "call_user"
 	ActionTypeType     ActionType = "type"
 	ActionTypeScroll   ActionType = "scroll"
-	ActionTypeHotkey   ActionType = "hotkey"
 )

uixt/ai/parser.go

@@ -110,7 +110,6 @@ func (p *ActionParser) parseActionText(actionText, thought string) ([]ParsedActi
 		"left_double":  regexp.MustCompile(`left_double\(start_box='([^']+)'\)`),
 		"right_single": regexp.MustCompile(`right_single\(start_box='([^']+)'\)`),
 		"drag":         regexp.MustCompile(`drag\(start_box='([^']+)', end_box='([^']+)'\)`),
-		"hotkey":       regexp.MustCompile(`hotkey\(key='([^']+)'\)`),
 		"type":         regexp.MustCompile(`type\(content='([^']+)'\)`),
 		"scroll":       regexp.MustCompile(`scroll\(start_box='([^']+)', direction='([^']+)'\)`),
 		"wait":         regexp.MustCompile(`wait\(\)`),
@@ -156,10 +155,6 @@ func (p *ActionParser) parseActionText(actionText, thought string) ([]ParsedActi
 				}
 				action.ActionInputs["endBox"] = endBox
 			}
-		case ActionTypeHotkey:
-			if len(matches) > 1 {
-				action.ActionInputs["key"] = matches[1]
-			}
 		case ActionTypeType:
 			if len(matches) > 1 {
 				action.ActionInputs["content"] = matches[1]

uixt/ai/planner.go

@@ -167,13 +167,8 @@ func processVLMResponse(actions []ParsedAction) (*PlanningResult, error) {
 		case "drag":
 			validateCoordinateAction(&actions[i], "startBox")
 			validateCoordinateAction(&actions[i], "endBox")
-		case "scroll":
-			validateCoordinateAction(&actions[i], "startBox")
-			validateScrollDirection(&actions[i])
 		case "type":
 			validateTypeContent(&actions[i])
-		case "hotkey":
-			validateHotkeyAction(&actions[i])
 		case "wait", "finished", "call_user":
 			// these actions do not need extra parameters
 		default:
@@ -208,22 +203,12 @@ func extractActionSummary(actions []ParsedAction) string {
 		return "点击操作"
 	case "drag":
 		return "拖拽操作"
-	case "left_double":
-		return "双击操作"
-	case "right_single":
-		return "右键点击操作"
-	case "scroll":
-		direction, _ := action.ActionInputs["direction"].(string)
-		return fmt.Sprintf("滚动操作 (%s)", direction)
 	case "type":
 		content, _ := action.ActionInputs["content"].(string)
 		if len(content) > 20 {
 			content = content[:20] + "..."
 		}
 		return fmt.Sprintf("输入文本: %s", content)
-	case "hotkey":
-		key, _ := action.ActionInputs["key"].(string)
-		return fmt.Sprintf("快捷键: %s", key)
 	case "wait":
 		return "等待操作"
 	case "finished":
@@ -240,22 +225,6 @@ func validateCoordinateAction(action *ParsedAction, boxField string) {
 	// TODO
 }
 
-// validateScrollDirection 验证滚动方向
-func validateScrollDirection(action *ParsedAction) {
-	if direction, ok := action.ActionInputs["direction"].(string); !ok || direction == "" {
-		// default to down
-		action.ActionInputs["direction"] = "down"
-	} else {
-		switch strings.ToLower(direction) {
-		case "up", "down", "left", "right":
-			// keep original direction
-		default:
-			action.ActionInputs["direction"] = "down"
-			log.Warn().Str("direction", direction).Msg("invalid scroll direction, set to default")
-		}
-	}
-}
-
 // validateTypeContent 验证输入文本内容
 func validateTypeContent(action *ParsedAction) {
 	if content, ok := action.ActionInputs["content"]; !ok || content == "" {
@@ -265,15 +234,6 @@ func validateTypeContent(action *ParsedAction) {
 	}
 }
 
-// validateHotkeyAction 验证快捷键动作
-func validateHotkeyAction(action *ParsedAction) {
-	if key, ok := action.ActionInputs["key"]; !ok || key == "" {
-		// 为空或缺失的键设置默认值
-		action.ActionInputs["key"] = "Enter"
-		log.Printf("警告: hotkey动作缺少key参数, 已设置默认值")
-	}
-}
-
 // SavePositionImg saves an image with position markers
 func SavePositionImg(params struct {
 	InputImgBase64 string

uixt/ai/prompt-ui-tars.go

@@ -8,16 +8,13 @@ Thought: ...
 Action: ...
 
 ## Action Space
-click(start_box='[x1, y1, x2, y2]')
+click(start_box='<|box_start|>(x1,y1)<|box_end|>')
-left_double(start_box='[x1, y1, x2, y2]')
+long_press(start_box='<|box_start|>(x1,y1)<|box_end|>', time='')
-right_single(start_box='[x1, y1, x2, y2]')
+type(content='')
-drag(start_box='[x1, y1, x2, y2]', end_box='[x3, y3, x4, y4]')
+drag(start_box='<|box_start|>(x1,y1)<|box_end|>', end_box='<|box_start|>(x3,y3)<|box_end|>')
-hotkey(key='')
+press_home()
-type(content='') #If you want to submit your input, use "\n" at the end of content.
+press_back()
-scroll(start_box='[x1, y1, x2, y2]', direction='down or up or right or left')
+finished(content='') # Submit the task regardless of whether it succeeds or fails.
-wait() #Sleep for 5s and take a screenshot to check for any changes.
-finished()
-call_user() # Submit the task and call the user when the task is unsolvable, or when you need the user's help.
 
 ## Note
 - Use Chinese in Thought part.