Tellurium中关于Container的优化

一.概述
在Tellurium中,所有的页面元素和方法都会写在groovy文件中,一个完整的groovy文件可以是这个样子的,以google首页为例:

public class GoogleHomeModule extends DslContext{
    
    public void defineGoogleHomeModule(){    
        ui.Container(uid: "root", clocator: [tag: "center"]){
            InputBox(uid: "input", clocator: [name: "q"])
            SubmitButton(uid: "gSearch", clocator: [name: "btnG"])
            SubmitButton(uid: "lucky", clocator: [name: "btnI"])
            Image(uid: "logo", clocator: [id: "logo"])       
    }
    }
    public void doGoogleSearch(String input){
        pause TelluriumConstant.PAUSELOW
        type "root.input", input
        pause TelluriumConstant.PAUSELOW
        click "root.gSearch"
        waitForPageToLoad TelluriumConstant.WAIT       
    }    
}

其中ui.Container这部分是我们可以做优化的地方

二.优化
首先,直接用Tellurium IDE产生出我们需要的代码,最初工具产生的代码是这样的:

ui.Container(uid: "Table", clocator: [tag: "table"]){
		SubmitButton(uid: "Google", clocator: [tag: "input", type: "submit", class: "lsb", value: "Google 搜索", name: "btnG"])
		Container(uid: "Section", clocator: [tag: "tr"]){
			Container(uid: "Part", clocator: [tag: "td", direct: "true"]){
				InputBox(uid: "Q", clocator: [tag: "input", title: "Google 搜索", class: "lst", name: "q"])
				SubmitButton(uid: "Input", clocator: [tag: "input", type: "submit", class: "lsb", value: " 手气不错 ", name: "btnI"])
					}
				}
			}

可以看到以上的Container中,我们所需要的InputBox在最底层,上面有3个父节点.如果要做一个搜索操作的话,写出来的代码会是这样:

type "Table.Selection.Part.Q", "james"
click "Table.Google"

这样的话,一个操作中对页面元素的依赖性很高,不容易维护.由于Tellurium也是(可以)用Xpath去解析一个元素,所以,比较理想的状况是这样去操作元素:

type "Table.Q", "james"
click "Table.Google"

所以,优化过的Container最好是这个样子:

 ui.Container(uid: "Table", clocator: [tag: "Table"]){
            InputBox(uid: "Q", clocator: [name: "q"])
            SubmitButton(uid: "Google", clocator: [name: "btnG"])
            SubmitButton(uid: "lucky", clocator: [name: "btnI"])
            Image(uid: "logo", clocator: [id: "logo"])       
    }

在以上代码中,将所有中间的Container,多余的元素属性,通通删除掉.让我们对所需要的元素的寻找,仅仅依赖于该元素的name,以及最上层Container的属性(tag,name).

三.总结
跟Selenium一样,我们的目标是要将页面元素的依赖性降低,最好的结果是仅仅依赖于元素的ID或者Name