更新

macaca 自带了获取控件坐标的功能

element = driver.element_by_xxx(xxx)
rect = element.rect
x = rect['x']
y = rect['y']
width = rect['width']
height = rect['height']

目的

为了不在 case 当中写死坐标, 研究了一下怎样获取控件坐标, 从而通过获取到的坐标实现滑动

用到的模块:

pip install lxml
pip install beautifulsoup4

思路

webdriver 中有一个获取页面 xml 的方法: xml = driver.source 但问题是, 这个 xml 的所有节点名字都是 node, 不太容易通过简单的 xpath 去获取里面的内容. 通过 app-inspector 发现每个节点的名字和 class 是相同的. 因此想到通过 beautifulsoup 把 xml 中的 node 都用 class 的内容替换, 然后在用 lxml 模块去读取其中的 bounds 坐标属性.

代码实现

from bs4 import BeautifulSoup
from lxml import etree

def getElementBounds(xml, xpath):
    '''
    xml: 通过 driver.source 获取
    '''
    soup = BeautifulSoup(xml, 'lxml')
    while soup.node is not None:
        soup.node.name = soup.node['class'][0]

    final_xml = str(soup).replace('$', '')
    # logging.info(final_xml)
    selector = etree.XML(final_xml)
    logging.info(selector.xpath(xpath))
    element = selector.xpath(xpath)[0]
    location_str = element.attrib['bounds'] # [0,1266][1080,1641]
    location_str = location_str[1:-1].replace('][', ',') # 0,1266,1080,1641
    location_list = location_str.split(',') # ['0', '1266', '1080', '1641']
    logging.info(location_list)
    for i in xrange(len(location_list)):
        location_list[i] = int(location_list[i])
    # logging.info(location_list) # [0, 1266, 1080, 1641]
    return location_list

def move(from_xpath, to_xpath):
    xml = driver.source
    from_location = getElementBounds(xml, from_xpath)
    to_location = getElementBounds(xml, to_xpath)
    actions = {}
    actions.update({'fromX': int(from_location[2] / 2), 'fromY': from_location[1]}) # 起点控件的上边缘的中间位置
    actions.update({'toX': int(to_location[2] / 2), 'toY': to_location[3]}) # 终点控件的下边缘的中间位置
    actions.update({'steps': 50})
    driver.touch('drag', actions)

上面代码实现的是从起点控件的上边缘, 移动到终点控件的下边缘, 大家可以根据自己的情况控制起始点位置.