增加了search()和search_all()

DrissionPage/config.py

@@ -10,17 +10,17 @@ Path(global_tmp_path).mkdir(parents=True, exist_ok=True)
 
 global_driver_options = {
     # ---------------已打开的浏览器---------------
-    # 'debuggerAddress': '127.0.0.1:9222',
+    'debuggerAddress': '127.0.0.1:9222',
     # ---------------chromedriver路径---------------
     'chromedriver_path': r'D:\python\Google Chrome\Chrome\chromedriver.exe',
     # ---------------手动指定使用的浏览器位置---------------
     # 'binary_location': r'C:\Program Files (x86)\Google\Chrome\Application\chrome.exe',
     # ---------------启动参数---------------
     'arguments': [
-        '--headless',  # 隐藏浏览器窗口
+        # '--headless',  # 隐藏浏览器窗口
         '--mute-audio',  # 静音
         '--no-sandbox',
-        '--blink-settings=imagesEnabled=false',  # 不加载图片
+        # '--blink-settings=imagesEnabled=false',  # 不加载图片
         # r'--user-data-dir="E:\tmp\chrome_tmp"',  # 指定用户文件夹路径
         # '-–disk-cache-dir=""',  # 指定缓存路径
         'zh_CN.UTF-8',  # 编码格式

DrissionPage/drission.py

@@ -12,7 +12,7 @@ from selenium import webdriver
 from selenium.common.exceptions import WebDriverException
 from selenium.webdriver.chrome.options import Options
 
-from DrissionPage.config import global_driver_options, global_session_options
+from .config import global_driver_options, global_session_options
 
 
 def _get_chrome_options(options: dict) -> Options:

DrissionPage/driver_page.py

@@ -95,7 +95,6 @@ class DriverPage(object):
             elif mode == 'visible':
                 msg = '元素不可见或不存在'
                 ele = wait.until(EC.visibility_of_element_located(loc))
-            # print(loc,ele)
             return ele
         except:
             if show_errmsg:
@@ -105,6 +104,25 @@ class DriverPage(object):
         """查找符合条件的所有元素"""
         return self.find(loc, mode='all', timeout=timeout, show_errmsg=show_errmsg)
 
+    def search(self, value: str, mode: str = None, timeout: float = 10):
+        mode = mode if mode else 'single'
+        if mode not in ['single', 'all']:
+            raise ValueError("mode须在'single', 'all'中选择")
+        ele = []
+        try:
+            loc = 'xpath', f'//*[contains(text(),"{value}")]'
+            wait = WebDriverWait(self.driver, timeout=timeout)
+            if mode == 'single':
+                ele = wait.until(EC.presence_of_element_located(loc))
+            elif mode == 'all':
+                ele = wait.until(EC.presence_of_all_elements_located(loc))
+            return ele
+        except:
+            return ele
+
+    def search_all(self, value: str, timeout: float = 10):
+        return self.search(value, mode='all', timeout=timeout)
+
     def get_attr(self, loc_or_ele: Union[WebElement, tuple], attr: str) -> str:
         """获取元素属性"""
         ele = self._get_ele(loc_or_ele)

DrissionPage/mix_element.py

@@ -0,0 +1,224 @@
+#!/usr/bin/env python
+# -*- coding:utf-8 -*-
+import re
+from html import unescape
+from time import sleep
+from typing import Union
+
+from requests_html import Element
+from selenium.webdriver.remote.webelement import WebElement
+from selenium.webdriver.support import expected_conditions as EC
+from selenium.webdriver.support.wait import WebDriverWait
+
+from .config import global_tmp_path
+from .session_page import _translate_loc
+
+
+class MixElement(object):
+    def __init__(self, ele: Union[WebElement, Element]):
+        self._ele = ele
+
+    @property
+    def ele(self):
+        return self._ele
+
+    @property
+    def text(self):
+        if isinstance(self._ele, Element):
+            return unescape(self._ele.text).replace('\xa0', ' ')
+        else:
+            return unescape(self.attr('innerText')).replace('\xa0', ' ')
+
+    @property
+    def html(self):
+        if isinstance(self._ele, Element):
+            html = unescape(self._ele.html).replace('\xa0', ' ')
+            r = re.match(r'<.*?>(.*)</.*?>', html, flags=re.DOTALL)
+            return r.group(1)
+        else:
+            return unescape(self.attr('innerHTML')).replace('\xa0', ' ')
+
+    @property
+    def tag_name(self):
+        if isinstance(self._ele, Element):
+            html = unescape(self._ele.html).replace('\xa0', ' ')
+            r = re.match(r'^<(.*?)\s+', html, flags=re.DOTALL)
+            return r.group(1)
+        else:
+            return self._ele.tag_name
+
+    def attr(self, attr):
+        if isinstance(self._ele, Element):
+            try:
+                if attr == 'href':
+                    # 如直接获取attr只能获取相对地址
+                    for link in self._ele.absolute_links:
+                        return link
+                elif attr == 'class':
+                    class_str = ''
+                    for key, i in enumerate(self._ele.attrs['class']):
+                        class_str += ' ' if key > 0 else ''
+                        class_str += i
+                    return class_str
+                else:
+                    return self._ele.attrs[attr]
+            except:
+                return ''
+        else:
+            return self._ele.get_attribute(attr)
+
+    def find(self, loc: tuple, mode: str = None, show_errmsg: bool = True):
+        """根据loc获取元素"""
+        if isinstance(self._ele, Element):
+            mode = mode if mode else 'single'
+            if mode not in ['single', 'all']:
+                raise ValueError("mode须在'single', 'all'中选择")
+            loc_by, loc_str = _translate_loc(loc)
+            msg = first = None
+            try:
+                if mode == 'single':
+                    msg = '未找到元素'
+                    first = True
+                elif mode == 'all':
+                    msg = '未找到元素s'
+                    first = False
+                if loc_by == 'xpath':
+                    ele = self.ele.xpath(loc_str, first=first, _encoding='utf-8')
+                else:
+                    ele = self.ele.find(loc_str, first=first, _encoding='utf-8')
+                return MixElement(ele)
+            except:
+                if show_errmsg:
+                    print(msg, loc)
+                    raise
+        else:  # d模式
+            mode = mode if mode else 'single'
+            if mode not in ['single', 'all', 'visible']:
+                raise ValueError("mode须在'single', 'all', 'visible'中选择")
+            msg = ele = None
+            try:
+                wait = WebDriverWait(self.ele.parent, timeout=10)
+                if mode == 'single':
+                    msg = '未找到元素'
+                    ele = wait.until(EC.presence_of_element_located(loc))
+                elif mode == 'all':
+                    msg = '未找到元素s'
+                    ele = wait.until(EC.presence_of_all_elements_located(loc))
+                elif mode == 'visible':
+                    msg = '元素不可见或不存在'
+                    ele = wait.until(EC.visibility_of_element_located(loc))
+                return MixElement(ele)
+            except:
+                if show_errmsg:
+                    print(msg, loc)
+
+    def find_all(self, loc: tuple, show_errmsg: bool = True):
+        return self.find(loc, mode='all', show_errmsg=show_errmsg)
+
+    def search(self, value: str, mode: str = None):
+        """根据内容获取元素"""
+        mode = mode if mode else 'single'
+        if mode not in ['single', 'all']:
+            raise ValueError("mode须在'single', 'all'中选择")
+        if isinstance(self._ele, Element):
+            try:
+                if mode == 'single':
+                    ele = self.ele.xpath(f'.//*[contains(text(),"{value}")]', first=True)
+                    return MixElement(ele)
+                elif mode == 'all':
+                    eles = self.ele.xpath(f'.//*[contains(text(),"{value}")]')
+                    return [MixElement(ele) for ele in eles]
+            except:
+                return None
+        else:  # d模式
+            try:
+                loc = 'xpath', f'.//*[contains(text(),"{value}")]'
+                wait = WebDriverWait(self.ele.parent, timeout=10)
+                if mode == 'single':
+                    ele = wait.until(EC.presence_of_element_located(loc))
+                    return MixElement(ele)
+                elif mode == 'all':
+                    eles = wait.until(EC.presence_of_all_elements_located(loc))
+                    return [MixElement(ele) for ele in eles]
+            except:
+                return None
+
+    def search_all(self, value: str):
+        return self.search(value, mode='all')
+
+    # -----------------以下为d模式独占-------------------
+    def click(self):
+        """点击"""
+        for _ in range(10):
+            try:
+                self.ele.click()
+                return True
+            except Exception as e:
+                print(e)
+                sleep(0.2)
+        # 若点击失败，用js方式点击
+        print('用js点击')
+        try:
+            self.run_script('arguments[0].click()')
+            return True
+        except:
+            raise
+
+    def input(self, value, clear: bool = True):
+        try:
+            if clear:
+                self.run_script("arguments[0].value=''")
+            self.ele.send_keys(value)
+            return True
+        except:
+            raise
+
+    def run_script(self, script: str):
+        self.ele.parent.execute_script(script, self.ele)
+
+    def submit(self):
+        self.ele.submit()
+
+    def clear(self):
+        self.ele.clear()
+
+    def is_selected(self):
+        return self.ele.is_selected()
+
+    def is_enabled(self):
+        return self.ele.is_enabled()
+
+    def is_displayed(self):
+        return self.ele.is_displayed()
+
+    @property
+    def size(self):
+        return self.ele.size
+
+    @property
+    def location(self):
+        return self.ele.location
+
+    def screenshot(self, path: str = None, filename: str = None):
+        path = path if path else global_tmp_path
+        name = filename if filename else self.tag_name
+        # 等待元素加载完成
+        if self.tag_name == 'img':
+            js = 'return arguments[0].complete && typeof arguments[0].naturalWidth ' \
+                 '!= "undefined" && arguments[0].naturalWidth > 0'
+            while not self.run_script(js):
+                pass
+        img_path = f'{path}\\{name}.png'
+        self.ele.screenshot(img_path)
+        return img_path
+
+    def select(self, value: str):
+        pass
+
+    def set_attr(self, attr, value):
+        """设置元素属性"""
+        try:
+            self.run_script(f"arguments[0].{attr} = '{value}';")
+            return True
+        except:
+            raise

DrissionPage/mix_page.py

@@ -12,9 +12,9 @@ from requests_html import Element, HTMLSession
 from selenium.webdriver.chrome.webdriver import WebDriver
 from selenium.webdriver.remote.webelement import WebElement
 
-from DrissionPage.drission import Drission
+from .drission import Drission
-from DrissionPage.driver_page import DriverPage
+from .driver_page import DriverPage
-from DrissionPage.session_page import SessionPage
+from .session_page import SessionPage
 
 
 class Null(object):
@@ -183,6 +183,18 @@ class MixPage(Null, SessionPage, DriverPage):
         elif self._mode == 'd':
             return super(SessionPage, self).find_all(loc, timeout=timeout, show_errmsg=show_errmsg)
 
+    def search(self, value: str, mode: str = None, timeout: float = 10):
+        if self._mode == 's':
+            return super().search(value, mode=mode)
+        elif self._mode == 'd':
+            return super(SessionPage, self).search(value, mode=mode, timeout=timeout)
+
+    def search_all(self, value: str, timeout: float = 10):
+        if self._mode == 's':
+            return super().search_all(value)
+        elif self._mode == 'd':
+            return super(SessionPage, self).search_all(value, timeout=timeout)
+
     def get_attr(self, loc_or_ele: Union[WebElement, Element, tuple], attr: str) -> str:
         """获取元素属性值"""
         if self._mode == 's':

DrissionPage/mix_page_old.py

@@ -1,232 +0,0 @@
-#!/usr/bin/env python
-# -*- coding:utf-8 -*-
-"""
-旧版MixPage，已弃用
-在MixPage类中使用DriverPage和SessionPage对象，使用时根据模式调用相应对象的函数
-问题是须要在MixPage类中为这两个类中的函数写一一对应的调用函数
-新版中直接继承这两个类，只须要为这两个类共有的函数写调用函数即可
-"""
-from abc import abstractmethod
-from typing import Union
-from urllib import parse
-
-from requests_html import Element
-from selenium.webdriver.remote.webelement import WebElement
-
-from DrissionPage.drission import Drission
-from DrissionPage.driver_page import DriverPage
-from DrissionPage.session_page import SessionPage
-
-
-class MixPage:
-    def __init__(self, drission: Drission, locs=None, mode='d'):
-        self._drission = drission
-        self._session = None
-        self._driver = None
-        self._session_page = None
-        self._driver_page = None
-        self._url = None
-        self._session_url = None
-        self._locs = locs
-        self._mode = mode
-        if mode == 's':
-            self._session_page = self.s_page
-        else:
-            self._driver_page = self.d_page
-        self._open_self_url()
-
-    @abstractmethod
-    def _open_self_url(self):
-        pass
-
-    @property
-    def mode(self):
-        return self._mode
-
-    @mode.setter
-    def mode(self, value):
-        self._mode = value
-
-    def change_mode(self, mode=None):
-        if mode == self.mode:
-            return
-        self.mode = 's' if self.mode == 'd' else 'd'
-
-    @property
-    def drission(self):
-        return self._drission
-
-    @property
-    def response(self):
-        return self.s_page.response
-
-    @property
-    def session(self):
-        if self._session is None:
-            self._session = self._drission.session
-        return self._session
-
-    @property
-    def driver(self):
-        if self._driver is None:
-            self._driver = self._drission.driver
-        return self._driver
-
-    @property
-    def d_page(self):
-        if self._driver_page is None:
-            self._driver_page = DriverPage(self.driver)
-            if self._url:
-                self._init_page()
-        return self._driver_page
-
-    @property
-    def s_page(self):
-        if self._session_page is None:
-            self._session_page = SessionPage(self.session)
-            if self._url:
-                self._init_page()
-        self.refresh_url()  # 每次调用session页面时，使url和driver页面保持一致
-        return self._session_page
-
-    @property
-    def url(self):
-        if self.mode == 'd':
-            return self.d_page.url
-        else:
-            return self._url
-
-    def _init_page(self):
-        if self._session_page:
-            self.cookies_to_driver(self._url)
-            self.d_page.get(self._url)
-        elif self._driver_page:
-            self.cookies_to_session()
-            self.s_page.get(self._url)
-
-    def goto(self, url: str, url_data: dict = None):
-        """跳转到一个url"""
-        to_url = f'{url}?{parse.urlencode(url_data)}' if url_data else url
-        if self._url == to_url:
-            return
-        now_url = self._url
-        self._url = to_url
-        if self._driver_page:
-            if self._session_page:
-                self.cookies_to_driver(now_url)
-            self._driver_page.get(to_url, url_data)
-            if not self._session_page:
-                return self.check_driver_url()
-        if self._session_page:
-            self._session_url = to_url
-            if self._session_page:
-                self.cookies_to_session()
-            return self.s_page.goto(to_url, url_data)
-
-    def check_driver_url(self) -> bool:
-        """由子类依据不同的页面自行实现"""
-        return True
-
-    def refresh_url(self):
-        """使session的url与driver当前保持一致，并复制cookies到session"""
-        if self._driver and (self._url != self._driver.current_url or self._session_url != self._driver.current_url):
-            self._url = self._driver.current_url
-            self._session_url = self._driver.current_url
-            self.cookies_to_session()
-            self._session_page.get(self._url)
-
-    def cookies_to_session(self):
-        self._drission.cookies_to_session()
-
-    def cookies_to_driver(self, url=None):
-        u = url if url else self._url
-        self._drission.cookies_to_driver(u)
-
-    # ----------------以下为共用函数-----------------------
-    def find(self, loc, timeout=10, show_errmsg=True) -> Union[WebElement, Element]:
-        if self._mode == 's':
-            return self.s_page.find(loc, show_errmsg)
-        elif self._mode == 'd':
-            return self.d_page.find(loc, timeout, show_errmsg)
-
-    def find_all(self, loc, timeout=10, show_errmsg=True) -> list:
-        if self._mode == 's':
-            return self.s_page.find_all(loc, show_errmsg)
-        elif self._mode == 'd':
-            return self.d_page.find_all(loc, timeout, show_errmsg)
-
-    def get_attr(self, loc_or_ele, attr) -> str:
-        if self._mode == 's':
-            return self.s_page.get_attr(loc_or_ele, attr)
-        elif self._mode == 'd':
-            return self.d_page.get_attr(loc_or_ele, attr)
-
-    def get_html(self, loc_or_ele) -> str:
-        if self._mode == 's':
-            return self.s_page.get_html(loc_or_ele)
-        elif self._mode == 'd':
-            return self.d_page.get_html(loc_or_ele)
-
-    def get_text(self, loc_or_ele) -> str:
-        if self._mode == 's':
-            return self.s_page.get_text(loc_or_ele)
-        elif self._mode == 'd':
-            return self.d_page.get_text(loc_or_ele)
-
-    def get_source(self):
-        if self._mode == 's':
-            return self.s_page.get_html()
-        elif self._mode == 'd':
-            return self.d_page.get_html()
-
-    def get_cookies(self):
-        if self._mode == 's':
-            return self.s_page.cookies
-        elif self._mode == 'd':
-            return self.d_page.cookies
-
-    # ----------------以下为driver page专用函数-----------------
-    def input(self, loc_or_ele, value: str, clear=True) -> bool:
-        return self.d_page.input(loc_or_ele, value, clear)
-
-    def click(self, loc_or_ele) -> bool:
-        return self.d_page.click(loc_or_ele)
-
-    def set_attr(self, loc_or_ele, attribute: str, value: str) -> bool:
-        return self.d_page.set_attr(loc_or_ele, attribute, value)
-
-    def run_script(self, loc_or_ele, script: str):
-        return self.d_page.run_script(loc_or_ele, script)
-
-    def get_tabs_sum(self) -> int:
-        return self.d_page.get_tabs_sum()
-
-    def get_tab_num(self) -> int:
-        return self.d_page.get_tab_num()
-
-    def to_tab(self, index: int = 0):
-        return self.d_page.to_tab(index)
-
-    def close_current_tab(self):
-        return self.d_page.close_current_tab()
-
-    def close_other_tabs(self, tab_index: int = None):
-        return self.d_page.close_other_tabs(tab_index)
-
-    def to_iframe(self, loc_or_ele):
-        return self.d_page.to_iframe(loc_or_ele)
-
-    def get_screen(self, loc_or_ele, path: str, file_name: str = None) -> str:
-        return self.d_page.get_screen(loc_or_ele, path, file_name)
-
-    def choose_select_list(self, loc_or_ele, text):
-        return self.d_page.choose_select_list(loc_or_ele, text)
-
-    def refresh(self):
-        return self.d_page.refresh()
-
-    def back(self):
-        return self.d_page.back()
-
-    def set_window_size(self, x: int = None, y: int = None):
-        return self.d_page.set_window_size(x, y)

DrissionPage/session_page.py

@@ -11,7 +11,7 @@ from urllib import parse
 
 from requests_html import Element, HTMLSession, HTMLResponse
 
-from DrissionPage.config import global_session_options
+from .config import global_session_options
 
 
 def _translate_loc(loc):
@@ -115,6 +115,23 @@ class SessionPage(object):
         """查找符合条件的所有元素"""
         return self.find(loc, mode='all', show_errmsg=True)
 
+    def search(self, value: str, mode: str = None):
+        mode = mode if mode else 'single'
+        if mode not in ['single', 'all']:
+            raise ValueError("mode须在'single', 'all'中选择")
+        try:
+            if mode == 'single':
+                ele = self.response.html.xpath(f'.//*[contains(text(),"{value}")]', first=True)
+                return ele
+            elif mode == 'all':
+                eles = self.response.html.xpath(f'.//*[contains(text(),"{value}")]')
+                return eles
+        except:
+            return None
+
+    def search_all(self, value: str):
+        return self.search(value, mode='all')
+
     def _get_ele(self, loc_or_ele: Union[Element, tuple]) -> Element:
         """获取loc或元素实例，返回元素实例"""
         # ======================================

README.md

@@ -2,36 +2,6 @@
 
 #### 介绍
 一个整合了selenium和requests_html的模块，封装了常用页面操作，可实现两种模式的无缝切换。兼顾selenium的易用性和requests的高性能，也可直接用于PO模式。  
-
+适用于网页自动化，有效减少代码量。
-#### 软件架构
-软件架构说明
 
 
-#### 安装教程
-
-1.  xxxx
-2.  xxxx
-3.  xxxx
-
-#### 使用说明
-
-1.  xxxx
-2.  xxxx
-3.  xxxx
-
-#### 参与贡献
-
-1.  Fork 本仓库
-2.  新建 Feat_xxx 分支
-3.  提交代码
-4.  新建 Pull Request
-
-
-#### 码云特技
-
-1.  使用 Readme\_XXX.md 来支持不同的语言，例如 Readme\_en.md, Readme\_zh.md
-2.  码云官方博客 [blog.gitee.com](https://blog.gitee.com)
-3.  你可以 [https://gitee.com/explore](https://gitee.com/explore) 这个地址来了解码云上的优秀开源项目
-4.  [GVP](https://gitee.com/gvp) 全称是码云最有价值开源项目，是码云综合评定出的优秀开源项目
-5.  码云官方提供的使用手册 [https://gitee.com/help](https://gitee.com/help)
-6.  码云封面人物是一档用来展示码云会员风采的栏目 [https://gitee.com/gitee-stars/](https://gitee.com/gitee-stars/)

setup.py

@@ -0,0 +1,22 @@
+#!/usr/bin/env python
+# -*- coding:utf-8 -*-
+
+from setuptools import setup, find_packages
+
+setup(
+    name="DrissionPage",
+    version="1.0.0 Alpha",
+    author="g1879",
+    author_email="g1879@qq.com",
+    description="page contains selenium and requests.",
+    license="BSD",
+    keywords="DrissionPage",
+    url="https://gitee.com/g1879/DrissionPage",
+    packages=find_packages(),
+    long_description="Long descrition is actually short...",
+    classifiers=[
+        "Development Status :: 1 - Alpha",
+        "Topic :: Utilities",
+        "License :: OSI Approved :: BSD License",
+    ],
+)