fanyi 01-12

百度翻译/baidufanyi.py

-#coding:utf-8
+#coding:utf-8
@@ -35,31 +35,27 @@ class Translate():
     def close(self):
         self.browser.quit()
 
+    def is_website_link(self,string):
+        pattern = r"^(http|https)?(://)?[a-zA-Z0-9-]+\.[a-zA-Z0-9-.]+(/[a-zA-Z0-9-_.?=/]*)?$"
+        if re.match(pattern, string):
+            return True
+        else:
+            return False
+
     def createDriver(self):
-        # chrome_driver = r'F:\spider\117\chromedriver-win64\chromedriver.exe'
-        # path = Service(chrome_driver)
-        # chrome_options = webdriver.ChromeOptions()
-        # chrome_options.add_argument('--disable-gpu')
-        # chrome_options.add_argument('--ignore-certificate-errors')
-        # chrome_options.add_argument("--disable-blink-features=AutomationControlled")
-        # chrome_options.add_argument("--start-maximized")
         proxy_ = baseCore.get_proxy()
-        # chrome_options.add_argument('--proxy-server=' + proxy['http'].split('://')[1])
-        # chrome_options.add_argument(
-        #     'user-agent=' + 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36')
-        #
-        # browser = webdriver.Chrome(service=path, chrome_options=chrome_options)
-        proxy= Proxy()
-        proxy.proxy_type = ProxyType.MANUAL
-        proxy.http_proxy = proxy_['http'].split('://')[1]
-        proxy.ssl_proxy = proxy_['http'].split('://')[1]
+        profile = webdriver.FirefoxProfile()
+        profile.set_preference('network.proxy.type',1)
+        profile.set_preference('network.proxy.http',proxy_['http'].split('://')[1].split(':')[0])
+        profile.set_preference('network.proxy.http_port',int(proxy_['http'].split('://')[1].split(':')[1]))
+        profile.set_preference('network.proxy.ssl',proxy_['http'].split('://')[1].split(':')[0])
+        profile.set_preference('network.proxy.ssl_port',int(proxy_['http'].split('://')[1].split(':')[1]))
+        profile.update_preferences()
         service = Service(r'F:\spider\firefox\geckodriver_1.exe')
         options = Options()
         options.set_preference("general.useragent.override",
                                "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/58.0.3029.110 Safari/537.3")
-        capabilities = webdriver.DesiredCapabilities.FIREFOX
-        proxy.add_to_capabilities(capabilities)
-        browser = webdriver.Firefox(options=options, service=service,capabilities=capabilities)
+        browser = webdriver.Firefox(firefox_profile=profile, service=service,options=options)
         return browser
 
     def translate(self, sentence, lang):
@@ -158,6 +154,8 @@ class Translate():
         for tag in tag_list:
             if tag.strip() == '':
                 continue
+            if self.is_website_link(str(tag).strip()):
+                continue
             sentence += f'{tag}😊'
             num += 1
         #print(num)
@@ -170,6 +168,9 @@ class Translate():
         #print(sentence)
         result = ''
         while True:
+            if len(sentence.strip()) == 1 and self.is_punctuation(sentence.strip()):
+                result += sentence
+                break
             if len(sentence) > 1000:
                 index_1000 = sentence[999]
                 # 判断该字符是不是逗号或句号