新增url解析工具类

zhanghe06 · zhanghe06 · commit 22768d8deb86 · 2015-11-05T11:17:59.000+08:00
diff --git a/test/test_urlparse.py b/test/test_urlparse.py
@@ -0,0 +1,10 @@
+# encoding: utf-8
+__author__ = 'zhanghe'
+
+import sys
+sys.path.append('..')
+from tools.url import get_next_url
+
+print get_next_url('http://www.163.com/mail/index.htm', 'http://www.163.com/about.htm')
+print get_next_url('http://www.163.com/mail/index.htm', '/about.htm')
+print get_next_url('http://www.163.com/mail/index.htm', 'about.htm')
diff --git a/tools/url.py b/tools/url.py
@@ -0,0 +1,33 @@
+# encoding: utf-8
+__author__ = 'zhanghe'
+
+from urlparse import urlparse
+from urlparse import urlunparse
+from urlparse import urljoin
+
+
+def get_next_url(current_url, path):
+    """
+    &#32452;&#35013;url
+    :param current_url:
+    :param path:
+    :return:
+    """
+    if path is None or path == '':
+        return ''
+    if path.startswith('http'):
+        return path
+    if path.startswith('/'):
+        url = urlparse(current_url)
+        return urlunparse((url.scheme, url.netloc, path, url.params, url.query, url.fragment))
+    return urljoin(current_url, path)
+
+
+def test():
+    print get_next_url('http://www.163.com/mail/index.htm', 'http://www.163.com/about.htm')
+    print get_next_url('http://www.163.com/mail/index.htm', '/about.htm')
+    print get_next_url('http://www.163.com/mail/index.htm', 'about.htm')
+
+
+if __name__ == '__main__':
+    test()
