Add GetResponses() and related funcionality

tools/chrome_proxy/webdriver/common.py

Index: tools/chrome_proxy/webdriver/common.py
diff --git a/tools/chrome_proxy/webdriver/common.py b/tools/chrome_proxy/webdriver/common.py
index 29c7e58a6ccd1201162126d8092990618ea9b4f8..f508d23c117a1c9e4e280471185089535da09e73 100644
--- a/tools/chrome_proxy/webdriver/common.py
+++ b/tools/chrome_proxy/webdriver/common.py
@@ -5,6 +5,7 @@
 import argparse
 import json
 import os
+import re
 import shlex
 import sys
 import time
@@ -15,13 +16,16 @@ sys.path.append(os.path.join(os.path.dirname(__file__), os.pardir, os.pardir,
 from selenium import webdriver
 from selenium.webdriver.chrome.options import Options
 
-# TODO(robertogden) add logging
+# TODO(robertogden): Add logging.
 
 
 def ParseFlags():
   """
-  Parses the given command line arguments and returns an object with the flags
-  as properties.
+  Parses the given command line arguments.

[sclittle, 2016/12/06 18:11:30]

tiny, tiny nit: Thanks for changing the style here! Could you also move the
one-line summary up a line, so that it's on the same line as the triple quotes?
Same with the other function/class comments.

https://google.github.io/styleguide/pyguide.html?showone=Comments#Comments

[Robert Ogden, 2016/12/06 18:43:00]

Done.

+
+  Returns:
+    A new Namespace object with class properties for each argument added below.
+    See pydoc for argparse.
   """
   parser = argparse.ArgumentParser()
   parser.add_argument('--browser_args', nargs=1, type=str, help='Override '
@@ -34,13 +38,15 @@ def ParseFlags():
   parser.add_argument('chrome_driver', nargs=1, type=str, help='The path to '
     'the ChromeDriver executable. If not given, the default system chrome '
     'will be used.')
-  # TODO(robertogden) make this a logging statement
-  print 'DEBUG: Args=', json.dumps(vars(parser.parse_args(sys.argv[1:])))
+  # TODO(robertogden): Log sys.argv here.
   return parser.parse_args(sys.argv[1:])
 
 def HandleException(test_name=None):
   """
   Writes the exception being handled and a stack trace to stderr.
+
+  Args:
+    test_name: The string name of the test that led to this exception.
   """
   sys.stderr.write("**************************************\n")
   sys.stderr.write("**************************************\n")
@@ -56,10 +62,18 @@ def HandleException(test_name=None):
 
 class TestDriver:
   """
+  The main driver for an integration test.
+
   This class is the tool that is used by every integration test to interact with
   the Chromium browser and validate proper functionality. This class sits on top
   of the Selenium Chrome Webdriver with added utility and helper functions for
   Chrome-Proxy. This class should be used with Python's 'with' operator.
+
+  Attributes:
+    _flags: A Namespace object from the call to ParseFlags()
+    _driver: A reference to the driver object from the Chrome Driver library.
+    _chrome_args: A set of string arguments to start Chrome with.
+    _url: The string URL that Chrome will navigate to for this test.
   """
 
   def __init__(self):
@@ -78,7 +92,9 @@ class TestDriver:
   def _OverrideChromeArgs(self):
     """
     Overrides any given arguments in the code with those given on the command
-    line. Arguments that need to be overridden may contain different values for
+    line.
+
+    Arguments that need to be overridden may contain different values for
     a flag given in the code. In that case, check by the flag whether to
     override the argument.
     """
@@ -89,7 +105,7 @@ class TestDriver:
       original_args = {}
       for arg in self._chrome_args:
           original_args[GetDictKey(arg)] = arg
-      # Override by flag.
+      # Override flags given in code with any command line arguments.
       for override_arg in shlex.split(self._flags.browser_args[0]):
         arg_key = GetDictKey(override_arg)
         if arg_key in original_args:
@@ -119,11 +135,14 @@ class TestDriver:
     Nicely stops the ChromeDriver.
     """
     self._driver.quit()
-    del self._driver
+    self._driver = None
 
   def AddChromeArgs(self, args):
     """
     Adds multiple arguments that will be passed to Chromium at start.
+
+    Args:
+      args: An iterable of strings, each an argument to pass to Chrome at start.
     """
     for arg in args:
       self._chrome_args.add(arg)
@@ -131,6 +150,9 @@ class TestDriver:
   def AddChromeArg(self, arg):
     """
     Adds a single argument that will be passed to Chromium at start.
+
+    Args:
+      arg: a string argument to pass to Chrome at start
     """
     self._chrome_args.add(arg)
 
@@ -138,6 +160,10 @@ class TestDriver:
     """
     Removes multiple arguments that will no longer be passed to Chromium at
     start.
+
+    Args:
+        args: An iterable of strings to no longer use the next time Chrome
+          starts.
     """
     for arg in args:
       self._chrome_args.discard(arg)
@@ -146,6 +172,9 @@ class TestDriver:
     """
     Removes a single argument that will no longer be passed to Chromium at
     start.
+
+    Args:
+      arg: A string flag to no longer use the next time Chrome starts.
     """
     self._chrome_args.discard(arg)
 
@@ -157,7 +186,9 @@ class TestDriver:
 
   def ClearCache(self):
     """
-    Clears the browser cache. Important note: ChromeDriver automatically starts
+    Clears the browser cache.
+
+    Important note: ChromeDriver automatically starts
     a clean copy of Chrome on every instantiation.
     """
     self.ExecuteJavascript('if(window.chrome && chrome.benchmarking && '
@@ -168,33 +199,164 @@ class TestDriver:
   def SetURL(self, url):
     """
     Sets the URL that the browser will navigate to during the test.
+
+    Args:
+      url: The string URL to navigate to
     """
     self._url = url
 
-  # TODO(robertogden) add timeout
+  # TODO(robertogden): Add timeout.
   def LoadPage(self):
     """
     Starts Chromium with any arguments previously given and navigates to the
-    previously given URL.
+    given URL.
     """
     if not self._driver:
       self._StartDriver()
     self._driver.get(self._url)
 
-  # TODO(robertogden) add timeout
+  # TODO(robertogden): Add timeout.
   def ExecuteJavascript(self, script):
     """
     Executes the given javascript in the browser's current page as if it were on
-    the console. Returns a string of whatever the evaluation was.
+    the console.
+
+    Args:
+      script: A string of Javascript code.
+    Returns:
+      A string of the verbatim output from the Javascript execution.
     """
     if not self._driver:
       self._StartDriver()
     return self._driver.execute_script("return " + script)
 
+  def GetPerformanceLogs(self, method_filter=r'Network\.responseReceived'):
+    """
+    Returns all logged Performance events from Chrome.
+
+    Args:
+      method_filter: A regex expression to match the method of logged events
+        against. Only logs who's method matches the regex will be returned.
+    Returns:
+      Performance logs as a list of dicts, since the last time this function was
+      called.
+    """
+    all_messages = []
+    for log in self._driver.execute('getLog', {'type': 'performance'})['value']:
+      message = json.loads(log['message'])['message']
+      if re.match(method_filter, message['method']):
+        all_messages.append(message)
+    return all_messages
+
+  def GetHTTPResponses(self, include_favicon=False):
+    """
+    Parses the Performance Logs and returns a list of HTTPResponse objects.
+
+    This function should be called exactly once after every page load.
+
+    Args:
+      include_favicon: A bool that if True will include responses for favicons.
+    Returns:
+      A list of HTTPResponse objects, each representing a single completed HTTP
+      transaction by Chrome.
+    """
+    def MakeHTTPResponse(log_dict):
+      params = log_dict['params']
+      response_dict = params['response']
+      http_response_dict = {
+        'response_headers': response_dict['headers'],
+        'request_headers': response_dict['requestHeaders'],
+        'url': response_dict['url'],
+        'protocol': response_dict['requestHeaders'][':scheme'],
+        'status': response_dict['status'],
+        'request_type': params['type']
+      }
+      return HTTPResponse(**http_response_dict)
+    all_responses = []
+    for message in self.GetPerformanceLogs():
+      response = MakeHTTPResponse(message)
+      is_favicon = response.url.endswith('favicon.ico')
+      if not is_favicon or include_favicon:
+        all_responses.append(response)
+    return all_responses
+
+class HTTPResponse:
+  """
+  This class represents a single HTTP transaction (request and response) by
+  Chrome.
+
+  This class also includes several convenience functions for ChromeProxy
+  specific assertions.
+
+  Attributes:
+    _response_headers: A dict of response headers.
+    _request_headers: A dict of request headers.
+    _url: the fetched url
+    _protocol: The protocol used for the request. (HTTP, HTTPS, etc)
+    _status: The integer status code of the response
+    _request_type: What caused this request (Document, XHR, etc)
+    _flags: A Namespace object from ParseFlags()
+  """
+
+  def __init__(self, response_headers, request_headers, url, protocol,
+      status, request_type):
+    self._response_headers = response_headers
+    self._request_headers = request_headers
+    self._url = url
+    self._protocol = protocol
+    self._status = status
+    self._request_type = request_type
+    self._flags = ParseFlags()
+
+  def __str__(self):
+    self_dict = {
+      'response_headers': self._response_headers,
+      'request_headers': self._request_headers,
+      'url': self._url,
+      'protocol': self._protocol,
+      'status': self._status,
+      'request_type': self._request_type
+    }
+    return json.dumps(self_dict)
+
+  @property
+  def response_headers(self):
+    return self._response_headers
+
+  @property
+  def request_headers(self):
+    return self._request_headers
+
+  @property
+  def url(self):
+    return self._url
+
+  @property
+  def protocol(self):
+    return self._protocol
+
+  @property
+  def status(self):
+    return self._status
+
+  @property
+  def request_type(self):
+    return self._request_type
+
+  def UsedHTTPS(self):
+    return self._protocol == 'https'

[sclittle, 2016/12/06 18:11:30]

As per offline discussion, it looks like this returns the protocol of the page,
not the protocol that the proxy used.

I'd recommend removing this for now and adding support for determining secure
proxy vs. insecure proxy in a later CL.

[Robert Ogden, 2016/12/06 18:43:00]

Done.

+
+  def ResponseHasViaHeader(self):
+    return 'via' in self._response_headers and (self._response_headers['via'] ==
+      self._flags.via_header_value)
+
+  def WasXHR(self):
+    return self.request_type == 'XHR'
 
 class IntegrationTest:
   """
   A parent class for all integration tests with utility and assertion methods.
+
   All methods starting with the word 'test' (ignoring case) will be called with
   the RunAllTests() method which can be used in place of a main method.
   """
@@ -209,12 +371,18 @@ class IntegrationTest:
       try:
         getattr(self, method)()
       except Exception as e:
-        # Uses the Exception tuple from sys.exec_info()
+        # Uses the Exception tuple from sys.exec_info().
         HandleException(method)
 
-  # TODO(robertogden) add some nice assertion functions
+  # TODO(robertogden): Add some nice assertion functions.
 
   def Fail(self, msg):
+    """
+    Called when a test fails an assertion.
+
+    Args:
+      msg: The string message to print to stderr
+    """
     sys.stderr.write("**************************************\n")
     sys.stderr.write("**************************************\n")
     sys.stderr.write("**                                  **\n")