Source Code for Module tlib.base.WptHelper

  1  import time 
  2  from tlib.asset.wpt_batch_lib import wpt_batch_lib 
  3  import xml.etree.ElementTree as elementTree 
  4   
  5  #Constants used for creating the results object 
  6  SUMMARY = 'summary'                 #:Link to the summary report 
  7  RUNS = 'runs'                       #:How many test runs were executed 
  8  HEADERS_DATA = 'headers'            #:Link to raw headers data 
  9  PAGE_DATA = 'pageData'              #:Link to raw page data 
 10  REQUESTS_DATA = 'requestsData'      #:Link to raw requests data 
 11  UTILIZATION_DATA = 'Utilization'    #:Link to raw utilization data 
 12  PAGE_SPEED_DATA = 'pageSpeedData'   #:Link to raw page speed data 
 13  FIRST_VIEW = 'firstView'            #:Results for first view (with empty cache) 
 14  REPEAT_VIEW = 'repeatView'          #:Results for repeat view (with cache populated) 
 15  AVERAGE = 'average'                 #:Response time average 
 16  STD_DEV = 'standardDeviation'       #:Response time standard deviation 
 17  MEDIAN = 'median'                   #:Response time median 
 18  TTFB = 'TTFB'                       #:Time to first byte 
 19  REQUESTS = 'requests'               #:Total number of requests 
 20  REQUESTS_DOC = 'requestsDoc'        #:Number of requests before document is complete 
 21  LOAD_TIME = 'loadTime'              #:Page load time until document is complete 
 22  FULLY_LOADED = 'fullyLoaded'        #Page load time until there is no traffic for 2 seconds 
 23   
 24 -class WptHelper(object): 
 25      """ 
 26      Helper class to send jobs to WebPagetest and get job results 
 27      """ 
 28      _wpt_server = None 
 29      _wpt_params = {} 
 30      _logger = None 
 31   
 32 -    def __init__(self, logger, wpt_server, wpt_params): 
 33          """ 
 34          Constructor for class  
 35   
 36          @param logger: instance of a logging object configured in testing project 
 37          @type logger: logging.Logger 
 38          @param wpt_server: URL of the WebPagetest server used for testing, e.g. http://172.27.108.148:8080 
 39          @type wpt_server: str 
 40          @param wpt_params: Dictionarry with test settings. See https://sites.google.com/a/webpagetest.org/docs/advanced-features/webpagetest-restful-apis for more details 
 41          @type wpt_params: dict 
 42          """ 
 43          logger.debug("Initializing WptOrchestrator") 
 44          logger.debug("WebPagetest server: %s" % wpt_server) 
 45          logger.debug("WebPagetest test parameters: %s" % wpt_params) 
 46   
 47          self._logger = logger 
 48          self._wpt_server = wpt_server 
 49          self._wpt_params = wpt_params 
 50   
 51 -    def test_urls_and_wait(self, requested_urls): 
 52          """ 
 53          Submits a test request for a list of URLs and waits until results are returned 
 54   
 55          @param requested_urls: List of URLs to test 
 56          @type requested_urls: list 
 57          @return: Results as generated by WebPagetest 
 58              For each item in the response, URL is the key and a dom object is the value 
 59          @rtype: dict 
 60          """ 
 61          self._logger.debug("Requested URLs\n%s" % requested_urls) 
 62   
 63          #Submit test request 
 64          id_url_dict = wpt_batch_lib.SubmitBatch(requested_urls, self._wpt_params, self._wpt_server) 
 65          self._logger.debug("SubmitBatch returned\n%s" % id_url_dict) 
 66   
 67          #Check all URLs were submitted to test server 
 68          submitted_urls = set(id_url_dict.values()) 
 69          self._logger.debug("Submitted URLs\n%(submitted_urls)s" % {"submitted_urls": submitted_urls}) 
 70          for url in requested_urls: 
 71              if url not in submitted_urls: 
 72                  self._logger.error('URL submission failed: %(url)s' % {"url": url}) 
 73   
 74          #Wait until all tests are completed 
 75          pending_test_ids = id_url_dict.keys() 
 76          results = [] 
 77          while pending_test_ids: 
 78              id_status_dict = wpt_batch_lib.CheckBatchStatus(pending_test_ids, server_url=self._wpt_server) 
 79   
 80              #Get list of completed tests 
 81              completed_test_ids = [] 
 82              for test_id, test_status in id_status_dict.iteritems(): 
 83                  # We could get 4 different status codes with different meanings as 
 84                  # as follows: 
 85                  # 1XX: Test in progress 
 86                  # 200: Test complete 
 87                  # 4XX: Test request not found 
 88                  if int(test_status) >= 400: 
 89                      self._logger.error('Tests failed with status %(test_id)s: %(test_status)s' % {'test_id':test_id,'test_status':test_status}) 
 90                      pending_test_ids.remove(test_id) 
 91                      continue 
 92   
 93                  if int(test_status) == 200: 
 94                      self._logger.debug("Test %(test_id)s completed, removing from list of pending tests" % {'test_id': test_id}) 
 95                      pending_test_ids.remove(test_id) 
 96                      completed_test_ids.append(test_id) 
 97   
 98                  else: 
 99                      self._logger.debug("Test %(test_id)s has not completed, status=%(test_status)s" % {'test_id':test_id,'test_status':test_status}) 
100   
101              #Get results for completed tests 
102              test_results = wpt_batch_lib.GetXMLResult(completed_test_ids, server_url=self._wpt_server) 
103   
104              #Check we got one result for each of the completed tests 
105              result_test_ids = set(test_results.keys()) 
106              for test_id in completed_test_ids: 
107                  if test_id not in result_test_ids: 
108                      self._logger.error('The XML failed to retrieve: %(test_id)s' % {'test_id': test_id}) 
109   
110              for test_id, dom in test_results.iteritems(): 
111                  self._logger.debug("Test id %(test_id)s result:\n%(xml)s" % {'test_id': test_id, 'xml': dom.toxml()}) 
112                  test_result = self._get_test_results(dom.toxml()) 
113                  results.append(test_result) 
114   
115              if pending_test_ids: 
116                  time.sleep(5) 
117   
118          return results 
119   
120 -    def test_script_and_wait(self, script): 
121          """ 
122          Submits a test request with a script and waits until results are returned 
123   
124          @param script: script 
125          @type script: str 
126          @return: Results as generated by WebPagetest 
127              For each item in the response, URL is the key and a dom object is the value 
128          @rtype: dict 
129          """ 
130          self._logger.debug("Script to execute\n%(script)s" % {"script": script}) 
131          #Create a new instance of _wpt_params to not change the original parameters 
132          test_params = dict(self._wpt_params) 
133          try: 
134              self._wpt_params['script'] = script 
135              result = self.test_urls_and_wait([""]) 
136          finally: 
137              #Restore WPT params 
138              self._wpt_params = dict(test_params) 
139   
140          return result 
141   
142 -    def test_url_and_wait(self, url): 
143          """ 
144          Submits a test request for a URL and waits until results are returned 
145   
146          @param url: URLs to test 
147          @type url: str 
148          @return: Results as generated by WebPagetest 
149              For each item in the response, URL is the key and a dom object is the value 
150          @rtype: dict 
151          """ 
152          self._logger.debug("URL to test\n%(url)s" % {"url": url}) 
153          return self.test_urls_and_wait([url]) 
154   
155 -    def _get_test_results(self, xml_text): 
156          """ 
157          Extracts global test metrics from DOM object returned by WPT and returns them as a dictionary 
158   
159          @param xml_text: XML returned by WPT server 
160          @type xml_text: str 
161          @return: Returns a dictionary with the metrics extracted from XML 
162          @rtype: dict 
163          """ 
164          root = elementTree.fromstring(xml_text) 
165   
166          result = dict() 
167          #Get data node 
168          result[SUMMARY] = root.find("./data/summary").text 
169   
170          #Number of test runs 
171          runs = int(root.find("./data/runs").text) 
172          result[RUNS] = runs 
173   
174          #Get fist view and last view aggregated metrics 
175          #It's possible that RepeatView doesn't exist depending on how test was ran 
176          for view in [FIRST_VIEW, REPEAT_VIEW]: 
177              #Get aggegated metrics 
178              for metric in [AVERAGE, STD_DEV, MEDIAN]: 
179                  view_node = root.find("./data/%(metric)s/%(view)s" % {'metric': metric, 'view': view}) 
180                  #Check there is an element "view", otherwise skip 
181                  if view_node is not None: 
182                      result.setdefault(view, {}).setdefault(LOAD_TIME, {}) 
183                      result[view][LOAD_TIME][metric] = int(view_node.find("./%(LOAD_TIME)s" % {'LOAD_TIME': LOAD_TIME}).text) 
184   
185                      result.setdefault(view, {}).setdefault(TTFB, {}) 
186                      result[view][TTFB][metric] = int(view_node.find("./%(TTFB)s" % {'TTFB': TTFB}).text) 
187   
188                      result.setdefault(view, {}).setdefault(REQUESTS, {}) 
189                      result[view][REQUESTS][metric] = int(view_node.find("./%(REQUESTS)s" % {'REQUESTS': REQUESTS}).text) 
190   
191                      result.setdefault(view, {}).setdefault(REQUESTS_DOC, {}) 
192                      result[view][REQUESTS_DOC][metric] = int(view_node.find("./%(REQUESTS_DOC)s" % {'REQUESTS_DOC': REQUESTS_DOC}).text) 
193   
194                      result.setdefault(view, {}).setdefault(FULLY_LOADED, {}) 
195                      result[view][FULLY_LOADED][metric] = int(view_node.find("./%(FULLY_LOADED)s" % {'FULLY_LOADED': FULLY_LOADED}).text) 
196   
197              #If element "view" doesn't exist, skip getting test runs for that "view" 
198              if view in result: 
199                  #Get results for each run 
200                  run_nodes = root.findall("./data/run") 
201                  for i in range(0, runs): 
202                      #Create missing keys or there will be an exception when assigning values 
203                      result[view].setdefault(RUNS, {}).setdefault(i, {}) 
204   
205                      data_node = run_nodes[i].find("./%(view)s" % {'view': view}) 
206                      #If there is no 'run'.'view' element, it means that test run failed for that view 
207                      if data_node is not None: 
208                          result[view][RUNS][i][HEADERS_DATA] = data_node.find('./rawData/headers').text 
209                          result[view][RUNS][i][PAGE_DATA] = data_node.find('./rawData/pageData').text 
210                          result[view][RUNS][i][REQUESTS_DATA] = data_node.find('./rawData/requestsData').text 
211                          result[view][RUNS][i][UTILIZATION_DATA] = data_node.find('./rawData/utilization').text 
212                          result[view][RUNS][i][PAGE_SPEED_DATA] = data_node.find('./rawData/PageSpeedData').text 
213                      else: 
214                          result[view][RUNS][i][HEADERS_DATA] = '' 
215                          result[view][RUNS][i][PAGE_DATA] = '' 
216                          result[view][RUNS][i][REQUESTS_DATA] = '' 
217                          result[view][RUNS][i][UTILIZATION_DATA] = '' 
218                          result[view][RUNS][i][PAGE_SPEED_DATA] = '' 
219   
220          self._logger.debug('Result:\n%(result)s' % {'result': result}) 
221          return result 
222