djmoffat@718: #!/usr/bin/python djmoffat@718: # -*- coding: utf-8 -*- djmoffat@718: djmoffat@718: import xml.etree.ElementTree as ET djmoffat@718: import os # for getting files from directory djmoffat@718: import operator # for sorting data with multiple keys djmoffat@718: import sys # for accessing command line arguments djmoffat@718: djmoffat@718: # Command line arguments djmoffat@718: assert len(sys.argv)<3, "evaluation_stats takes at most 1 command line argument\n"+\ djmoffat@718: "Use: python evaluation_stats.py [results_folder]" djmoffat@718: djmoffat@718: # XML results files location djmoffat@718: if len(sys.argv) == 1: djmoffat@718: folder_name = "../saves" # Looks in 'saves/' folder from 'scripts/' folder djmoffat@718: print "Use: python evaluation_stats.py [results_folder]" djmoffat@718: print "Using default path: " + folder_name djmoffat@718: elif len(sys.argv) == 2: djmoffat@718: folder_name = sys.argv[1] # First command line argument is folder djmoffat@718: djmoffat@718: # Turn number of seconds (int) to '[minutes] min [seconds] s' (string) djmoffat@718: def seconds2timestr(time_in_seconds): djmoffat@718: time_in_minutes = int(time_in_seconds/60) djmoffat@718: remaining_seconds = int(time_in_seconds%60) djmoffat@718: return str(time_in_minutes) + " min " + str(remaining_seconds) + " s" djmoffat@718: djmoffat@718: # stats initialisation djmoffat@718: number_of_XML_files = 0 djmoffat@718: number_of_pages = 0 djmoffat@718: number_of_fragments = 0 djmoffat@718: total_empty_comments = 0 djmoffat@718: total_not_played = 0 djmoffat@718: total_not_moved = 0 djmoffat@718: time_per_page_accum = 0 djmoffat@718: djmoffat@718: # arrays initialisation djmoffat@718: page_names = [] djmoffat@718: page_count = [] djmoffat@718: duration_page = [] # duration of experiment in function of page content djmoffat@718: duration_order = [] # duration of experiment in function of page number djmoffat@718: fragments_per_page = [] # number of fragments for corresponding page djmoffat@718: djmoffat@718: # get every XML file in folder djmoffat@718: files_list = os.listdir(folder_name) djmoffat@718: for file in files_list: # iterate over all files in files_list djmoffat@718: if file.endswith(".xml"): # check if XML file djmoffat@718: number_of_XML_files += 1 djmoffat@718: tree = ET.parse(folder_name + '/' + file) djmoffat@718: root = tree.getroot() djmoffat@718: djmoffat@718: print file # print file name (subject name) djmoffat@718: djmoffat@718: # reset for new subject djmoffat@718: total_duration = 0 djmoffat@718: page_number = 0 djmoffat@718: djmoffat@718: # get list of all page names djmoffat@718: for audioholder in root.findall("./audioholder"): # iterate over pages djmoffat@718: page_name = audioholder.get('id') # get page name djmoffat@718: djmoffat@718: if page_name is None: # ignore 'empty' audio_holders djmoffat@718: print "WARNING: " + file + " contains empty audio holder. (evaluation_stats.py)" djmoffat@718: break # move on to next djmoffat@718: djmoffat@718: number_of_comments = 0 # for this page djmoffat@718: number_of_missing_comments = 0 # for this page djmoffat@718: not_played = 0 # for this page djmoffat@718: not_moved = 0 # for this page djmoffat@718: djmoffat@718: # 'testTime' keeps total duration: subtract time so far for duration of this audioholder djmoffat@718: duration = float(audioholder.find("./metric/metricresult[@id='testTime']").text) - total_duration djmoffat@718: djmoffat@718: # total duration of test djmoffat@718: total_duration += duration djmoffat@718: djmoffat@718: # number of audio elements djmoffat@718: audioelements = audioholder.findall("./audioelement") # get audioelements djmoffat@718: number_of_fragments += len(audioelements) # add length of this list to total djmoffat@718: djmoffat@718: # number of comments (interesting if comments not mandatory) djmoffat@718: for audioelement in audioelements: djmoffat@718: response = audioelement.find("./comment/response") djmoffat@718: was_played = audioelement.find("./metric/metricresult/[@name='elementFlagListenedTo']") djmoffat@718: was_moved = audioelement.find("./metric/metricresult/[@name='elementFlagMoved']") djmoffat@718: if response.text is not None and len(response.text) > 1: djmoffat@718: number_of_comments += 1 djmoffat@718: else: djmoffat@718: number_of_missing_comments += 1 djmoffat@718: if was_played is not None and was_played.text == 'false': djmoffat@718: not_played += 1 djmoffat@718: if was_moved is not None and was_moved.text == 'false': djmoffat@718: not_moved += 1 djmoffat@718: djmoffat@718: # update global counters djmoffat@718: total_empty_comments += number_of_missing_comments djmoffat@718: total_not_played += not_played djmoffat@718: total_not_moved += not_moved djmoffat@718: djmoffat@718: # print audioholder id and duration djmoffat@718: print " " + page_name + ": " + seconds2timestr(duration) + ", "\ djmoffat@718: + str(number_of_comments)+"/"\ djmoffat@718: +str(number_of_comments+number_of_missing_comments)+" comments" djmoffat@718: djmoffat@718: # number of audio elements not played djmoffat@718: if not_played > 1: djmoffat@718: print 'ATTENTION: '+str(not_played)+' fragments were not listened to!' djmoffat@718: if not_played == 1: djmoffat@718: print 'ATTENTION: one fragment was not listened to!' djmoffat@718: djmoffat@718: # number of audio element markers not moved djmoffat@718: if not_moved > 1: djmoffat@718: print 'ATTENTION: '+str(not_moved)+' markers were not moved!' djmoffat@718: if not_moved == 1: djmoffat@718: print 'ATTENTION: one marker was not moved!' djmoffat@718: djmoffat@718: # keep track of duration in function of page index djmoffat@718: if len(duration_order)>page_number: djmoffat@718: duration_order[page_number].append(duration) djmoffat@718: else: djmoffat@718: duration_order.append([duration]) djmoffat@718: djmoffat@718: # keep list of audioholder ids and count how many times each audioholder id djmoffat@718: # was tested, how long it took, and how many fragments there were (if number of djmoffat@718: # fragments is different, store as different audioholder id) djmoffat@718: if page_name in page_names: djmoffat@718: page_index = page_names.index(page_name) # get index djmoffat@718: # check if number of audioelements the same djmoffat@718: if len(audioelements) == fragments_per_page[page_index]: djmoffat@718: page_count[page_index] += 1 djmoffat@718: duration_page[page_index].append(duration) djmoffat@718: else: # make new entry djmoffat@718: alt_page_name = page_name+"("+str(len(audioelements))+")" djmoffat@718: if alt_page_name in page_names: # if already there djmoffat@718: alt_page_index = page_names.index(alt_page_name) # get index djmoffat@718: page_count[alt_page_index] += 1 djmoffat@718: duration_page[alt_page_index].append(duration) djmoffat@718: else: djmoffat@718: page_names.append(alt_page_name) djmoffat@718: page_count.append(1) djmoffat@718: duration_page.append([duration]) djmoffat@718: fragments_per_page.append(len(audioelements)) djmoffat@718: else: djmoffat@718: page_names.append(page_name) djmoffat@718: page_count.append(1) djmoffat@718: duration_page.append([duration]) djmoffat@718: fragments_per_page.append(len(audioelements)) djmoffat@718: djmoffat@718: # bookkeeping djmoffat@718: page_number += 1 # increase page count for this specific test djmoffat@718: number_of_pages += 1 # increase total number of pages djmoffat@718: time_per_page_accum += duration # total duration (for average time spent per page) djmoffat@718: djmoffat@718: # print total duration of this test djmoffat@718: print " TOTAL: " + seconds2timestr(total_duration) djmoffat@718: djmoffat@718: djmoffat@718: # PRINT EVERYTHING djmoffat@718: djmoffat@718: print "Number of XML files: " + str(number_of_XML_files) djmoffat@718: print "Number of pages: " + str(number_of_pages) djmoffat@718: print "Number of fragments: " + str(number_of_fragments) djmoffat@718: print "Number of empty comments: " + str(total_empty_comments) +\ djmoffat@718: " (" + str(round(100.0*total_empty_comments/number_of_fragments,2)) + "%)" djmoffat@718: print "Number of unplayed fragments: " + str(total_not_played) +\ djmoffat@718: " (" + str(round(100.0*total_not_played/number_of_fragments,2)) + "%)" djmoffat@718: print "Number of unmoved markers: " + str(total_not_moved) +\ djmoffat@718: " (" + str(round(100.0*total_not_moved/number_of_fragments,2)) + "%)" djmoffat@718: print "Average time per page: " + seconds2timestr(time_per_page_accum/number_of_pages) djmoffat@718: djmoffat@718: # Pages and number of times tested djmoffat@718: page_count_strings = list(str(x) for x in page_count) djmoffat@718: count_list = page_names + page_count_strings djmoffat@718: count_list[::2] = page_names djmoffat@718: count_list[1::2] = page_count_strings djmoffat@718: print "Pages tested: " + str(count_list) djmoffat@718: djmoffat@718: # Average duration for first, second, ... page djmoffat@718: print "Average duration per page:" djmoffat@718: for page_number in range(len(duration_order)): djmoffat@718: print " page " + str(page_number+1) + ": " +\ djmoffat@718: seconds2timestr(sum(duration_order[page_number])/len(duration_order[page_number])) +\ djmoffat@718: " ("+str(len(duration_order[page_number]))+" subjects)" djmoffat@718: djmoffat@718: djmoffat@718: # Sort pages by number of audioelements, then by duration djmoffat@718: djmoffat@718: # average duration and number of subjects per page djmoffat@718: average_duration_page = [] djmoffat@718: number_of_subjects_page = [] djmoffat@718: for line in duration_page: djmoffat@718: number_of_subjects_page.append(len(line)) djmoffat@718: average_duration_page.append(sum(line)/len(line)) djmoffat@718: djmoffat@718: # combine and sort in function of number of audioelements and duration djmoffat@718: combined_list = [page_names, average_duration_page, fragments_per_page, number_of_subjects_page] djmoffat@718: combined_list = sorted(zip(*combined_list), key=operator.itemgetter(1, 2)) # sort djmoffat@718: djmoffat@718: # Show average duration for all songs djmoffat@718: print "Average duration per audioholder:" djmoffat@718: for page_index in range(len(page_names)): djmoffat@718: print " "+combined_list[page_index][0] + ": " \ djmoffat@718: + seconds2timestr(combined_list[page_index][1]) \ djmoffat@718: + " (" + str(combined_list[page_index][3]) + " subjects, " \ djmoffat@718: + str(combined_list[page_index][2]) + " fragments)" djmoffat@718: djmoffat@718: djmoffat@718: #TODO djmoffat@718: # time per page in function of number of fragments (plot) djmoffat@718: # time per participant in function of number of pages djmoffat@718: # plot total time for each participant djmoffat@718: # plot total time djmoffat@718: # show 'count' per page (in order) djmoffat@718: djmoffat@718: # clear up page_index <> page_count <> page_number confusion djmoffat@718: djmoffat@718: # LaTeX -> PDF print out