nicholas@763: #!/usr/bin/python nicholas@763: # -*- coding: utf-8 -*- nicholas@763: nicholas@763: import xml.etree.ElementTree as ET nicholas@763: import os # for getting files from directory nicholas@763: import operator # for sorting data with multiple keys nicholas@763: import sys # for accessing command line arguments nicholas@763: nicholas@763: # Command line arguments nicholas@763: assert len(sys.argv)<3, "evaluation_stats takes at most 1 command line argument\n"+\ nicholas@763: "Use: python evaluation_stats.py [results_folder]" nicholas@763: nicholas@763: # XML results files location nicholas@763: if len(sys.argv) == 1: nicholas@763: folder_name = "../saves" # Looks in 'saves/' folder from 'scripts/' folder nicholas@763: print "Use: python evaluation_stats.py [results_folder]" nicholas@763: print "Using default path: " + folder_name nicholas@763: elif len(sys.argv) == 2: nicholas@763: folder_name = sys.argv[1] # First command line argument is folder nicholas@763: nicholas@763: # Turn number of seconds (int) to '[minutes] min [seconds] s' (string) nicholas@763: def seconds2timestr(time_in_seconds): nicholas@763: time_in_minutes = int(time_in_seconds/60) nicholas@763: remaining_seconds = int(time_in_seconds%60) nicholas@763: return str(time_in_minutes) + " min " + str(remaining_seconds) + " s" nicholas@763: nicholas@763: # stats initialisation nicholas@763: number_of_XML_files = 0 nicholas@763: number_of_pages = 0 nicholas@763: number_of_fragments = 0 nicholas@763: total_empty_comments = 0 nicholas@763: total_not_played = 0 nicholas@763: total_not_moved = 0 nicholas@763: time_per_page_accum = 0 nicholas@763: nicholas@763: # arrays initialisation nicholas@763: page_names = [] nicholas@763: page_count = [] nicholas@763: duration_page = [] # duration of experiment in function of page content nicholas@763: duration_order = [] # duration of experiment in function of page number nicholas@763: fragments_per_page = [] # number of fragments for corresponding page nicholas@763: nicholas@763: # get every XML file in folder nicholas@763: files_list = os.listdir(folder_name) nicholas@763: for file in files_list: # iterate over all files in files_list nicholas@763: if file.endswith(".xml"): # check if XML file nicholas@763: number_of_XML_files += 1 nicholas@763: tree = ET.parse(folder_name + '/' + file) nicholas@763: root = tree.getroot() nicholas@763: nicholas@763: print file # print file name (subject name) nicholas@763: nicholas@763: # reset for new subject nicholas@763: total_duration = 0 nicholas@763: page_number = 0 nicholas@763: nicholas@763: # get list of all page names nicholas@763: for audioholder in root.findall("./audioholder"): # iterate over pages nicholas@763: page_name = audioholder.get('id') # get page name nicholas@763: nicholas@763: if page_name is None: # ignore 'empty' audio_holders nicholas@763: print "WARNING: " + file + " contains empty audio holder. (evaluation_stats.py)" nicholas@763: break # move on to next nicholas@763: nicholas@763: number_of_comments = 0 # for this page nicholas@763: number_of_missing_comments = 0 # for this page nicholas@763: not_played = 0 # for this page nicholas@763: not_moved = 0 # for this page nicholas@763: nicholas@763: # 'testTime' keeps total duration: subtract time so far for duration of this audioholder nicholas@763: duration = float(audioholder.find("./metric/metricresult[@id='testTime']").text) - total_duration nicholas@763: nicholas@763: # total duration of test nicholas@763: total_duration += duration nicholas@763: nicholas@763: # number of audio elements nicholas@763: audioelements = audioholder.findall("./audioelement") # get audioelements nicholas@763: number_of_fragments += len(audioelements) # add length of this list to total nicholas@763: nicholas@763: # number of comments (interesting if comments not mandatory) nicholas@763: for audioelement in audioelements: nicholas@763: response = audioelement.find("./comment/response") nicholas@763: was_played = audioelement.find("./metric/metricresult/[@name='elementFlagListenedTo']") nicholas@763: was_moved = audioelement.find("./metric/metricresult/[@name='elementFlagMoved']") nicholas@763: if response.text is not None and len(response.text) > 1: nicholas@763: number_of_comments += 1 nicholas@763: else: nicholas@763: number_of_missing_comments += 1 nicholas@763: if was_played is not None and was_played.text == 'false': nicholas@763: not_played += 1 nicholas@763: if was_moved is not None and was_moved.text == 'false': nicholas@763: not_moved += 1 nicholas@763: nicholas@763: # update global counters nicholas@763: total_empty_comments += number_of_missing_comments nicholas@763: total_not_played += not_played nicholas@763: total_not_moved += not_moved nicholas@763: nicholas@763: # print audioholder id and duration nicholas@763: print " " + page_name + ": " + seconds2timestr(duration) + ", "\ nicholas@763: + str(number_of_comments)+"/"\ nicholas@763: +str(number_of_comments+number_of_missing_comments)+" comments" nicholas@763: nicholas@763: # number of audio elements not played nicholas@763: if not_played > 1: nicholas@763: print 'ATTENTION: '+str(not_played)+' fragments were not listened to!' nicholas@763: if not_played == 1: nicholas@763: print 'ATTENTION: one fragment was not listened to!' nicholas@763: nicholas@763: # number of audio element markers not moved nicholas@763: if not_moved > 1: nicholas@763: print 'ATTENTION: '+str(not_moved)+' markers were not moved!' nicholas@763: if not_moved == 1: nicholas@763: print 'ATTENTION: one marker was not moved!' nicholas@763: nicholas@763: # keep track of duration in function of page index nicholas@763: if len(duration_order)>page_number: nicholas@763: duration_order[page_number].append(duration) nicholas@763: else: nicholas@763: duration_order.append([duration]) nicholas@763: nicholas@763: # keep list of audioholder ids and count how many times each audioholder id nicholas@763: # was tested, how long it took, and how many fragments there were (if number of nicholas@763: # fragments is different, store as different audioholder id) nicholas@763: if page_name in page_names: nicholas@763: page_index = page_names.index(page_name) # get index nicholas@763: # check if number of audioelements the same nicholas@763: if len(audioelements) == fragments_per_page[page_index]: nicholas@763: page_count[page_index] += 1 nicholas@763: duration_page[page_index].append(duration) nicholas@763: else: # make new entry nicholas@763: alt_page_name = page_name+"("+str(len(audioelements))+")" nicholas@763: if alt_page_name in page_names: # if already there nicholas@763: alt_page_index = page_names.index(alt_page_name) # get index nicholas@763: page_count[alt_page_index] += 1 nicholas@763: duration_page[alt_page_index].append(duration) nicholas@763: else: nicholas@763: page_names.append(alt_page_name) nicholas@763: page_count.append(1) nicholas@763: duration_page.append([duration]) nicholas@763: fragments_per_page.append(len(audioelements)) nicholas@763: else: nicholas@763: page_names.append(page_name) nicholas@763: page_count.append(1) nicholas@763: duration_page.append([duration]) nicholas@763: fragments_per_page.append(len(audioelements)) nicholas@763: nicholas@763: # bookkeeping nicholas@763: page_number += 1 # increase page count for this specific test nicholas@763: number_of_pages += 1 # increase total number of pages nicholas@763: time_per_page_accum += duration # total duration (for average time spent per page) nicholas@763: nicholas@763: # print total duration of this test nicholas@763: print " TOTAL: " + seconds2timestr(total_duration) nicholas@763: nicholas@763: nicholas@763: # PRINT EVERYTHING nicholas@763: nicholas@763: print "Number of XML files: " + str(number_of_XML_files) nicholas@763: print "Number of pages: " + str(number_of_pages) nicholas@763: print "Number of fragments: " + str(number_of_fragments) nicholas@763: print "Number of empty comments: " + str(total_empty_comments) +\ nicholas@763: " (" + str(round(100.0*total_empty_comments/number_of_fragments,2)) + "%)" nicholas@763: print "Number of unplayed fragments: " + str(total_not_played) +\ nicholas@763: " (" + str(round(100.0*total_not_played/number_of_fragments,2)) + "%)" nicholas@763: print "Number of unmoved markers: " + str(total_not_moved) +\ nicholas@763: " (" + str(round(100.0*total_not_moved/number_of_fragments,2)) + "%)" nicholas@763: print "Average time per page: " + seconds2timestr(time_per_page_accum/number_of_pages) nicholas@763: nicholas@763: # Pages and number of times tested nicholas@763: page_count_strings = list(str(x) for x in page_count) nicholas@763: count_list = page_names + page_count_strings nicholas@763: count_list[::2] = page_names nicholas@763: count_list[1::2] = page_count_strings nicholas@763: print "Pages tested: " + str(count_list) nicholas@763: nicholas@763: # Average duration for first, second, ... page nicholas@763: print "Average duration per page:" nicholas@763: for page_number in range(len(duration_order)): nicholas@763: print " page " + str(page_number+1) + ": " +\ nicholas@763: seconds2timestr(sum(duration_order[page_number])/len(duration_order[page_number])) +\ nicholas@763: " ("+str(len(duration_order[page_number]))+" subjects)" nicholas@763: nicholas@763: nicholas@763: # Sort pages by number of audioelements, then by duration nicholas@763: nicholas@763: # average duration and number of subjects per page nicholas@763: average_duration_page = [] nicholas@763: number_of_subjects_page = [] nicholas@763: for line in duration_page: nicholas@763: number_of_subjects_page.append(len(line)) nicholas@763: average_duration_page.append(sum(line)/len(line)) nicholas@763: nicholas@763: # combine and sort in function of number of audioelements and duration nicholas@763: combined_list = [page_names, average_duration_page, fragments_per_page, number_of_subjects_page] nicholas@763: combined_list = sorted(zip(*combined_list), key=operator.itemgetter(1, 2)) # sort nicholas@763: nicholas@763: # Show average duration for all songs nicholas@763: print "Average duration per audioholder:" nicholas@763: for page_index in range(len(page_names)): nicholas@763: print " "+combined_list[page_index][0] + ": " \ nicholas@763: + seconds2timestr(combined_list[page_index][1]) \ nicholas@763: + " (" + str(combined_list[page_index][3]) + " subjects, " \ nicholas@763: + str(combined_list[page_index][2]) + " fragments)" nicholas@763: nicholas@763: nicholas@763: #TODO nicholas@763: # time per page in function of number of fragments (plot) nicholas@763: # time per participant in function of number of pages nicholas@763: # plot total time for each participant nicholas@763: # plot total time nicholas@763: # show 'count' per page (in order) nicholas@763: nicholas@763: # clear up page_index <> page_count <> page_number confusion nicholas@763: nicholas@763: # LaTeX -> PDF print out