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