giuliomoro@1088: #!/usr/bin/python
giuliomoro@1088: # -*- coding: utf-8 -*-
giuliomoro@1088: 
giuliomoro@1088: import xml.etree.ElementTree as ET
giuliomoro@1088: import os       # for getting files from directory
giuliomoro@1088: import operator # for sorting data with multiple keys
giuliomoro@1088: import sys      # for accessing command line arguments
giuliomoro@1088: import subprocess # for calling pdflatex
giuliomoro@1088: import shlex # for calling pdflatex
giuliomoro@1088: import matplotlib.pyplot as plt # plots
giuliomoro@1088: import numpy as np # numbers
giuliomoro@1088: 
giuliomoro@1088: # Command line arguments
giuliomoro@1088: assert len(sys.argv)<4, "generate_report takes at most 2 command line arguments\n"+\
giuliomoro@1088:                         "Use: python generate_report.py [results_folder] [no_render | -nr]"
giuliomoro@1088: 
giuliomoro@1088: render_figures = True
giuliomoro@1088: 
giuliomoro@1088: # XML results files location
giuliomoro@1088: if len(sys.argv) == 1:
giuliomoro@1088:     folder_name = "../saves/"    # Looks in 'saves/' folder from 'scripts/' folder
giuliomoro@1088:     print "Use: python generate_report.py [results_folder] [no_render | -nr]"
giuliomoro@1088:     print "Using default path: " + folder_name
giuliomoro@1088: elif len(sys.argv) == 2:
giuliomoro@1088:     folder_name = sys.argv[1]   # First command line argument is folder
giuliomoro@1088: elif len(sys.argv) == 3:
giuliomoro@1088:     folder_name = sys.argv[1]   # First command line argument is folder
giuliomoro@1088:     assert sys.argv[2] in ('no_render','-nr'), "Second argument not recognised. \n" +\
giuliomoro@1088:            "Use: python generate_report.py [results_folder] [no_render | -nr]"
giuliomoro@1088:     # Second command line argument is [no_render | -nr]
giuliomoro@1088:     render_figures = False
giuliomoro@1088: 
giuliomoro@1088: def isNaN(num):
giuliomoro@1088:     return num != num
giuliomoro@1088: 
giuliomoro@1088: # Turn number of seconds (int) to '[minutes] min [seconds] s' (string)
giuliomoro@1088: def seconds2timestr(time_in_seconds):
giuliomoro@1088:     if time_in_seconds is not None and not isNaN(time_in_seconds): 
giuliomoro@1088:         time_in_minutes = int(time_in_seconds/60)
giuliomoro@1088:         remaining_seconds = int(time_in_seconds%60)
giuliomoro@1088:         return str(time_in_minutes) + " min " + str(remaining_seconds) + " s"
giuliomoro@1088:     else:
giuliomoro@1088:         return 'N/A'
giuliomoro@1088: 
giuliomoro@1088: # stats initialisation
giuliomoro@1088: number_of_XML_files  = 0
giuliomoro@1088: number_of_pages      = 0
giuliomoro@1088: number_of_fragments  = 0
giuliomoro@1088: total_empty_comments = 0
giuliomoro@1088: total_not_played     = 0
giuliomoro@1088: total_not_moved      = 0
giuliomoro@1088: time_per_page_accum  = 0
giuliomoro@1088: 
giuliomoro@1088: # arrays initialisation
giuliomoro@1088: page_names       = []
giuliomoro@1088: real_page_names  = [] # regardless of differing numbers of fragments
giuliomoro@1088: subject_count    = [] # subjects per audioholder name
giuliomoro@1088: page_count       = []
giuliomoro@1088: duration_page    = []      # duration of experiment in function of page content
giuliomoro@1088: duration_order   = []      # duration of experiment in function of page number
giuliomoro@1088: fragments_per_page = []    # number of fragments for corresponding page
giuliomoro@1088: 
giuliomoro@1088: # survey stats
giuliomoro@1088: gender = []
giuliomoro@1088: age    = []
giuliomoro@1088: 
giuliomoro@1088: # get username if available
giuliomoro@1088: for name in ('LOGNAME', 'USER', 'LNAME', 'USERNAME'):
giuliomoro@1088:     user = os.environ.get(name)
giuliomoro@1088:     if user:
giuliomoro@1088:         break
giuliomoro@1088:     else:
giuliomoro@1088:         user = ''
giuliomoro@1088: 
giuliomoro@1088: 
giuliomoro@1088: # begin LaTeX document
giuliomoro@1088: header = r'''\documentclass[11pt, oneside]{article} 
giuliomoro@1088:           \usepackage{geometry}
giuliomoro@1088:           \geometry{a4paper}
giuliomoro@1088:           \usepackage[parfill]{parskip} % empty line instead of indent
giuliomoro@1088:           \usepackage{graphicx}    % figures
giuliomoro@1088:           \usepackage[space]{grffile} % include figures with spaces in paths
giuliomoro@1088:           \usepackage{hyperref}
giuliomoro@1088:           \usepackage{tikz}           % pie charts
giuliomoro@1088:           \title{Report}
giuliomoro@1088:           \author{'''+\
giuliomoro@1088:           user+\
giuliomoro@1088:           r'''}
giuliomoro@1088:           \graphicspath{{'''+\
giuliomoro@1088:           folder_name+\
giuliomoro@1088:           r'''}}
giuliomoro@1088:           %\setcounter{section}{-1} % Summary section 0 so number of sections equals number of files
giuliomoro@1088:           \begin{document}
giuliomoro@1088:           \maketitle
giuliomoro@1088:           This is an automatically generated report using the `generate\_report.py' Python script 
giuliomoro@1088:           included with the Web Audio Evaluation Tool \cite{WAET} distribution which can be found 
giuliomoro@1088:           at \texttt{code.soundsoftware.ac.uk/projects/webaudioevaluationtool}.
giuliomoro@1088:           \tableofcontents
giuliomoro@1088:           
giuliomoro@1088:           '''
giuliomoro@1088:           
giuliomoro@1088: footer = '\n\t\t'+r'''\begin{thebibliography}{9}
giuliomoro@1088:          \bibitem{WAET} % reference to accompanying publication
giuliomoro@1088:         Nicholas Jillings, Brecht De Man, David Moffat and Joshua D. Reiss, 
giuliomoro@1088:         ``Web Audio Evaluation Tool: A browser-based listening test environment,'' 
giuliomoro@1088:         presented at the 12th Sound and Music Computing Conference, July 2015.
giuliomoro@1088:         \end{thebibliography}
giuliomoro@1088:         \end{document}'''
giuliomoro@1088: 
giuliomoro@1088: body = ''
giuliomoro@1088: 
giuliomoro@1088: # make sure folder_name ends in '/'
giuliomoro@1088: folder_name = os.path.join(folder_name, '')
giuliomoro@1088: 
giuliomoro@1088: # generate images for later use
giuliomoro@1088: if render_figures:
giuliomoro@1088:     subprocess.call("python timeline_view_movement.py '"+folder_name+"'", shell=True)
giuliomoro@1088:     subprocess.call("python score_parser.py '"+folder_name+"'", shell=True)
giuliomoro@1088:     subprocess.call("python score_plot.py '"+folder_name+"ratings/'", shell=True)
giuliomoro@1088: 
giuliomoro@1088: # get every XML file in folder
giuliomoro@1088: files_list = os.listdir(folder_name)
giuliomoro@1088: for file in files_list: # iterate over all files in files_list
giuliomoro@1088:     if file.endswith(".xml"): # check if XML file
giuliomoro@1088:         number_of_XML_files += 1
giuliomoro@1088:         tree = ET.parse(folder_name + file)
giuliomoro@1088:         root = tree.getroot()
giuliomoro@1088:         
giuliomoro@1088:         # PRINT name as section
giuliomoro@1088:         body+= '\n\section{'+file[:-4].capitalize()+'}\n' # make section header from name without extension
giuliomoro@1088:         
giuliomoro@1088:         # reset for new subject
giuliomoro@1088:         total_duration = 0
giuliomoro@1088:         page_number = 0
giuliomoro@1088:         
giuliomoro@1088:         individual_table = '\n' # table with stats for this individual test file
giuliomoro@1088:         timeline_plots = '' # plots of timeline (movements and plays)
giuliomoro@1088:         
giuliomoro@1088:         # DEMO survey stats
giuliomoro@1088:         # get gender
giuliomoro@1088:         this_subjects_gender = root.find("./posttest/radio/[@id='gender']")
giuliomoro@1088:         if this_subjects_gender is not None:
giuliomoro@1088:             gender.append(this_subjects_gender.get("name"))
giuliomoro@1088:         else:
giuliomoro@1088:             gender.append('UNAVAILABLE')
giuliomoro@1088:         # get age
giuliomoro@1088:         this_subjects_age = root.find("./posttest/number/[@id='age']")
giuliomoro@1088:         if this_subjects_age is not None:
giuliomoro@1088:             age.append(this_subjects_age.text)
giuliomoro@1088:         #TODO add plot of age
giuliomoro@1088:                 
giuliomoro@1088:         # get list of all page names
giuliomoro@1088:         for audioholder in root.findall("./audioholder"):   # iterate over pages
giuliomoro@1088:             page_name = audioholder.get('id')               # get page name
giuliomoro@1088:             
giuliomoro@1088:             if page_name is None: # ignore 'empty' audio_holders
giuliomoro@1088:                 print "WARNING: " + file + " contains empty audio holder. (evaluation_stats.py)"
giuliomoro@1088:                 break # move on to next
giuliomoro@1088:             
giuliomoro@1088:             number_of_comments = 0 # for this page
giuliomoro@1088:             number_of_missing_comments = 0 # for this page
giuliomoro@1088:             not_played = [] # for this page
giuliomoro@1088:             not_moved = [] # for this page
giuliomoro@1088:             
giuliomoro@1088:             if audioholder.find("./metric/metricresult[@id='testTime']") is not None: # check if time is included
giuliomoro@1088:                 # 'testTime' keeps total duration: subtract time so far for duration of this audioholder
giuliomoro@1088:                 duration = float(audioholder.find("./metric/metricresult[@id='testTime']").text) - total_duration
giuliomoro@1088:             
giuliomoro@1088:                 # total duration of test
giuliomoro@1088:                 total_duration += duration
giuliomoro@1088:             else: 
giuliomoro@1088:                 duration = float('nan')
giuliomoro@1088:                 total_duration = float('nan')
giuliomoro@1088:             
giuliomoro@1088:             # number of audio elements
giuliomoro@1088:             audioelements = audioholder.findall("./audioelement") # get audioelements
giuliomoro@1088:             number_of_fragments += len(audioelements) # add length of this list to total
giuliomoro@1088:             
giuliomoro@1088:             # number of comments (interesting if comments not mandatory)
giuliomoro@1088:             for audioelement in audioelements:
giuliomoro@1088:                 response = audioelement.find("./comment/response")
giuliomoro@1088:                 was_played = audioelement.find("./metric/metricresult/[@name='elementFlagListenedTo']")
giuliomoro@1088:                 was_moved = audioelement.find("./metric/metricresult/[@name='elementFlagMoved']")
giuliomoro@1088:                 if response.text is not None and len(response.text) > 1: 
giuliomoro@1088:                     number_of_comments += 1
giuliomoro@1088:                 else: 
giuliomoro@1088:                     number_of_missing_comments += 1
giuliomoro@1088:                 if was_played is not None and was_played.text == 'false': 
giuliomoro@1088:                     not_played.append(audioelement.get('id'))
giuliomoro@1088:                 if was_moved is not None and was_moved.text == 'false': 
giuliomoro@1088:                     not_moved.append(audioelement.get('id'))
giuliomoro@1088:             
giuliomoro@1088:             # update global counters
giuliomoro@1088:             total_empty_comments += number_of_missing_comments
giuliomoro@1088:             total_not_played += len(not_played)
giuliomoro@1088:             total_not_moved += len(not_moved)
giuliomoro@1088:             
giuliomoro@1088:             # PRINT alerts when elements not played or markers not moved
giuliomoro@1088:             # number of audio elements not played
giuliomoro@1088:             if len(not_played) > 1:
giuliomoro@1088:                 body += '\t\t\\emph{\\textbf{ATTENTION: '+str(len(not_played))+\
giuliomoro@1088:                         ' fragments were not listened to in '+page_name+'! }}'+\
giuliomoro@1088:                         ', '.join(not_played)+'\\\\ \n'
giuliomoro@1088:             if len(not_played) == 1: 
giuliomoro@1088:                 body += '\t\t\\emph{\\textbf{ATTENTION: one fragment was not listened to in '+page_name+'! }}'+\
giuliomoro@1088:                         not_played[0]+'\\\\ \n'
giuliomoro@1088:             
giuliomoro@1088:             # number of audio element markers not moved
giuliomoro@1088:             if len(not_moved) > 1:
giuliomoro@1088:                 body += '\t\t\\emph{\\textbf{ATTENTION: '+str(len(not_moved))+\
giuliomoro@1088:                         ' markers were not moved in '+page_name+'! }}'+\
giuliomoro@1088:                         ', '.join(not_moved)+'\\\\ \n'
giuliomoro@1088:             if len(not_moved) == 1: 
giuliomoro@1088:                 body += '\t\t\\emph{\\textbf{ATTENTION: one marker was not moved in '+page_name+'! }}'+\
giuliomoro@1088:                         not_moved[0]+'\\\\ \n'
giuliomoro@1088:             
giuliomoro@1088:             # PRINT song-specific statistic
giuliomoro@1088:             individual_table += '\t\t'+page_name+'&'+\
giuliomoro@1088:                                 str(number_of_comments) + '/' +\
giuliomoro@1088:                                 str(number_of_comments+number_of_missing_comments)+'&'+\
giuliomoro@1088:                                 seconds2timestr(duration)+'\\\\\n'
giuliomoro@1088:             
giuliomoro@1088:             # get timeline for this audioholder
giuliomoro@1088:             img_path = 'timelines_movement/'+file[:-4]+'-'+page_name+'.pdf'
giuliomoro@1088:             
giuliomoro@1088:             # check if available
giuliomoro@1088:             if os.path.isfile(folder_name+img_path):
giuliomoro@1088:                 # SHOW timeline image
giuliomoro@1088:                 timeline_plots += '\\includegraphics[width=\\textwidth]{'+\
giuliomoro@1088:                          folder_name+img_path+'}\n\t\t'
giuliomoro@1088:             
giuliomoro@1088:             # keep track of duration in function of page index
giuliomoro@1088:             if len(duration_order)>page_number:
giuliomoro@1088:                 duration_order[page_number].append(duration)
giuliomoro@1088:             else:
giuliomoro@1088:                 duration_order.append([duration])
giuliomoro@1088:             
giuliomoro@1088:             # keep list of audioholder ids and count how many times each audioholder id
giuliomoro@1088:             # was tested, how long it took, and how many fragments there were 
giuliomoro@1088:             # (if number of fragments is different, store as different audioholder id)
giuliomoro@1088:             if page_name in page_names: 
giuliomoro@1088:                 page_index = page_names.index(page_name) # get index
giuliomoro@1088:                 # check if number of audioelements the same
giuliomoro@1088:                 if len(audioelements) == fragments_per_page[page_index]: 
giuliomoro@1088:                     page_count[page_index] += 1
giuliomoro@1088:                     duration_page[page_index].append(duration)
giuliomoro@1088:                 else: # make new entry
giuliomoro@1088:                     alt_page_name = page_name+"("+str(len(audioelements))+")"
giuliomoro@1088:                     if alt_page_name in page_names: # if already there
giuliomoro@1088:                         alt_page_index = page_names.index(alt_page_name) # get index
giuliomoro@1088:                         page_count[alt_page_index] += 1
giuliomoro@1088:                         duration_page[alt_page_index].append(duration)
giuliomoro@1088:                     else: 
giuliomoro@1088:                         page_names.append(alt_page_name)
giuliomoro@1088:                         page_count.append(1)
giuliomoro@1088:                         duration_page.append([duration])
giuliomoro@1088:                         fragments_per_page.append(len(audioelements))
giuliomoro@1088:             else: 
giuliomoro@1088:                 page_names.append(page_name)
giuliomoro@1088:                 page_count.append(1)
giuliomoro@1088:                 duration_page.append([duration])
giuliomoro@1088:                 fragments_per_page.append(len(audioelements))
giuliomoro@1088:             
giuliomoro@1088:             # number of subjects per audioholder regardless of differing numbers of 
giuliomoro@1088:             # fragments (for inclusion in box plots)
giuliomoro@1088:             if page_name in real_page_names:
giuliomoro@1088:                 page_index = real_page_names.index(page_name) # get index
giuliomoro@1088:                 subject_count[page_index] += 1
giuliomoro@1088:             else: 
giuliomoro@1088:                 real_page_names.append(page_name)
giuliomoro@1088:                 subject_count.append(1)
giuliomoro@1088:             
giuliomoro@1088:             # bookkeeping
giuliomoro@1088:             page_number += 1 # increase page count for this specific test
giuliomoro@1088:             number_of_pages += 1 # increase total number of pages
giuliomoro@1088:             time_per_page_accum += duration # total duration (for average time spent per page)
giuliomoro@1088: 
giuliomoro@1088:         # PRINT table with statistics about this test
giuliomoro@1088:         body += '\t\t'+r'''\begin{tabular}{|p{3.5cm}|c|p{2.5cm}|}
giuliomoro@1088:                  \hline
giuliomoro@1088:                  \textbf{Song name} & \textbf{Comments} & \textbf{Duration} \\ \hline '''+\
giuliomoro@1088:                  individual_table+'\t\t'+\
giuliomoro@1088:                  r'''\hline
giuliomoro@1088:                   \textbf{TOTAL} & & \textbf{'''+\
giuliomoro@1088:                   seconds2timestr(total_duration)+\
giuliomoro@1088:                  r'''}\\
giuliomoro@1088:                   \hline 
giuliomoro@1088:                   \end{tabular}
giuliomoro@1088:                   
giuliomoro@1088:                   '''
giuliomoro@1088:         # PRINT timeline plots
giuliomoro@1088:         body += timeline_plots
giuliomoro@1088: 
giuliomoro@1088: # join to footer
giuliomoro@1088: footer = body + footer
giuliomoro@1088: 
giuliomoro@1088: # empty body again
giuliomoro@1088: body = ''
giuliomoro@1088: 
giuliomoro@1088: # PRINT summary of everything (at start) 
giuliomoro@1088: #       unnumbered so that number of sections equals number of files
giuliomoro@1088: body += '\section*{Summary}\n\t\t\\addcontentsline{toc}{section}{Summary}\n'
giuliomoro@1088: 
giuliomoro@1088: # PRINT table with statistics
giuliomoro@1088: body += '\t\t\\begin{tabular}{ll}\n\t\t\t'
giuliomoro@1088: body += r'Number of XML files: &' + str(number_of_XML_files) + r'\\'+'\n\t\t\t'
giuliomoro@1088: body += r'Number of pages: &' + str(number_of_pages) + r'\\'+'\n\t\t\t'
giuliomoro@1088: body += r'Number of fragments: &' + str(number_of_fragments) + r'\\'+'\n\t\t\t'
giuliomoro@1088: body += r'Number of empty comments: &' + str(total_empty_comments) +\
giuliomoro@1088:       " (" + str(round(100.0*total_empty_comments/number_of_fragments,2)) + r"\%)\\"+'\n\t\t\t'
giuliomoro@1088: body += r'Number of unplayed fragments: &' + str(total_not_played) +\
giuliomoro@1088:       " (" + str(round(100.0*total_not_played/number_of_fragments,2)) + r"\%)\\"+'\n\t\t\t'
giuliomoro@1088: body += r'Number of unmoved markers: &' + str(total_not_moved) +\
giuliomoro@1088:       " (" + str(round(100.0*total_not_moved/number_of_fragments,2)) + r"\%)\\"+'\n\t\t\t'
giuliomoro@1088: body += r'Average time per page: &' + seconds2timestr(time_per_page_accum/number_of_pages) + r"\\"+'\n\t\t'
giuliomoro@1088: body += '\\end{tabular} \\vspace{1.5cm} \\\\ \n'
giuliomoro@1088: 
giuliomoro@1088: # Average duration for first, second, ... page
giuliomoro@1088: body += "\t\t\\vspace{.5cm} \n\n\t\tAverage duration per page (see also Figure \\ref{fig:avgtimeperpage}): \\\\ \n\t\t"
giuliomoro@1088: body += r'''\begin{tabular}{lll}
giuliomoro@1088:                     \textbf{Page} & \textbf{Duration} & \textbf{\# subjects}\\'''
giuliomoro@1088: tpp_averages = [] # store average time per page
giuliomoro@1088: for page_number in range(len(duration_order)): 
giuliomoro@1088:     body += '\n\t\t\t'+str(page_number+1) + "&" +\
giuliomoro@1088:         seconds2timestr(sum(duration_order[page_number])/len(duration_order[page_number])) +\
giuliomoro@1088:             "&"+str(len(duration_order[page_number]))+r"\\"
giuliomoro@1088:     tpp_averages.append(sum(duration_order[page_number])/len(duration_order[page_number]))
giuliomoro@1088:             
giuliomoro@1088: body += '\n\t\t\\end{tabular} \\vspace{1.5cm} \\\\ \n\n\t\t'
giuliomoro@1088: 
giuliomoro@1088: # SHOW bar plot of average time per page
giuliomoro@1088: plt.bar(range(1,len(duration_order)+1), np.array(tpp_averages)/60)
giuliomoro@1088: plt.xlabel('Page order')
giuliomoro@1088: plt.xlim(.8, len(duration_order)+1)
giuliomoro@1088: plt.xticks(np.arange(1,len(duration_order)+1)+.4, range(1,len(duration_order)+1))
giuliomoro@1088: plt.ylabel('Average time [minutes]')
giuliomoro@1088: plt.savefig(folder_name+"time_per_page.pdf", bbox_inches='tight')
giuliomoro@1088: plt.close()
giuliomoro@1088: #TODO add error bars
giuliomoro@1088: 
giuliomoro@1088: 
giuliomoro@1088: # Sort pages by number of audioelements, then by duration
giuliomoro@1088: 
giuliomoro@1088: # average duration and number of subjects per page
giuliomoro@1088: average_duration_page = []
giuliomoro@1088: number_of_subjects_page = []
giuliomoro@1088: for line in duration_page:
giuliomoro@1088:     number_of_subjects_page.append(len(line))
giuliomoro@1088:     average_duration_page.append(sum(line)/len(line))
giuliomoro@1088: 
giuliomoro@1088: # combine and sort in function of number of audioelements and duration
giuliomoro@1088: combined_list = [page_names, average_duration_page, fragments_per_page, number_of_subjects_page]
giuliomoro@1088: combined_list = sorted(zip(*combined_list), key=operator.itemgetter(1, 2)) # sort
giuliomoro@1088: 
giuliomoro@1088: # Show average duration for all songs
giuliomoro@1088: body += r'''\vspace{.5cm}
giuliomoro@1088:                 Average duration per audioholder (see also Figure \ref{fig:avgtimeperaudioholder}): \\
giuliomoro@1088:                 \begin{tabular}{llll}
giuliomoro@1088:                         \textbf{Audioholder} & \textbf{Duration} & \textbf{\# subjects} & \textbf{\# fragments} \\'''
giuliomoro@1088: audioholder_names_ordered = []
giuliomoro@1088: average_duration_audioholder_ordered = []
giuliomoro@1088: number_of_subjects = []
giuliomoro@1088: for page_index in range(len(page_names)):
giuliomoro@1088:     audioholder_names_ordered.append(combined_list[page_index][0])
giuliomoro@1088:     average_duration_audioholder_ordered.append(combined_list[page_index][1])
giuliomoro@1088:     number_of_subjects.append(combined_list[page_index][3])
giuliomoro@1088:     body +=  '\n\t\t\t'+combined_list[page_index][0] + "&" +\
giuliomoro@1088:              seconds2timestr(combined_list[page_index][1]) + "&" +\
giuliomoro@1088:              str(combined_list[page_index][3]) + "&" +\
giuliomoro@1088:              str(combined_list[page_index][2]) + r"\\"
giuliomoro@1088: body += '\n\t\t\\end{tabular}\n'
giuliomoro@1088: 
giuliomoro@1088: # SHOW bar plot of average time per page
giuliomoro@1088: plt.bar(range(1,len(audioholder_names_ordered)+1), np.array(average_duration_audioholder_ordered)/60)
giuliomoro@1088: plt.xlabel('Audioholder')
giuliomoro@1088: plt.xlim(.8, len(audioholder_names_ordered)+1)
giuliomoro@1088: plt.xticks(np.arange(1,len(audioholder_names_ordered)+1)+.4, audioholder_names_ordered, rotation=90)
giuliomoro@1088: plt.ylabel('Average time [minutes]')
giuliomoro@1088: plt.savefig(folder_name+"time_per_audioholder.pdf", bbox_inches='tight')
giuliomoro@1088: plt.close()
giuliomoro@1088: 
giuliomoro@1088: # SHOW bar plot of average time per page
giuliomoro@1088: plt.bar(range(1,len(audioholder_names_ordered)+1), number_of_subjects)
giuliomoro@1088: plt.xlabel('Audioholder')
giuliomoro@1088: plt.xlim(.8, len(audioholder_names_ordered)+1)
giuliomoro@1088: plt.xticks(np.arange(1,len(audioholder_names_ordered)+1)+.4, audioholder_names_ordered, rotation=90)
giuliomoro@1088: plt.ylabel('Number of subjects')
giuliomoro@1088: ax = plt.gca()
giuliomoro@1088: ylims = ax.get_ylim()
giuliomoro@1088: yint = np.arange(int(np.floor(ylims[0])), int(np.ceil(ylims[1]))+1)
giuliomoro@1088: plt.yticks(yint)
giuliomoro@1088: plt.savefig(folder_name+"subjects_per_audioholder.pdf", bbox_inches='tight')
giuliomoro@1088: plt.close()
giuliomoro@1088: 
giuliomoro@1088: # SHOW both figures
giuliomoro@1088: body += r'''
giuliomoro@1088:          \begin{figure}[htbp]
giuliomoro@1088:          \begin{center}
giuliomoro@1088:          \includegraphics[width=.65\textwidth]{'''+\
giuliomoro@1088:          folder_name+'time_per_page.pdf'+\
giuliomoro@1088:         r'''}
giuliomoro@1088:         \caption{Average time spent per page.}
giuliomoro@1088:         \label{fig:avgtimeperpage}
giuliomoro@1088:          \end{center}
giuliomoro@1088:          \end{figure}
giuliomoro@1088:          
giuliomoro@1088:          '''
giuliomoro@1088: body += r'''\begin{figure}[htbp]
giuliomoro@1088:          \begin{center}
giuliomoro@1088:          \includegraphics[width=.65\textwidth]{'''+\
giuliomoro@1088:          folder_name+'time_per_audioholder.pdf'+\
giuliomoro@1088:         r'''}
giuliomoro@1088:         \caption{Average time spent per audioholder.}
giuliomoro@1088:         \label{fig:avgtimeperaudioholder}
giuliomoro@1088:          \end{center}
giuliomoro@1088:          \end{figure}
giuliomoro@1088:          
giuliomoro@1088:          '''
giuliomoro@1088: body += r'''\begin{figure}[htbp]
giuliomoro@1088:          \begin{center}
giuliomoro@1088:          \includegraphics[width=.65\textwidth]{'''+\
giuliomoro@1088:          folder_name+'subjects_per_audioholder.pdf'+\
giuliomoro@1088:         r'''}
giuliomoro@1088:         \caption{Number of subjects per audioholder.}
giuliomoro@1088:         \label{fig:subjectsperaudioholder}
giuliomoro@1088:          \end{center}
giuliomoro@1088:          \end{figure}
giuliomoro@1088:          
giuliomoro@1088:          '''
giuliomoro@1088: #TODO add error bars
giuliomoro@1088: #TODO layout of figures
giuliomoro@1088: 
giuliomoro@1088: # SHOW boxplot per audioholder
giuliomoro@1088: #TODO order in decreasing order of participants
giuliomoro@1088: for audioholder_name in page_names: # get each name
giuliomoro@1088:     # plot boxplot if exists (not so for the 'alt' names)
giuliomoro@1088:     if os.path.isfile(folder_name+'ratings/'+audioholder_name+'-ratings-box.pdf'):
giuliomoro@1088:         body += r'''\begin{figure}[htbp]
giuliomoro@1088:              \begin{center}
giuliomoro@1088:              \includegraphics[width=.65\textwidth]{'''+\
giuliomoro@1088:              folder_name+"ratings/"+audioholder_name+'-ratings-box.pdf'+\
giuliomoro@1088:             r'''}
giuliomoro@1088:             \caption{Box plot of ratings for audioholder '''+\
giuliomoro@1088:             audioholder_name+' ('+str(subject_count[real_page_names.index(audioholder_name)])+\
giuliomoro@1088:             ''' participants).}
giuliomoro@1088:             \label{fig:boxplot'''+audioholder_name.replace(" ", "")+'''}
giuliomoro@1088:              \end{center}
giuliomoro@1088:              \end{figure}
giuliomoro@1088:              
giuliomoro@1088:              '''
giuliomoro@1088: 
giuliomoro@1088: # DEMO pie chart of gender distribution among subjects
giuliomoro@1088: genders = ['male', 'female', 'other', 'preferNotToSay', 'UNAVAILABLE']
giuliomoro@1088: # TODO: get the above automatically
giuliomoro@1088: gender_distribution = ''
giuliomoro@1088: for item in genders:
giuliomoro@1088:     number = gender.count(item)
giuliomoro@1088:     if number>0:
giuliomoro@1088:         gender_distribution += str("{:.2f}".format((100.0*number)/len(gender)))+\
giuliomoro@1088:                                '/'+item.capitalize()+' ('+str(number)+'),\n'
giuliomoro@1088: 
giuliomoro@1088: body += r'''
giuliomoro@1088:         % Pie chart of gender distribution
giuliomoro@1088:         \def\angle{0}
giuliomoro@1088:         \def\radius{3}
giuliomoro@1088:         \def\cyclelist{{"orange","blue","red","green"}}
giuliomoro@1088:         \newcount\cyclecount \cyclecount=-1
giuliomoro@1088:         \newcount\ind \ind=-1
giuliomoro@1088:         \begin{figure}[htbp]
giuliomoro@1088:         \begin{center}\begin{tikzpicture}[nodes = {font=\sffamily}]
giuliomoro@1088:         \foreach \percent/\name in {'''+\
giuliomoro@1088:         gender_distribution+\
giuliomoro@1088:         r'''} {\ifx\percent\empty\else               % If \percent is empty, do nothing
giuliomoro@1088:         \global\advance\cyclecount by 1     % Advance cyclecount
giuliomoro@1088:         \global\advance\ind by 1            % Advance list index
giuliomoro@1088:         \ifnum6<\cyclecount                 % If cyclecount is larger than list
giuliomoro@1088:           \global\cyclecount=0              %   reset cyclecount and
giuliomoro@1088:           \global\ind=0                     %   reset list index
giuliomoro@1088:         \fi
giuliomoro@1088:         \pgfmathparse{\cyclelist[\the\ind]} % Get color from cycle list
giuliomoro@1088:         \edef\color{\pgfmathresult}         %   and store as \color
giuliomoro@1088:         % Draw angle and set labels
giuliomoro@1088:         \draw[fill={\color!50},draw={\color}] (0,0) -- (\angle:\radius)
giuliomoro@1088:           arc (\angle:\angle+\percent*3.6:\radius) -- cycle;
giuliomoro@1088:         \node at (\angle+0.5*\percent*3.6:0.7*\radius) {\percent\,\%};
giuliomoro@1088:         \node[pin=\angle+0.5*\percent*3.6:\name]
giuliomoro@1088:           at (\angle+0.5*\percent*3.6:\radius) {};
giuliomoro@1088:         \pgfmathparse{\angle+\percent*3.6}  % Advance angle
giuliomoro@1088:         \xdef\angle{\pgfmathresult}         %   and store in \angle
giuliomoro@1088:         \fi
giuliomoro@1088:         };
giuliomoro@1088:         \end{tikzpicture}
giuliomoro@1088:         \caption{Representation of gender across subjects}
giuliomoro@1088:         \label{default}
giuliomoro@1088:         \end{center}
giuliomoro@1088:         \end{figure}
giuliomoro@1088:         
giuliomoro@1088:         '''
giuliomoro@1088: # problem: some people entered twice? 
giuliomoro@1088: 
giuliomoro@1088: #TODO
giuliomoro@1088: # time per page in function of number of fragments (plot)
giuliomoro@1088: # time per participant in function of number of pages
giuliomoro@1088: # plot total time for each participant
giuliomoro@1088: # show 'count' per page (in order)
giuliomoro@1088: 
giuliomoro@1088: # clear up page_index <> page_count <> page_number confusion
giuliomoro@1088: 
giuliomoro@1088: 
giuliomoro@1088: texfile = header+body+footer # add bits together
giuliomoro@1088: 
giuliomoro@1088: print 'pdflatex -output-directory="'+folder_name+'"" "'+ folder_name + 'Report.tex"' # DEBUG
giuliomoro@1088: 
giuliomoro@1088: # write TeX file
giuliomoro@1088: with open(folder_name + 'Report.tex','w') as f:
giuliomoro@1088:     f.write(texfile)
giuliomoro@1088: proc=subprocess.Popen(shlex.split('pdflatex -output-directory="'+folder_name+'" "'+ folder_name + 'Report.tex"'))
giuliomoro@1088: proc.communicate()
giuliomoro@1088: # run again
giuliomoro@1088: proc=subprocess.Popen(shlex.split('pdflatex -output-directory="'+folder_name+'" "'+ folder_name + 'Report.tex"'))
giuliomoro@1088: proc.communicate()
giuliomoro@1088: 
giuliomoro@1088: #TODO remove auxiliary LaTeX files
giuliomoro@1088: try:
giuliomoro@1088:     os.remove(folder_name + 'Report.aux')
giuliomoro@1088:     os.remove(folder_name + 'Report.log')
giuliomoro@1088:     os.remove(folder_name + 'Report.out')
giuliomoro@1088:     os.remove(folder_name + 'Report.toc')
giuliomoro@1088: except OSError:
giuliomoro@1088:     pass
giuliomoro@1088: