giuliomoro@15: #!/usr/bin/python
giuliomoro@15: 
giuliomoro@15: import sys
giuliomoro@15: import os
giuliomoro@15: import csv
giuliomoro@15: import matplotlib.pyplot as plt
giuliomoro@15: import numpy as np
giuliomoro@15: import scipy as sp
giuliomoro@15: import scipy.stats
giuliomoro@15: 
giuliomoro@15: # COMMAND LINE ARGUMENTS
giuliomoro@15: 
giuliomoro@15: #TODO: Merge, implement this functionality
giuliomoro@15: #TODO: Control by CLI arguments (plot types, save and/or show, ...) 
giuliomoro@15: 
giuliomoro@15: assert len(sys.argv)<4, "score_plot takes at most 2 command line arguments\n"+\
giuliomoro@15:                         "Use: python score_plot.py [ratings_folder_location]."+\
giuliomoro@15:                         "Type 'python score_plot.py -h' for more options"
giuliomoro@15: 
giuliomoro@15: # initialise plot types (false by default) and options
giuliomoro@15: enable_boxplot    = False     # show box plot
giuliomoro@15: enable_confidence = False     # show confidence interval
giuliomoro@15: confidence        = 0.90      # confidence value (for confidence interval plot)
giuliomoro@15: enable_individual = False     # show all individual ratings
giuliomoro@15: show_individual   = []        # show specific individuals (empty: show all individuals found)
giuliomoro@15: show_legend       = False     # show names of individuals
giuliomoro@15: 
giuliomoro@15: # DEFAULT: Looks in 'saves/ratings/' folder from 'scripts/' folder
giuliomoro@15: rating_folder = "../saves/ratings/" 
giuliomoro@15: 
giuliomoro@15: # XML results files location
giuliomoro@15: if len(sys.argv) == 1: # no extra arguments
giuliomoro@15:     enable_boxplot    = True # show box plot
giuliomoro@15:     print("Use: python score_plot.py [rating folder] [plot_type] [-l/-legend]")
giuliomoro@15:     print("Type 'python score_plot.py -h' for help.")
giuliomoro@15:     print("Using default path: " + rating_folder + " with boxplot.")
giuliomoro@15: else:
giuliomoro@15:     for arg in sys.argv: # go over all arguments
giuliomoro@15:         if arg == '-h':
giuliomoro@15:             # show help
giuliomoro@15:             #TODO: replace with contents of helpfile score_plot.info (or similar)
giuliomoro@15:             print("Use: python score_plot.py [rating_folder] [plot_type] [-l] [confidence]")
giuliomoro@15:             print("   rating_folder:")
giuliomoro@15:             print("            folder where output of 'score_parser' can be found, and")
giuliomoro@15:             print("            where plots will be stored.")
giuliomoro@15:             print("            By default, '../saves/ratings/' is used.")
giuliomoro@15:             print("")
giuliomoro@15:             print("PLOT TYPES")
giuliomoro@15:             print(" Can be used in combination.")
giuliomoro@15:             print("    box | boxplot | -b")
giuliomoro@15:             print("            Enables the boxplot" )
giuliomoro@15:             print("    conf | confidence | -c")
giuliomoro@15:             print("            Enables the confidence interval plot" )
giuliomoro@15:             print("    ind | individual | -i")
giuliomoro@15:             print("            Enables plot of individual ratings" )
giuliomoro@15:             print("")
giuliomoro@15:             print("PLOT OPTIONS")
giuliomoro@15:             print("    leg | legend | -l")
giuliomoro@15:             print("            For individual plot: show legend with individual file names")
giuliomoro@15:             print("    numeric value between 0 and 1, e.g. 0.95")
giuliomoro@15:             print("            For confidence interval plot: confidence value")
giuliomoro@15:             assert False, ""# stop immediately after showing help #TODO cleaner way
giuliomoro@15:             
giuliomoro@15:         # PLOT TYPES
giuliomoro@15:         elif arg == 'box' or arg == 'boxplot' or arg == '-b':
giuliomoro@15:             enable_boxplot    = True     # show box plot
giuliomoro@15:         elif arg == 'conf' or arg == 'confidence' or arg == '-c':
giuliomoro@15:             enable_confidence = True     # show confidence interval
giuliomoro@15:             #TODO add confidence value input
giuliomoro@15:         elif arg == 'ind' or arg == 'individual' or arg == '-i':
giuliomoro@15:             enable_individual = True     # show all individual ratings
giuliomoro@15:             
giuliomoro@15:         # PLOT OPTIONS
giuliomoro@15:         elif arg == 'leg' or arg == 'legend' or arg == '-l':
giuliomoro@15:             if not enable_individual: 
giuliomoro@15:                 print("WARNING: The 'legend' option is only relevant to plots of "+\
giuliomoro@15:                       "individual ratings")
giuliomoro@15:             show_legend = True     # show all individual ratings
giuliomoro@15:         elif arg.isdigit():
giuliomoro@15:             if not enable_confidence: 
giuliomoro@15:                 print("WARNING: The numeric confidence value is only relevant when "+\
giuliomoro@15:                       "confidence plot is enabled")
giuliomoro@15:             if float(arg)>0 and float(arg)<1:
giuliomoro@15:                 confidence = float(arg)
giuliomoro@15:             else: 
giuliomoro@15:                 print("WARNING: The confidence value needs to be between 0 and 1")
giuliomoro@15:         
giuliomoro@15:         # FOLDER NAME
giuliomoro@15:         else: 
giuliomoro@15:              # assume it's the folder name
giuliomoro@15:              rating_folder = arg
giuliomoro@15: 
giuliomoro@15: # at least one plot type should be selected: box plot by default
giuliomoro@15: if not enable_boxplot and not enable_confidence and not enable_individual:
giuliomoro@15:     enable_boxplot = True
giuliomoro@15: 
giuliomoro@15: # check if folder_name exists
giuliomoro@15: if not os.path.exists(rating_folder):
giuliomoro@15:     #the file is not there
giuliomoro@15:     print("Folder '"+rating_folder+"' does not exist.")
giuliomoro@15:     sys.exit() # terminate script execution
giuliomoro@15: elif not os.access(os.path.dirname(rating_folder), os.W_OK):
giuliomoro@15:     #the file does exist but write rating_folder are not given
giuliomoro@15:     print("No write privileges in folder '"+rating_folder+"'.")
giuliomoro@15: 
giuliomoro@15: 
giuliomoro@15: # CONFIGURATION
giuliomoro@15: 
giuliomoro@15: # Font settings
giuliomoro@15: font = {'weight' : 'bold',
giuliomoro@15:         'size'   : 10}
giuliomoro@15: plt.rc('font', **font)
giuliomoro@15: 
giuliomoro@15: 
giuliomoro@15: # CODE
giuliomoro@15: 
giuliomoro@15: # get every csv file in folder
giuliomoro@15: for file in os.listdir(rating_folder):
giuliomoro@15:     if file.endswith(".csv"):
giuliomoro@15:         page_name = file[:-4] # file name (without extension) is page ID
giuliomoro@15: 
giuliomoro@15:         # get header (as text)
giuliomoro@15:         with open(rating_folder+file, 'rt') as readfile: # read this csv file
giuliomoro@15:             filereader = csv.reader(readfile, delimiter=',')
giuliomoro@15:             headerrow = next(filereader) # use headerrow as X-axis
giuliomoro@15:             headerrow = headerrow[1:]
giuliomoro@15: 
giuliomoro@15:         # read ratings into matrix (as bytes)
giuliomoro@15:         with open(rating_folder+file, 'rb') as readfile: # read this csv file
giuliomoro@15:             filereader = csv.reader(readfile, delimiter=',')
giuliomoro@15:             ratings = np.genfromtxt(readfile,
giuliomoro@15:                                    delimiter=",",
giuliomoro@15:                                    skip_header = 1,
giuliomoro@15:                                    converters = {3: lambda s: float(s or 'Nan')},
giuliomoro@15:                                    usecols=list(range(1,len(headerrow)+1))
giuliomoro@15:                                    )
giuliomoro@15:         
giuliomoro@15:         # assert at least 2 subjects (move on to next file if violated)
giuliomoro@15:         if ratings.shape[0]<2:
giuliomoro@15:             print("WARNING: Just one subject for " + page_name + ". Moving on to next file.")
giuliomoro@15:             break
giuliomoro@15: 
giuliomoro@15:         # BOXPLOT
giuliomoro@15:         if enable_boxplot:
giuliomoro@15:             plt.boxplot(ratings)
giuliomoro@15:             
giuliomoro@15:         # CONFIDENCE INTERVAL
giuliomoro@15:         if enable_confidence:
giuliomoro@15:             iterator = 0
giuliomoro@15:             for column in ratings.T: # iterate over transposed matrix
giuliomoro@15:                 # remove all 'Nan's from column
giuliomoro@15:                 column = column[~np.isnan(column)]
giuliomoro@15:             
giuliomoro@15:                 # get number of non-Nan ratings (= #subjects)
giuliomoro@15:                 n = column.size
giuliomoro@15:         
giuliomoro@15:                 # get mean
giuliomoro@15:                 mean_rating = np.mean(column)
giuliomoro@15:         
giuliomoro@15:                 # get errors
giuliomoro@15:                 err = scipy.stats.sem(column)* sp.stats.t._ppf((1+confidence)/2., n-1)
giuliomoro@15:         
giuliomoro@15:                 # draw plot
giuliomoro@15:                 plt.errorbar(iterator+1, 
giuliomoro@15:                             mean_rating, 
giuliomoro@15:                             yerr=err,
giuliomoro@15:                             marker="x",
giuliomoro@15:                             color ="k",
giuliomoro@15:                             markersize=12,
giuliomoro@15:                             linestyle='None')
giuliomoro@15:                             
giuliomoro@15:                 iterator += 1 # increase counter
giuliomoro@15:     
giuliomoro@15:     
giuliomoro@15:         # INDIVIDUAL PLOT
giuliomoro@15:         if enable_individual or show_individual:
giuliomoro@15:             # marker list and color map to cycle through
giuliomoro@15:             markerlist = ["x", ".", "o", "*", "+", "v", ">", "<", "8", "s", "p"]
giuliomoro@15:             colormap = ['b', 'r', 'g', 'c', 'm', 'y', 'k'] 
giuliomoro@15:             increment = 0
giuliomoro@15:             linehandles = []
giuliomoro@15:             legendnames = []
giuliomoro@15:             with open(rating_folder+file, 'r') as readfile: # read this csv file
giuliomoro@15:                 filereader = csv.reader(readfile, delimiter=',')
giuliomoro@15:                 headerrow = next(filereader) # use headerrow as X-axis
giuliomoro@15:                 headerrow = headerrow[1:]
giuliomoro@15:                 for row in filereader:
giuliomoro@15:                     subject_id = row[0][:-4] # read from beginning of line
giuliomoro@15:                     # assume plotting all individuals if no individual(s) specified
giuliomoro@15:                     if not show_individual or subject_id in show_individual:
giuliomoro@15:                         plothandle, = plt.plot(range(1,len(row)), # x-values
giuliomoro@15:                                  ratings[increment,:],#row[1:], # y-values: csv values except subject name
giuliomoro@15:                                  color=colormap[increment%len(colormap)],
giuliomoro@15:                                  marker=markerlist[increment%len(markerlist)],
giuliomoro@15:                                  markersize=10,
giuliomoro@15:                                  linestyle='None',
giuliomoro@15:                                  label=subject_id
giuliomoro@15:                                 )
giuliomoro@15:                         linehandles.append(plothandle)
giuliomoro@15:                         legendnames.append(subject_id)
giuliomoro@15:                         if show_legend:
giuliomoro@15:                             plt.legend(linehandles, legendnames,
giuliomoro@15:                                        loc='upper right',
giuliomoro@15:                                        bbox_to_anchor=(1.1, 1),
giuliomoro@15:                                        borderaxespad=0.,
giuliomoro@15:                                        numpoints=1 # remove extra marker
giuliomoro@15:                                        )
giuliomoro@15:                     increment += 1 # increase counter
giuliomoro@15: 
giuliomoro@15:         # TITLE, AXIS LABELS AND LIMITS
giuliomoro@15:         plt.title(page_name)
giuliomoro@15:         plt.xlabel('Fragment')
giuliomoro@15:         plt.xlim(0, len(headerrow)+1) # only show relevant region, leave space left & right)
giuliomoro@15:         plt.xticks(range(1, len(headerrow)+1), headerrow, rotation=90) # show fragment names
giuliomoro@15:         plt.ylabel('Rating')
giuliomoro@15:         plt.ylim(0,1)
giuliomoro@15:         
giuliomoro@15:         
giuliomoro@15: 
giuliomoro@15:         # SHOW PLOT
giuliomoro@15:         #plt.show()
giuliomoro@15:         #exit()
giuliomoro@15:     
giuliomoro@15:         # SAVE PLOT
giuliomoro@15:         # automatically 
giuliomoro@15:         plot_type = ("-box" if enable_boxplot else "") + \
giuliomoro@15:                     ("-conf" if enable_confidence else "") + \
giuliomoro@15:                     ("-ind" if enable_individual else "")
giuliomoro@15:         plt.savefig(rating_folder+page_name+plot_type+".pdf", bbox_inches='tight')
giuliomoro@15:         plt.close()