me@1942: #!/usr/bin/python
me@1942: 
me@1942: import xml.etree.ElementTree as ET
me@1942: import os
me@1942: import sys
me@1942: import csv
me@1942: 
me@1942: # COMMAND LINE ARGUMENTS
me@1942: 
me@1942: assert len(sys.argv)<3, "score_parser takes at most 1 command line argument\n"+\
me@1942:                         "Use: python score_parser.py [rating_folder_location]"
me@1942: 
me@1942: # XML results files location
me@1942: if len(sys.argv) == 1:
me@1942:     folder_name = "../saves"    # Looks in 'saves/' folder from 'scripts/' folder
me@1942:     print "Use: python score_parser.py [rating_folder_location]"
me@1942:     print "Using default path: " + folder_name
me@1942: elif len(sys.argv) == 2:
me@1942:     folder_name = sys.argv[1]   # First command line argument is folder
me@1942: 
me@1942: # check if folder_name exists
me@1942: if not os.path.exists(folder_name):
me@1942:     #the file is not there
me@1942:     print "Folder '"+folder_name+"' does not exist."
me@1942:     sys.exit() # terminate script execution
me@1942: elif not os.access(os.path.dirname(folder_name), os.W_OK):
me@1942:     #the file does exist but write privileges are not given
me@1942:     print "No write privileges in folder '"+folder_name+"'."
me@1942: 
me@1942:     
me@1942: # CODE
me@1942: 
me@1942: # remember which files have been opened this time
me@1942: file_history = []
me@1942: 
me@1942: # get every XML file in folder
me@1942: for file in os.listdir(folder_name):
me@1942:     if file.endswith(".xml"):
me@1942:         tree = ET.parse(folder_name + '/' + file)
me@1942:         root = tree.getroot()
me@1942: 
me@1942:         # get subject ID from XML file
me@1942:         subject_id = file[:-4] # file name (without extension) as subject ID
me@1942: 
me@1942:         # get list of all pages this subject evaluated
me@1942:         for audioholder in root.findall("./audioholder"):    # iterate over pages
me@1942:             page_name = audioholder.get('id') # get page name
me@1942:                        
me@1942:             if page_name is None: # ignore 'empty' audio_holders
me@1942:                 print "WARNING: " + file + " contains empty audio holder. (score_parser.py)"
me@1942:                 break
me@1942: 
me@1942:             file_name = folder_name+'/ratings/'+page_name+'-ratings.csv' # score file name
me@1942: 
me@1942:             # create folder 'ratings' if not yet created
me@1942:             if not os.path.exists(folder_name + '/ratings'):
me@1942:                 os.makedirs(folder_name + '/ratings')
me@1942: 
me@1942:             # header: fragment IDs in 'alphabetical' order
me@1942:             # go to fragment column, or create new column if it doesn't exist yet
me@1942: 
me@1942:             # get array of audio elements and number of audio elements
me@1942:             audiolist = audioholder.findall("./audioelement")
me@1942:             n_fragments = len(audiolist)
me@1942: 
me@1942:             # get alphabetical array of fragment IDs from this subject's XML
me@1942:             fragmentnamelist = []    # make empty list
me@1942:             for audioelement in audiolist: # iterate over all audioelements
me@1942:                 fragmentnamelist.append(audioelement.get('id')) # add to list
me@1942: 
me@1942:             # if file exists, get header and add 'new' fragments
me@1942:             if os.path.isfile(file_name):
me@1942:                 with open(file_name, 'r') as readfile:
me@1942:                     filereader = csv.reader(readfile, delimiter=',')
me@1942:                     headerrow = filereader.next()
me@1942: 
me@1942:                 # If file hasn't been opened yet this time, remove all rows except header
me@1942:                 if file_name not in file_history:
me@1942:                     with open(file_name, 'w') as writefile:
me@1942:                         filewriter = csv.writer(writefile, delimiter=',')
me@1942:                         headerrow = sorted(headerrow)
me@1942:                         filewriter.writerow(headerrow)
me@1942:                     file_history.append(file_name)
me@1942: 
me@1942:                 # Which of the fragmentes are in fragmentnamelist but not in headerrow?
me@1942:                 newfragments = list(set(fragmentnamelist)-set(headerrow))
me@1942:                 newfragments = sorted(newfragments) # new fragments in alphabetical order
me@1942:                 # If not empty, read file and rewrite adding extra columns
me@1942:                 if newfragments: # if not empty
me@1942:                     with open('temp.csv', 'w') as writefile:
me@1942:                         filewriter = csv.writer(writefile, delimiter=',')
me@1942:                         filewriter.writerow(headerrow + newfragments) # write new header
me@1942:                         with open(file_name, 'r') as readfile:
me@1942:                             filereader = csv.reader(readfile, delimiter=',')
me@1942:                             filereader.next() # skip header
me@1942:                             for row in filereader: # rewrite row plus empty cells for every new fragment name
me@1942:                                 filewriter.writerow(row + ['']*len(newfragments))
me@1942:                     os.rename('temp.csv', file_name) # replace old file with temp file
me@1942:                     headerrow = headerrow + newfragments
me@1942: 
me@1942:             # if not, create file and make header
me@1942:             else:
me@1942:                 headerrow = sorted(fragmentnamelist) # sort alphabetically
me@1942:                 headerrow.insert(0,'')
me@1942:                 fragmentnamelist = fragmentnamelist[1:] #HACKY FIX inserting in firstrow also affects fragmentnamelist
me@1942:                 with open(file_name, 'w') as writefile:
me@1942:                     filewriter = csv.writer(writefile, delimiter=',')
me@1942:                     filewriter.writerow(headerrow)
me@1942: 
me@1942:             # open file to write for this page
me@1942:             writefile = open(file_name, 'a')
me@1942:             filewriter = csv.writer(writefile, delimiter=',')
me@1942: 
me@1942:             # prepare row to be written for this subject for this page
me@1942:             ratingrow = [subject_id]
me@1942: 
me@1942:             # get scores related to fragment [id]
me@1942:             for fragmentname in headerrow[1:]: # iterate over fragments in header (skip first empty column)
me@1942:                 elementvalue = audioholder.find("./audioelement/[@id='"
me@1942:                                        + fragmentname
me@1942:                                        + "']/value")
me@1942:                 if hasattr(elementvalue, 'text'): # if rating for this fragment exists
me@1942:                     ratingrow.append(elementvalue.text) # add to rating row
me@1942:                 else: # if this subject has not rated this fragment
me@1942:                     ratingrow.append('') # append empty cell
me@1942: 
me@1942:             # write row: [subject ID, rating fragment ID 1, ..., rating fragment ID M]
me@1942:             if any(ratingrow[1:]): # append to file if row non-empty (except subject name)
me@1942:                 filewriter.writerow(ratingrow)