n@921: import xml.etree.ElementTree as ET
n@921: import os
n@921: import csv
n@921: 
n@921: #TODO Remove DEBUG statements
n@921: 
n@921: # get every XML file in folder
n@921: for file in os.listdir("."): # You have to put this in folder where output XML files are.
n@921:     if file.endswith(".xml"):
n@921:         tree = ET.parse(file)
n@921:         root = tree.getroot()
n@921:         #print ["DEBUG Reading " + file + "..."]
n@921: 
n@921:         # get subject ID from XML file
n@921:         subject_id = file # file name as subject ID
n@921: 
n@921:         # get list of all pages this subject evaluated
n@921:         for audioholder in root.findall("./audioholder"):    # iterate over pages
n@921:             page_name = audioholder.get('id') # get page name
n@921:             #print ["DEBUG    page " + page_name]
n@921: 
n@921:             file_name = 'ratings/'+page_name+'-ratings.csv' # score file name
n@921: 
n@921:             # create folder 'ratings if not yet created
n@921:             if not os.path.exists('ratings'):
n@921:                 os.makedirs('ratings')
n@921: 
n@921:             # header: fragment IDs in 'alphabetical' order
n@921:             # go to fragment column, or create new column if it doesn't exist yet
n@921: 
n@921:             # get array of audio elements and number of audio elements
n@921:             audiolist = root.findall("*/[@id='"+page_name+"']/audioelement")
n@921:             n_fragments = len(audiolist)
n@921: 
n@921:             # get alphabetical array of fragment IDs from this subject's XML
n@921:             fragmentnamelist = []    # make empty list
n@921:             for audioelement in audiolist: # iterate over all audioelements
n@921:                 fragmentnamelist.append(audioelement.get('id')) # add to list
n@921: 
n@921: 
n@921:             # if file exists, get header and add 'new' fragments
n@921:             if os.path.isfile(file_name):
n@921:                 #print ["DEBUG file " + file_name + " already exists - reading header"]
n@921:                 with open(file_name, 'r') as readfile:
n@921:                     filereader = csv.reader(readfile, delimiter=',')
n@921:                     headerrow = filereader.next()
n@921:                     #headerrow = headerrow[1:] # remove first column (empty)
n@921: 
n@921:                     # Which of the fragmentes are in fragmentnamelist but not in headerrow?
n@921:                     newfragments = list(set(fragmentnamelist)-set(headerrow))
n@921:                     newfragments = sorted(newfragments) # new fragments in alphabetical order
n@921:                     # If not empty, read file and rewrite adding extra columns
n@921:                     if newfragments: # if not empty
n@921:                         print ["DEBUG New fragments found: " + str(newfragments)]
n@921:                         with open('temp.csv', 'w') as writefile:
n@921:                             filewriter = csv.writer(writefile, delimiter=',')
n@921:                             filewriter.writerow(headerrow + newfragments) # write new header
n@921:                             for row in filereader: # rewrite row plus empty cells for every new fragment name
n@921:                                 #print ["DEBUG Old row: " + str(row)]
n@921:                                 filewriter.writerow(row + ['']*len(newfragments))
n@921:                                 #print ["DEBUG New row: " + str(row + ['']*len(newfragments))]
n@921:                         os.rename('temp.csv', file_name) # replace old file with temp file
n@921:                         headerrow = headerrow + newfragments
n@921:                         print ["DEBUG New header row: " + str(headerrow)]
n@921: 
n@921:             # if not, create file and make header
n@921:             else:
n@921:                 #print ["DEBUG file " + file_name + " doesn't exist yet - making new one"]
n@921:                 headerrow = sorted(fragmentnamelist) # sort alphabetically
n@921:                 headerrow.insert(0,'')
n@921:                 fragmentnamelist = fragmentnamelist[1:] #HACKY FIX inserting in firstrow also affects fragmentnamelist
n@921:                 with open(file_name, 'w') as writefile:
n@921:                     filewriter = csv.writer(writefile, delimiter=',')
n@921:                     filewriter.writerow(headerrow)
n@921: 
n@921:             # open file to write for this page
n@921:             writefile = open(file_name, 'a')
n@921:             filewriter = csv.writer(writefile, delimiter=',')
n@921: 
n@921:             # prepare row to be written for this subject for this page
n@921:             ratingrow = [subject_id]
n@921: 
n@921:             # get scores related to fragment [id]
n@921:             for fragmentname in headerrow[1:]: # iterate over fragments in header (skip first empty column)
n@921:                 elementvalue = root.find("*/[@id='"
n@921:                                        + page_name
n@921:                                        + "']/audioelement/[@id='"
n@921:                                        + fragmentname
n@921:                                        + "']/value")
n@921:                 if hasattr(elementvalue, 'text'): # if rating for this fragment exists
n@921:                     ratingrow.append(elementvalue.text) # add to rating row
n@921:                 else: # if this subject has not rated this fragment
n@921:                     ratingrow.append('') # append empty cell
n@921: 
n@921:             # write row: [subject ID, rating fragment ID 1, ..., rating fragment ID M]
n@921:             filewriter.writerow(ratingrow)
n@921: