b@1543: import xml.etree.ElementTree as ET
b@1543: import os
b@1543: import csv
b@1543: 
b@1543: #TODO Remove DEBUG statements
b@1543: 
b@1543: # get every XML file in folder
b@1543: for file in os.listdir("."): # You have to put this in folder where output XML files are.
b@1543:     if file.endswith(".xml"):
b@1543:         tree = ET.parse(file)
b@1543:         root = tree.getroot()
b@1543:         #print ["DEBUG Reading " + file + "..."]
b@1543: 
b@1543:         # get subject ID from XML file
b@1543:         subject_id = file # file name as subject ID
b@1543: 
b@1543:         # get list of all pages this subject evaluated
b@1543:         for audioholder in root.findall("./audioholder"):    # iterate over pages
b@1543:             page_name = audioholder.get('id') # get page name
b@1543:             #print ["DEBUG    page " + page_name]
b@1543:             
b@1543:             if page_name is None: # ignore 'empty' audio_holders
b@1543:                 break
b@1543: 
b@1543:             file_name = 'ratings/'+page_name+'-ratings.csv' # score file name
b@1543: 
b@1543:             # create folder 'ratings if not yet created
b@1543:             if not os.path.exists('ratings'):
b@1543:                 os.makedirs('ratings')
b@1543: 
b@1543:             # header: fragment IDs in 'alphabetical' order
b@1543:             # go to fragment column, or create new column if it doesn't exist yet
b@1543: 
b@1543:             # get array of audio elements and number of audio elements
b@1543:             audiolist = root.findall("*/[@id='"+page_name+"']/audioelement")
b@1543:             n_fragments = len(audiolist)
b@1543: 
b@1543:             # get alphabetical array of fragment IDs from this subject's XML
b@1543:             fragmentnamelist = []    # make empty list
b@1543:             for audioelement in audiolist: # iterate over all audioelements
b@1543:                 fragmentnamelist.append(audioelement.get('id')) # add to list
b@1543: 
b@1543: 
b@1543:             # if file exists, get header and add 'new' fragments
b@1543:             if os.path.isfile(file_name):
b@1543:                 #print ["DEBUG file " + file_name + " already exists - reading header"]
b@1543:                 with open(file_name, 'r') as readfile:
b@1543:                     filereader = csv.reader(readfile, delimiter=',')
b@1543:                     headerrow = filereader.next()
b@1543:                     #headerrow = headerrow[1:] # remove first column (empty)
b@1543: 
b@1543:                     # Which of the fragmentes are in fragmentnamelist but not in headerrow?
b@1543:                     newfragments = list(set(fragmentnamelist)-set(headerrow))
b@1543:                     newfragments = sorted(newfragments) # new fragments in alphabetical order
b@1543:                     # If not empty, read file and rewrite adding extra columns
b@1543:                     if newfragments: # if not empty
b@1543:                         print ["DEBUG New fragments found: " + str(newfragments)]
b@1543:                         with open('temp.csv', 'w') as writefile:
b@1543:                             filewriter = csv.writer(writefile, delimiter=',')
b@1543:                             filewriter.writerow(headerrow + newfragments) # write new header
b@1543:                             for row in filereader: # rewrite row plus empty cells for every new fragment name
b@1543:                                 #print ["DEBUG Old row: " + str(row)]
b@1543:                                 filewriter.writerow(row + ['']*len(newfragments))
b@1543:                                 #print ["DEBUG New row: " + str(row + ['']*len(newfragments))]
b@1543:                         os.rename('temp.csv', file_name) # replace old file with temp file
b@1543:                         headerrow = headerrow + newfragments
b@1543:                         print ["DEBUG New header row: " + str(headerrow)]
b@1543: 
b@1543:             # if not, create file and make header
b@1543:             else:
b@1543:                 #print ["DEBUG file " + file_name + " doesn't exist yet - making new one"]
b@1543:                 headerrow = sorted(fragmentnamelist) # sort alphabetically
b@1543:                 headerrow.insert(0,'')
b@1543:                 fragmentnamelist = fragmentnamelist[1:] #HACKY FIX inserting in firstrow also affects fragmentnamelist
b@1543:                 with open(file_name, 'w') as writefile:
b@1543:                     filewriter = csv.writer(writefile, delimiter=',')
b@1543:                     filewriter.writerow(headerrow)
b@1543: 
b@1543:             # open file to write for this page
b@1543:             writefile = open(file_name, 'a')
b@1543:             filewriter = csv.writer(writefile, delimiter=',')
b@1543: 
b@1543:             # prepare row to be written for this subject for this page
b@1543:             ratingrow = [subject_id]
b@1543: 
b@1543:             # get scores related to fragment [id]
b@1543:             for fragmentname in headerrow[1:]: # iterate over fragments in header (skip first empty column)
b@1543:                 elementvalue = root.find("*/[@id='"
b@1543:                                        + page_name
b@1543:                                        + "']/audioelement/[@id='"
b@1543:                                        + fragmentname
b@1543:                                        + "']/value")
b@1543:                 if hasattr(elementvalue, 'text'): # if rating for this fragment exists
b@1543:                     ratingrow.append(elementvalue.text) # add to rating row
b@1543:                 else: # if this subject has not rated this fragment
b@1543:                     ratingrow.append('') # append empty cell
b@1543: 
b@1543:             # write row: [subject ID, rating fragment ID 1, ..., rating fragment ID M]
b@1543:             filewriter.writerow(ratingrow)
b@1543: