b@230: #!/usr/bin/python
b@241: # -*- coding: utf-8 -*-
b@230: 
b@146: import xml.etree.ElementTree as ET
b@146: import os
b@146: import csv
b@290: import sys
b@284: 
b@284: # COMMAND LINE ARGUMENTS
b@284: 
b@284: assert len(sys.argv)<3, "comment_parser takes at most 1 command line argument\n"+\
b@284:                         "Use: python score_parser.py [rating_folder_location]"
b@284: 
b@284: # XML results files location
b@284: if len(sys.argv) == 1:
b@284:     folder_name = "../saves"    # Looks in 'saves/' folder from 'scripts/' folder
b@284:     print "Use: python  comment_parser.py [XML_files_location]"
b@284:     print "Using default path: " + folder_name
b@284: elif len(sys.argv) == 2:
b@284:     folder_name = sys.argv[1]   # First command line argument is folder
b@284: 
b@284: # check if folder_name exists
b@284: if not os.path.exists(folder_name):
b@284:     #the file is not there
b@284:     print "Folder '"+folder_name+"' does not exist."
b@284:     sys.exit() # terminate script execution
b@284: elif not os.access(os.path.dirname(folder_name), os.W_OK):
b@284:     #the file does exist but write privileges are not given
b@284:     print "No write privileges in folder '"+folder_name+"'."
b@284: 
b@284: 
b@284: # CODE
b@245: 
b@290: # remember which files have been opened this time
b@290: file_history = []
b@290: 
b@146: # get every XML file in folder
b@245: for file in os.listdir(folder_name): 
b@146:     if file.endswith(".xml"):
b@245:         tree = ET.parse(folder_name + '/' + file)
b@146:         root = tree.getroot()
b@251:         
b@147:         # get list of all page names
b@147:         for audioholder in root.findall("./audioholder"):   # iterate over pages
b@147:             page_name = audioholder.get('id')               # get page name
b@206:             
b@206:             if page_name is None: # ignore 'empty' audio_holders
b@251:                 print "WARNING: " + file + " contains empty audio holder. (comment_parser.py)"
b@206:                 break
b@146: 
b@147:             # create folder [page_name] if not yet created
b@245:             if not os.path.exists(folder_name + "/" + page_name):
b@245:                 os.makedirs(folder_name + "/" + page_name)
b@146: 
b@147:             # for page [page_name], print comments related to fragment [id]
b@245:             for audioelement in audioholder.findall("./audioelement"):
b@206:                 if audioelement is not None: # Check it exists
b@206:                     audio_id = str(audioelement.get('id'))
b@230:                     
b@245:                     csv_name = folder_name +'/' + page_name+'/'+page_name+'-comments-'+audio_id+'.csv'
b@146: 
b@290:                     # If file hasn't been opened yet this time, empty
b@290:                     if csv_name not in file_history:
b@290:                         csvfile = open(csv_name, 'w')
b@290:                         file_history.append(csv_name) # remember this file has been written to this time around
b@290:                     else: 
b@290:                         # append (!) to file [page_name]/[page_name]-comments-[id].csv
b@290:                         csvfile = open(csv_name, 'a')
b@290:                     writer = csv.writer(csvfile, 
b@290:                                         delimiter=',', 
b@290:                                         dialect="excel",
b@290:                                         quoting=csv.QUOTE_ALL)
b@290:                     commentstr = audioelement.find("./comment/response").text
b@245:                         
b@290:                     if commentstr is None:
b@290:                        commentstr = ''
b@290:                         
b@290:                     # anonymous comments:
b@290:                     #writer.writerow([commentstr.encode("utf-8")]) 
b@290:                     # comments with (file) name:
b@290:                     writer.writerow([file[:-4]] + [commentstr.encode("utf-8")]) 
b@146: 
b@290:                     #TODO Replace 'new line' in comment with something else?
b@230:                         
b@230: # PRO TIP: Change from csv to txt by running this in bash: 
b@230: # $ cd folder_where_csvs_are/
b@230: # $ for i in *.csv; do mv "$i" "${i/.csv}".txt; done