BrechtDeMan@878: #!/usr/bin/python
BrechtDeMan@880: # -*- coding: utf-8 -*-
BrechtDeMan@878: 
BrechtDeMan@927: import xml.etree.ElementTree as ET
BrechtDeMan@927: import os
BrechtDeMan@927: import csv
BrechtDeMan@1072: import sys
BrechtDeMan@1066: 
BrechtDeMan@1066: # COMMAND LINE ARGUMENTS
BrechtDeMan@1066: 
BrechtDeMan@1066: assert len(sys.argv)<3, "comment_parser takes at most 1 command line argument\n"+\
BrechtDeMan@1066:                         "Use: python score_parser.py [rating_folder_location]"
BrechtDeMan@1066: 
BrechtDeMan@1066: # XML results files location
BrechtDeMan@1066: if len(sys.argv) == 1:
BrechtDeMan@1066:     folder_name = "../saves"    # Looks in 'saves/' folder from 'scripts/' folder
BrechtDeMan@1066:     print "Use: python  comment_parser.py [XML_files_location]"
BrechtDeMan@1066:     print "Using default path: " + folder_name
BrechtDeMan@1066: elif len(sys.argv) == 2:
BrechtDeMan@1066:     folder_name = sys.argv[1]   # First command line argument is folder
BrechtDeMan@1066: 
BrechtDeMan@1066: # check if folder_name exists
BrechtDeMan@1066: if not os.path.exists(folder_name):
BrechtDeMan@1066:     #the file is not there
BrechtDeMan@1066:     print "Folder '"+folder_name+"' does not exist."
BrechtDeMan@1066:     sys.exit() # terminate script execution
BrechtDeMan@1066: elif not os.access(os.path.dirname(folder_name), os.W_OK):
BrechtDeMan@1066:     #the file does exist but write privileges are not given
BrechtDeMan@1066:     print "No write privileges in folder '"+folder_name+"'."
BrechtDeMan@1066: 
BrechtDeMan@1066: 
BrechtDeMan@1066: # CODE
BrechtDeMan@883: 
BrechtDeMan@1072: # remember which files have been opened this time
BrechtDeMan@1072: file_history = []
BrechtDeMan@1072: 
BrechtDeMan@927: # get every XML file in folder
BrechtDeMan@883: for file in os.listdir(folder_name): 
BrechtDeMan@927:     if file.endswith(".xml"):
BrechtDeMan@883:         tree = ET.parse(folder_name + '/' + file)
BrechtDeMan@927:         root = tree.getroot()
BrechtDeMan@1051:         
BrechtDeMan@928:         # get list of all page names
n@1109:         for audioholder in root.findall("./page"):   # iterate over pages
BrechtDeMan@928:             page_name = audioholder.get('id')               # get page name
BrechtDeMan@1030:             
BrechtDeMan@1030:             if page_name is None: # ignore 'empty' audio_holders
n@1109:                 print "WARNING: " + file + " contains empty page. (comment_parser.py)"
BrechtDeMan@1030:                 break
BrechtDeMan@927: 
BrechtDeMan@928:             # create folder [page_name] if not yet created
BrechtDeMan@883:             if not os.path.exists(folder_name + "/" + page_name):
BrechtDeMan@883:                 os.makedirs(folder_name + "/" + page_name)
BrechtDeMan@927: 
BrechtDeMan@928:             # for page [page_name], print comments related to fragment [id]
BrechtDeMan@883:             for audioelement in audioholder.findall("./audioelement"):
BrechtDeMan@1030:                 if audioelement is not None: # Check it exists
BrechtDeMan@1030:                     audio_id = str(audioelement.get('id'))
BrechtDeMan@878:                     
BrechtDeMan@883:                     csv_name = folder_name +'/' + page_name+'/'+page_name+'-comments-'+audio_id+'.csv'
BrechtDeMan@927: 
BrechtDeMan@1072:                     # If file hasn't been opened yet this time, empty
BrechtDeMan@1072:                     if csv_name not in file_history:
BrechtDeMan@1072:                         csvfile = open(csv_name, 'w')
BrechtDeMan@1072:                         file_history.append(csv_name) # remember this file has been written to this time around
BrechtDeMan@1072:                     else: 
BrechtDeMan@1072:                         # append (!) to file [page_name]/[page_name]-comments-[id].csv
BrechtDeMan@1072:                         csvfile = open(csv_name, 'a')
BrechtDeMan@1072:                     writer = csv.writer(csvfile, 
BrechtDeMan@1072:                                         delimiter=',', 
BrechtDeMan@1072:                                         dialect="excel",
BrechtDeMan@1072:                                         quoting=csv.QUOTE_ALL)
BrechtDeMan@1072:                     commentstr = audioelement.find("./comment/response").text
BrechtDeMan@883:                         
BrechtDeMan@1072:                     if commentstr is None:
BrechtDeMan@1072:                        commentstr = ''
BrechtDeMan@1072:                         
BrechtDeMan@1072:                     # anonymous comments:
BrechtDeMan@1072:                     #writer.writerow([commentstr.encode("utf-8")]) 
BrechtDeMan@1072:                     # comments with (file) name:
BrechtDeMan@1072:                     writer.writerow([file[:-4]] + [commentstr.encode("utf-8")]) 
BrechtDeMan@927: 
BrechtDeMan@1072:                     #TODO Replace 'new line' in comment with something else?
BrechtDeMan@878:                         
BrechtDeMan@878: # PRO TIP: Change from csv to txt by running this in bash: 
BrechtDeMan@878: # $ cd folder_where_csvs_are/
BrechtDeMan@878: # $ for i in *.csv; do mv "$i" "${i/.csv}".txt; done