d@0: #!/usr/bin/python
d@0: # -*- coding: utf-8 -*-
d@0: 
d@0: import xml.etree.ElementTree as ET
d@0: import os
d@0: import csv
d@0: 
d@0: # XML results files location (modify as needed):
d@0: folder_name = "../saves"    # Looks in 'saves/' folder from 'scripts/' folder
d@0: 
d@0: # get every XML file in folder
d@0: for file in os.listdir(folder_name): 
d@0:     if file.endswith(".xml"):
d@0:         tree = ET.parse(folder_name + '/' + file)
d@0:         root = tree.getroot()
d@0:         
d@0:         # get list of all page names
d@0:         for audioholder in root.findall("./audioholder"):   # iterate over pages
d@0:             page_name = audioholder.get('id')               # get page name
d@0:             
d@0:             if page_name is None: # ignore 'empty' audio_holders
d@0:                 print "WARNING: " + file + " contains empty audio holder. (comment_parser.py)"
d@0:                 break
d@0: 
d@0:             # create folder [page_name] if not yet created
d@0:             if not os.path.exists(folder_name + "/" + page_name):
d@0:                 os.makedirs(folder_name + "/" + page_name)
d@0: 
d@0:             # for page [page_name], print comments related to fragment [id]
d@0:             for audioelement in audioholder.findall("./audioelement"):
d@0:                 if audioelement is not None: # Check it exists
d@0:                     audio_id = str(audioelement.get('id'))
d@0:                     
d@0:                     csv_name = folder_name +'/' + page_name+'/'+page_name+'-comments-'+audio_id+'.csv'
d@0: 
d@0:                     # append (!) to file [page_name]/[page_name]-comments-[id].csv
d@0:                     with open(csv_name, 'a') as csvfile:
d@0:                         writer = csv.writer(csvfile, 
d@0:                                             delimiter=',', 
d@0:                                             dialect="excel",
d@0:                                             quoting=csv.QUOTE_ALL)
d@0:                         commentstr = audioelement.find("./comment/response").text
d@0:                         
d@0:                         if commentstr is None:
d@0:                            commentstr = '';
d@0:                             
d@0:                         # anonymous comments:
d@0:                         #writer.writerow([commentstr.encode("utf-8")]) 
d@0:                         # comments with (file) name:
d@0:                         writer.writerow([file[:-4]] + [commentstr.encode("utf-8")]) 
d@0: 
d@0:                         #TODO Replace 'new line' in comment with something else?
d@0:                         
d@0: # PRO TIP: Change from csv to txt by running this in bash: 
d@0: # $ cd folder_where_csvs_are/
d@0: # $ for i in *.csv; do mv "$i" "${i/.csv}".txt; done