annotate scripts/comment_parser.py @ 147:927d05a43a70
Score parsing: make csv from all XML files (one file per page ID, one column per fragment ID, one row per subject). Supports varying selections of pages and fragments across subjects. Generalised to 'page' (instead of 'song') and 'fragment' (instead of 'mix').
| author |
Brecht De Man <b.deman@qmul.ac.uk> |
| date |
Sun, 31 May 2015 14:45:30 +0100 |
| parents |
2d08d2025258 |
| children |
97ebdb6b266a |
| rev |
line source |
|
b@146
|
1 import xml.etree.ElementTree as ET
|
|
b@146
|
2 import os
|
|
b@146
|
3 import csv
|
|
b@146
|
4
|
|
b@146
|
5 # get every XML file in folder
|
|
b@146
|
6 for file in os.listdir("."): # You have to put this script in folder where output XML files are.
|
|
b@146
|
7 if file.endswith(".xml"):
|
|
b@146
|
8 tree = ET.parse(file)
|
|
b@146
|
9 root = tree.getroot()
|
|
b@146
|
10
|
|
b@147
|
11 # get list of all page names
|
|
b@147
|
12 for audioholder in root.findall("./audioholder"): # iterate over pages
|
|
b@147
|
13 page_name = audioholder.get('id') # get page name
|
|
b@146
|
14
|
|
b@147
|
15 # create folder [page_name] if not yet created
|
|
b@147
|
16 if not os.path.exists(page_name):
|
|
b@147
|
17 os.makedirs(page_name)
|
|
b@146
|
18
|
|
b@147
|
19 # for page [page_name], print comments related to fragment [id]
|
|
b@147
|
20 for audioelement in root.findall("*/[@id='"+page_name+"']/audioelement"):
|
|
b@146
|
21 audio_id = str(audioelement.get('id'))
|
|
b@147
|
22 # append to file [page_name]/[page_name]-comments-[id].csv
|
|
b@147
|
23 with open(page_name+'/'+page_name+'-comments-'+audio_id+'.csv', 'a') as csvfile:
|
|
b@146
|
24 commentstr = root.find("*/[@id='"
|
|
b@147
|
25 + page_name
|
|
b@146
|
26 + "']/audioelement/[@id='"
|
|
b@146
|
27 + audio_id
|
|
b@146
|
28 + "']/comment/response").text
|
|
b@146
|
29 writer = csv.writer(csvfile, delimiter=',')
|
|
b@146
|
30 writer.writerow([commentstr.encode("utf-8")])
|
|
b@146
|
31 #TODO Comma doesn't act as delimiter now!
|
|
b@146
|
32 # (when adding more than just a comment per line):
|
|
b@146
|
33 # writer.writerow([file + ',' + commentstr.encode("utf-8")])
|
|
b@146
|
34
|
|
b@146
|
35 #TODO Replace 'new line' with something else?
|
|
b@146
|
36
|
|
b@146
|
37 #TODO 'Append' means duplicate entries if run several times... |
