| Index: trunk/tools/editor_trends/kaggle/training.py |
| — | — | @@ -0,0 +1,25 @@ |
| | 2 | +import codecs |
| | 3 | +import os |
| | 4 | + |
| | 5 | + |
| | 6 | + |
| | 7 | +location = '/home/diederik/wikimedia/wikilytics/en/wiki/txt' |
| | 8 | +files = os.listdir(location) |
| | 9 | + |
| | 10 | +output = codecs.open('training.txt', 'w', 'utf-8') |
| | 11 | + |
| | 12 | +for filename in files: |
| | 13 | + fh = codecs.open(os.path.join(location, filename)) |
| | 14 | + for line in fh: |
| | 15 | + line = line.strip() |
| | 16 | + line = line.split('\t') |
| | 17 | + if len(line) != 13: |
| | 18 | + continue |
| | 19 | + username = line[12].lower() |
| | 20 | + if username.endswith('bot'): |
| | 21 | + line[5] = 1 |
| | 22 | + line = '\t'.join(line) |
| | 23 | + output.write(line) |
| | 24 | + |
| | 25 | + |
| | 26 | +output.close() |
| \ No newline at end of file |
| Property changes on: trunk/tools/editor_trends/kaggle/training.py |
| ___________________________________________________________________ |
| Added: svn:eol-style |
| 1 | 27 | + native |