| Index: trunk/tools/editor_trends/etl/extracter.py |
| — | — | @@ -1,283 +1,283 @@ |
| 2 | | -#!/usr/bin/python
|
| 3 | | -# -*- coding: utf-8 -*-
|
| 4 | | -'''
|
| 5 | | -Copyright (C) 2010 by Diederik van Liere (dvanliere@gmail.com)
|
| 6 | | -This program is free software; you can redistribute it and/or
|
| 7 | | -modify it under the terms of the GNU General Public License version 2
|
| 8 | | -as published by the Free Software Foundation.
|
| 9 | | -This program is distributed in the hope that it will be useful,
|
| 10 | | -but WITHOUT ANY WARRANTY; without even the implied warranty of
|
| 11 | | -MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
|
| 12 | | -See the GNU General Public License for more details, at
|
| 13 | | -http://www.fsf.org/licenses/gpl.html
|
| 14 | | -'''
|
| 15 | | -
|
| 16 | | -__author__ = '''\n'''.join(['Diederik van Liere (dvanliere@gmail.com)', ])
|
| 17 | | -__author__email = 'dvanliere at gmail dot com'
|
| 18 | | -__date__ = '2010-12-13'
|
| 19 | | -__version__ = '0.1'
|
| 20 | | -
|
| 21 | | -import sys
|
| 22 | | -import re
|
| 23 | | -import json
|
| 24 | | -import os
|
| 25 | | -import xml.etree.cElementTree as cElementTree
|
| 26 | | -
|
| 27 | | -sys.path.append('..')
|
| 28 | | -import configuration
|
| 29 | | -settings = configuration.Settings()
|
| 30 | | -
|
| 31 | | -import wikitree.parser
|
| 32 | | -from bots import bots
|
| 33 | | -from utils import utils
|
| 34 | | -
|
| 35 | | -try:
|
| 36 | | - import psyco
|
| 37 | | - psyco.full()
|
| 38 | | -except ImportError:
|
| 39 | | - pass
|
| 40 | | -
|
| 41 | | -
|
| 42 | | -RE_NUMERIC_CHARACTER = re.compile('&#(\d+);')
|
| 43 | | -
|
| 44 | | -
|
| 45 | | -def remove_numeric_character_references(text):
|
| 46 | | - return re.sub(RE_NUMERIC_CHARACTER, lenient_deccharref, text).encode('utf-8')
|
| 47 | | -
|
| 48 | | -
|
| 49 | | -def lenient_deccharref(m):
|
| 50 | | - try:
|
| 51 | | - return unichr(int(m.group(1)))
|
| 52 | | - except ValueError:
|
| 53 | | - '''
|
| 54 | | - There are a few articles that raise a Value Error here, the reason is
|
| 55 | | - that I am using a narrow Python build (UCS2) instead of a wide build
|
| 56 | | - (UCS4). The quick fix is to return an empty string...
|
| 57 | | - Real solution is to rebuild Python with UCS4 support.....
|
| 58 | | - '''
|
| 59 | | - return ''
|
| 60 | | -
|
| 61 | | -
|
| 62 | | -def remove_namespace(element, namespace):
|
| 63 | | - '''Remove namespace from the XML document.'''
|
| 64 | | - ns = u'{%s}' % namespace
|
| 65 | | - nsl = len(ns)
|
| 66 | | - for elem in element.getiterator():
|
| 67 | | - if elem.tag.startswith(ns):
|
| 68 | | - elem.tag = elem.tag[nsl:]
|
| 69 | | - return element
|
| 70 | | -
|
| 71 | | -
|
| 72 | | -def load_namespace(language):
|
| 73 | | - file = '%s_ns.json' % language
|
| 74 | | - fh = utils.create_txt_filehandle(settings.namespace_location, file, 'r', settings.encoding)
|
| 75 | | - ns = json.load(fh)
|
| 76 | | - fh.close()
|
| 77 | | - ns = ns['query']['namespaces']
|
| 78 | | - return ns
|
| 79 | | -
|
| 80 | | -
|
| 81 | | -def build_namespaces_locale(namespaces, include=['0']):
|
| 82 | | - '''
|
| 83 | | - @include is a list of namespace keys that should not be ignored, the default
|
| 84 | | - setting is to ignore all namespaces except the main namespace.
|
| 85 | | - '''
|
| 86 | | - ns = []
|
| 87 | | - for namespace in namespaces:
|
| 88 | | - if namespace not in include:
|
| 89 | | - value = namespaces[namespace].get(u'*', None)
|
| 90 | | - ns.append(value)
|
| 91 | | - return ns
|
| 92 | | -
|
| 93 | | -
|
| 94 | | -def parse_comments(revisions, function):
|
| 95 | | - for revision in revisions:
|
| 96 | | - comment = revision.find('{%s}comment' % settings.xml_namespace)
|
| 97 | | - #timestamp = revision.find('{%s}timestamp' % settings.xml_namespace).text
|
| 98 | | - if comment != None and comment.text != None:
|
| 99 | | - comment.text = function(comment.text)
|
| 100 | | - return revisions
|
| 101 | | -
|
| 102 | | -
|
| 103 | | -def is_article_main_namespace(elem, namespace):
|
| 104 | | - '''
|
| 105 | | - checks whether the article belongs to the main namespace
|
| 106 | | - '''
|
| 107 | | - title = elem.text
|
| 108 | | - for ns in namespace:
|
| 109 | | - if title.startswith(ns):
|
| 110 | | - return False
|
| 111 | | - return True
|
| 112 | | -
|
| 113 | | -def validate_hostname(address):
|
| 114 | | - '''
|
| 115 | | - This is not a foolproof solution at all. The problem is that it's really hard
|
| 116 | | - to determine whether a string is a hostname or not **reliably**. This is a
|
| 117 | | - very fast rule of thumb. Will lead to false positives, but that's life :)
|
| 118 | | - '''
|
| 119 | | - parts = address.split(".")
|
| 120 | | - if len(parts) > 2:
|
| 121 | | - return True
|
| 122 | | - else:
|
| 123 | | - return False
|
| 124 | | -
|
| 125 | | -
|
| 126 | | -def validate_ip(address):
|
| 127 | | - parts = address.split(".")
|
| 128 | | - if len(parts) != 4:
|
| 129 | | - return False
|
| 130 | | - parts = parts[:3]
|
| 131 | | - for item in parts:
|
| 132 | | - try:
|
| 133 | | - if not 0 <= int(item) <= 255:
|
| 134 | | - return False
|
| 135 | | - except ValueError:
|
| 136 | | - return False
|
| 137 | | - return True
|
| 138 | | -
|
| 139 | | -
|
| 140 | | -def determine_username_is_bot(contributor, **kwargs):
|
| 141 | | - '''
|
| 142 | | - #contributor is an xml element containing the id of the contributor
|
| 143 | | - @bots should have a dict with all the bot ids and bot names
|
| 144 | | - @Return False if username id is not in bot dict id or True if username id
|
| 145 | | - is a bot id.
|
| 146 | | - '''
|
| 147 | | - bots = kwargs.get('bots')
|
| 148 | | - username = contributor.find('username')
|
| 149 | | - if username == None:
|
| 150 | | - return 0
|
| 151 | | - else:
|
| 152 | | - if username in bots:
|
| 153 | | - return 1
|
| 154 | | - else:
|
| 155 | | - return 0
|
| 156 | | -
|
| 157 | | -
|
| 158 | | -def extract_username(contributor, **kwargs):
|
| 159 | | - contributor = contributor.find('username')
|
| 160 | | - if contributor != None:
|
| 161 | | - return contributor.text
|
| 162 | | - else:
|
| 163 | | - return None
|
| 164 | | -
|
| 165 | | -
|
| 166 | | -def extract_contributor_id(contributor, **kwargs):
|
| 167 | | - '''
|
| 168 | | - @contributor is the xml contributor node containing a number of attributes
|
| 169 | | - Currently, we are only interested in registered contributors, hence we
|
| 170 | | - ignore anonymous editors.
|
| 171 | | - '''
|
| 172 | | - if contributor.get('deleted'):
|
| 173 | | - return None # ASK: Not sure if this is the best way to code deleted contributors.
|
| 174 | | - elem = contributor.find('id')
|
| 175 | | - if elem != None:
|
| 176 | | - return {'id':elem.text}
|
| 177 | | - else:
|
| 178 | | - elem = contributor.find('ip')
|
| 179 | | - if elem != None and elem.text != None and validate_ip(elem.text) == False and validate_hostname(elem.text) == False:
|
| 180 | | - return {'username':elem.text, 'id': elem.text}
|
| 181 | | - else:
|
| 182 | | - return None
|
| 183 | | -
|
| 184 | | -
|
| 185 | | -def output_editor_information(revisions, page, bots):
|
| 186 | | - '''
|
| 187 | | - @elem is an XML element containing 1 revision from a page
|
| 188 | | - @output is where to store the data, a filehandle
|
| 189 | | - @**kwargs contains extra information
|
| 190 | | -
|
| 191 | | - the variable tags determines which attributes are being parsed, the values in
|
| 192 | | - this dictionary are the functions used to extract the data.
|
| 193 | | - '''
|
| 194 | | - headers = ['id', 'date', 'article', 'username']
|
| 195 | | - tags = {'contributor': {'id': extract_contributor_id,
|
| 196 | | - 'bot': determine_username_is_bot,
|
| 197 | | - 'username': extract_username,
|
| 198 | | - },
|
| 199 | | - 'timestamp': {'date': wikitree.parser.extract_text},
|
| 200 | | - }
|
| 201 | | - vars = {}
|
| 202 | | - flat = []
|
| 203 | | -
|
| 204 | | - for x, revision in enumerate(revisions):
|
| 205 | | - #print len(revision.getchildren())
|
| 206 | | - vars[x] = {}
|
| 207 | | - vars[x]['article'] = page
|
| 208 | | - for tag in tags:
|
| 209 | | - el = revision.find('%s' % tag)
|
| 210 | | - if el == None:
|
| 211 | | - #print cElementTree.tostring(revision, settings.encoding)
|
| 212 | | - del vars[x]
|
| 213 | | - break
|
| 214 | | - for function in tags[tag].keys():
|
| 215 | | - f = tags[tag][function]
|
| 216 | | - value = f(el, bots=bots)
|
| 217 | | - if type(value) == type({}):
|
| 218 | | - for kw in value:
|
| 219 | | - vars[x][kw] = value[kw]
|
| 220 | | - else:
|
| 221 | | - vars[x][function] = value
|
| 222 | | -
|
| 223 | | - '''
|
| 224 | | - This loop determines for each observation whether it should be stored or not.
|
| 225 | | - '''
|
| 226 | | - for x in vars:
|
| 227 | | - if vars[x]['bot'] == 1 or vars[x]['id'] == None or vars[x]['username'] == None:
|
| 228 | | - continue
|
| 229 | | - else:
|
| 230 | | - f = []
|
| 231 | | - for head in headers:
|
| 232 | | - f.append(vars[x][head])
|
| 233 | | - flat.append(f)
|
| 234 | | -
|
| 235 | | - return flat
|
| 236 | | -
|
| 237 | | -
|
| 238 | | -def parse_dumpfile(project, language_code, namespaces=['0']):
|
| 239 | | - bot_ids = bots.retrieve_bots(language_code)
|
| 240 | | - ns = load_namespace(language_code)
|
| 241 | | - ns = build_namespaces_locale(ns, namespaces)
|
| 242 | | -
|
| 243 | | - location = os.path.join(settings.input_location, language_code, project)
|
| 244 | | - fh = utils.create_txt_filehandle(location, 'enwiki-latest-stub-meta-history.xml', 'r', settings.encoding)
|
| 245 | | - for page in wikitree.parser.read_input(fh):
|
| 246 | | - title = page.find('title')
|
| 247 | | - if is_article_main_namespace(title, ns):
|
| 248 | | - #cElementTree.dump(page)
|
| 249 | | - article_id = page.find('id').text
|
| 250 | | - revisions = page.findall('revision')
|
| 251 | | - revisions = parse_comments(revisions, remove_numeric_character_references)
|
| 252 | | - output = output_editor_information(revisions, article_id, bot_ids)
|
| 253 | | - write_output(output, project, language_code)
|
| 254 | | - page.clear()
|
| 255 | | - fh.close()
|
| 256 | | -
|
| 257 | | -
|
| 258 | | -def write_output(output, project, language_code):
|
| 259 | | - location = os.path.join(settings.input_location, language_code, project, 'txt')
|
| 260 | | - for o in output:
|
| 261 | | - file = '%s.csv' % hash(o[0])
|
| 262 | | - try:
|
| 263 | | - fh = utils.create_txt_filehandle(location, file, 'a', settings.encoding)
|
| 264 | | - utils.write_list_to_csv(o, fh)
|
| 265 | | - fh.close()
|
| 266 | | - except Exception, error:
|
| 267 | | - print error
|
| 268 | | -
|
| 269 | | -
|
| 270 | | -def hash(id):
|
| 271 | | - '''
|
| 272 | | - A very simple hash function based on modulo. The except clause has been
|
| 273 | | - addde because there are instances where the username is stored in userid
|
| 274 | | - tag and hence that's a string and not an integer.
|
| 275 | | - '''
|
| 276 | | - try:
|
| 277 | | - return int(id) % 500
|
| 278 | | - except:
|
| 279 | | - return sum([ord(i) for i in id]) % 500
|
| 280 | | -
|
| 281 | | -if __name__ == '__main__':
|
| 282 | | - project = 'wiki'
|
| 283 | | - language_code = 'en'
|
| 284 | | - parse_dumpfile(project, language_code)
|
| | 2 | +#!/usr/bin/python |
| | 3 | +# -*- coding: utf-8 -*- |
| | 4 | +''' |
| | 5 | +Copyright (C) 2010 by Diederik van Liere (dvanliere@gmail.com) |
| | 6 | +This program is free software; you can redistribute it and/or |
| | 7 | +modify it under the terms of the GNU General Public License version 2 |
| | 8 | +as published by the Free Software Foundation. |
| | 9 | +This program is distributed in the hope that it will be useful, |
| | 10 | +but WITHOUT ANY WARRANTY; without even the implied warranty of |
| | 11 | +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. |
| | 12 | +See the GNU General Public License for more details, at |
| | 13 | +http://www.fsf.org/licenses/gpl.html |
| | 14 | +''' |
| | 15 | + |
| | 16 | +__author__ = '''\n'''.join(['Diederik van Liere (dvanliere@gmail.com)', ]) |
| | 17 | +__author__email = 'dvanliere at gmail dot com' |
| | 18 | +__date__ = '2010-12-13' |
| | 19 | +__version__ = '0.1' |
| | 20 | + |
| | 21 | +import sys |
| | 22 | +import re |
| | 23 | +import json |
| | 24 | +import os |
| | 25 | +import xml.etree.cElementTree as cElementTree |
| | 26 | + |
| | 27 | +sys.path.append('..') |
| | 28 | +import configuration |
| | 29 | +settings = configuration.Settings() |
| | 30 | + |
| | 31 | +import wikitree.parser |
| | 32 | +from bots import bots |
| | 33 | +from utils import utils |
| | 34 | + |
| | 35 | +try: |
| | 36 | + import psyco |
| | 37 | + psyco.full() |
| | 38 | +except ImportError: |
| | 39 | + pass |
| | 40 | + |
| | 41 | + |
| | 42 | +RE_NUMERIC_CHARACTER = re.compile('&#(\d+);') |
| | 43 | + |
| | 44 | + |
| | 45 | +def remove_numeric_character_references(text): |
| | 46 | + return re.sub(RE_NUMERIC_CHARACTER, lenient_deccharref, text).encode('utf-8') |
| | 47 | + |
| | 48 | + |
| | 49 | +def lenient_deccharref(m): |
| | 50 | + try: |
| | 51 | + return unichr(int(m.group(1))) |
| | 52 | + except ValueError: |
| | 53 | + ''' |
| | 54 | + There are a few articles that raise a Value Error here, the reason is |
| | 55 | + that I am using a narrow Python build (UCS2) instead of a wide build |
| | 56 | + (UCS4). The quick fix is to return an empty string... |
| | 57 | + Real solution is to rebuild Python with UCS4 support..... |
| | 58 | + ''' |
| | 59 | + return '' |
| | 60 | + |
| | 61 | + |
| | 62 | +def remove_namespace(element, namespace): |
| | 63 | + '''Remove namespace from the XML document.''' |
| | 64 | + ns = u'{%s}' % namespace |
| | 65 | + nsl = len(ns) |
| | 66 | + for elem in element.getiterator(): |
| | 67 | + if elem.tag.startswith(ns): |
| | 68 | + elem.tag = elem.tag[nsl:] |
| | 69 | + return element |
| | 70 | + |
| | 71 | + |
| | 72 | +def load_namespace(language): |
| | 73 | + file = '%s_ns.json' % language |
| | 74 | + fh = utils.create_txt_filehandle(settings.namespace_location, file, 'r', settings.encoding) |
| | 75 | + ns = json.load(fh) |
| | 76 | + fh.close() |
| | 77 | + ns = ns['query']['namespaces'] |
| | 78 | + return ns |
| | 79 | + |
| | 80 | + |
| | 81 | +def build_namespaces_locale(namespaces, include=['0']): |
| | 82 | + ''' |
| | 83 | + @include is a list of namespace keys that should not be ignored, the default |
| | 84 | + setting is to ignore all namespaces except the main namespace. |
| | 85 | + ''' |
| | 86 | + ns = [] |
| | 87 | + for namespace in namespaces: |
| | 88 | + if namespace not in include: |
| | 89 | + value = namespaces[namespace].get(u'*', None) |
| | 90 | + ns.append(value) |
| | 91 | + return ns |
| | 92 | + |
| | 93 | + |
| | 94 | +def parse_comments(revisions, function): |
| | 95 | + for revision in revisions: |
| | 96 | + comment = revision.find('{%s}comment' % settings.xml_namespace) |
| | 97 | + #timestamp = revision.find('{%s}timestamp' % settings.xml_namespace).text |
| | 98 | + if comment != None and comment.text != None: |
| | 99 | + comment.text = function(comment.text) |
| | 100 | + return revisions |
| | 101 | + |
| | 102 | + |
| | 103 | +def is_article_main_namespace(elem, namespace): |
| | 104 | + ''' |
| | 105 | + checks whether the article belongs to the main namespace |
| | 106 | + ''' |
| | 107 | + title = elem.text |
| | 108 | + for ns in namespace: |
| | 109 | + if title.startswith(ns): |
| | 110 | + return False |
| | 111 | + return True |
| | 112 | + |
| | 113 | +def validate_hostname(address): |
| | 114 | + ''' |
| | 115 | + This is not a foolproof solution at all. The problem is that it's really hard |
| | 116 | + to determine whether a string is a hostname or not **reliably**. This is a |
| | 117 | + very fast rule of thumb. Will lead to false positives, but that's life :) |
| | 118 | + ''' |
| | 119 | + parts = address.split(".") |
| | 120 | + if len(parts) > 2: |
| | 121 | + return True |
| | 122 | + else: |
| | 123 | + return False |
| | 124 | + |
| | 125 | + |
| | 126 | +def validate_ip(address): |
| | 127 | + parts = address.split(".") |
| | 128 | + if len(parts) != 4: |
| | 129 | + return False |
| | 130 | + parts = parts[:3] |
| | 131 | + for item in parts: |
| | 132 | + try: |
| | 133 | + if not 0 <= int(item) <= 255: |
| | 134 | + return False |
| | 135 | + except ValueError: |
| | 136 | + return False |
| | 137 | + return True |
| | 138 | + |
| | 139 | + |
| | 140 | +def determine_username_is_bot(contributor, **kwargs): |
| | 141 | + ''' |
| | 142 | + #contributor is an xml element containing the id of the contributor |
| | 143 | + @bots should have a dict with all the bot ids and bot names |
| | 144 | + @Return False if username id is not in bot dict id or True if username id |
| | 145 | + is a bot id. |
| | 146 | + ''' |
| | 147 | + bots = kwargs.get('bots') |
| | 148 | + username = contributor.find('username') |
| | 149 | + if username == None: |
| | 150 | + return 0 |
| | 151 | + else: |
| | 152 | + if username in bots: |
| | 153 | + return 1 |
| | 154 | + else: |
| | 155 | + return 0 |
| | 156 | + |
| | 157 | + |
| | 158 | +def extract_username(contributor, **kwargs): |
| | 159 | + contributor = contributor.find('username') |
| | 160 | + if contributor != None: |
| | 161 | + return contributor.text |
| | 162 | + else: |
| | 163 | + return None |
| | 164 | + |
| | 165 | + |
| | 166 | +def extract_contributor_id(contributor, **kwargs): |
| | 167 | + ''' |
| | 168 | + @contributor is the xml contributor node containing a number of attributes |
| | 169 | + Currently, we are only interested in registered contributors, hence we |
| | 170 | + ignore anonymous editors. |
| | 171 | + ''' |
| | 172 | + if contributor.get('deleted'): |
| | 173 | + return None # ASK: Not sure if this is the best way to code deleted contributors. |
| | 174 | + elem = contributor.find('id') |
| | 175 | + if elem != None: |
| | 176 | + return {'id':elem.text} |
| | 177 | + else: |
| | 178 | + elem = contributor.find('ip') |
| | 179 | + if elem != None and elem.text != None and validate_ip(elem.text) == False and validate_hostname(elem.text) == False: |
| | 180 | + return {'username':elem.text, 'id': elem.text} |
| | 181 | + else: |
| | 182 | + return None |
| | 183 | + |
| | 184 | + |
| | 185 | +def output_editor_information(revisions, page, bots): |
| | 186 | + ''' |
| | 187 | + @elem is an XML element containing 1 revision from a page |
| | 188 | + @output is where to store the data, a filehandle |
| | 189 | + @**kwargs contains extra information |
| | 190 | + |
| | 191 | + the variable tags determines which attributes are being parsed, the values in |
| | 192 | + this dictionary are the functions used to extract the data. |
| | 193 | + ''' |
| | 194 | + headers = ['id', 'date', 'article', 'username'] |
| | 195 | + tags = {'contributor': {'id': extract_contributor_id, |
| | 196 | + 'bot': determine_username_is_bot, |
| | 197 | + 'username': extract_username, |
| | 198 | + }, |
| | 199 | + 'timestamp': {'date': wikitree.parser.extract_text}, |
| | 200 | + } |
| | 201 | + vars = {} |
| | 202 | + flat = [] |
| | 203 | + |
| | 204 | + for x, revision in enumerate(revisions): |
| | 205 | + #print len(revision.getchildren()) |
| | 206 | + vars[x] = {} |
| | 207 | + vars[x]['article'] = page |
| | 208 | + for tag in tags: |
| | 209 | + el = revision.find('%s' % tag) |
| | 210 | + if el == None: |
| | 211 | + #print cElementTree.tostring(revision, settings.encoding) |
| | 212 | + del vars[x] |
| | 213 | + break |
| | 214 | + for function in tags[tag].keys(): |
| | 215 | + f = tags[tag][function] |
| | 216 | + value = f(el, bots=bots) |
| | 217 | + if type(value) == type({}): |
| | 218 | + for kw in value: |
| | 219 | + vars[x][kw] = value[kw] |
| | 220 | + else: |
| | 221 | + vars[x][function] = value |
| | 222 | + |
| | 223 | + ''' |
| | 224 | + This loop determines for each observation whether it should be stored or not. |
| | 225 | + ''' |
| | 226 | + for x in vars: |
| | 227 | + if vars[x]['bot'] == 1 or vars[x]['id'] == None or vars[x]['username'] == None: |
| | 228 | + continue |
| | 229 | + else: |
| | 230 | + f = [] |
| | 231 | + for head in headers: |
| | 232 | + f.append(vars[x][head]) |
| | 233 | + flat.append(f) |
| | 234 | + |
| | 235 | + return flat |
| | 236 | + |
| | 237 | + |
| | 238 | +def parse_dumpfile(project, language_code, namespaces=['0']): |
| | 239 | + bot_ids = bots.retrieve_bots(language_code) |
| | 240 | + ns = load_namespace(language_code) |
| | 241 | + ns = build_namespaces_locale(ns, namespaces) |
| | 242 | + |
| | 243 | + location = os.path.join(settings.input_location, language_code, project) |
| | 244 | + fh = utils.create_txt_filehandle(location, 'enwiki-latest-stub-meta-history.xml', 'r', settings.encoding) |
| | 245 | + for page in wikitree.parser.read_input(fh): |
| | 246 | + title = page.find('title') |
| | 247 | + if is_article_main_namespace(title, ns): |
| | 248 | + #cElementTree.dump(page) |
| | 249 | + article_id = page.find('id').text |
| | 250 | + revisions = page.findall('revision') |
| | 251 | + revisions = parse_comments(revisions, remove_numeric_character_references) |
| | 252 | + output = output_editor_information(revisions, article_id, bot_ids) |
| | 253 | + write_output(output, project, language_code) |
| | 254 | + page.clear() |
| | 255 | + fh.close() |
| | 256 | + |
| | 257 | + |
| | 258 | +def write_output(output, project, language_code): |
| | 259 | + location = os.path.join(settings.input_location, language_code, project, 'txt') |
| | 260 | + for o in output: |
| | 261 | + file = '%s.csv' % hash(o[0]) |
| | 262 | + try: |
| | 263 | + fh = utils.create_txt_filehandle(location, file, 'a', settings.encoding) |
| | 264 | + utils.write_list_to_csv(o, fh) |
| | 265 | + fh.close() |
| | 266 | + except Exception, error: |
| | 267 | + print error |
| | 268 | + |
| | 269 | + |
| | 270 | +def hash(id): |
| | 271 | + ''' |
| | 272 | + A very simple hash function based on modulo. The except clause has been |
| | 273 | + addde because there are instances where the username is stored in userid |
| | 274 | + tag and hence that's a string and not an integer. |
| | 275 | + ''' |
| | 276 | + try: |
| | 277 | + return int(id) % 500 |
| | 278 | + except: |
| | 279 | + return sum([ord(i) for i in id]) % 500 |
| | 280 | + |
| | 281 | +if __name__ == '__main__': |
| | 282 | + project = 'wiki' |
| | 283 | + language_code = 'en' |
| | 284 | + parse_dumpfile(project, language_code) |
| Property changes on: trunk/tools/editor_trends/etl/extracter.py |
| ___________________________________________________________________ |
| Added: svn:eol-style |
| 285 | 285 | + native |