3 # original wikiq headers are: title articleid revid date_time anon
4 # editor editor_id minor text_size text_entropy text_md5 reversion
5 # additions_size deletions_size
12 from subprocess import Popen, PIPE
13 from collections import deque
14 from hashlib import sha1
16 from mw.xml_dump import Iterator
17 from mw.lib import persistence
18 from mw.lib import reverts
22 def calculate_persistence(tokens_added):
23 return(sum([(len(x.revisions)-1) for x in tokens_added]),
26 class WikiqIterator():
27 def __init__(self, fh, collapse_user=False):
29 self.collapse_user = collapse_user
30 self.mwiterator = Iterator.from_file(self.fh)
31 self.__pages = self.load_pages()
34 for page in self.mwiterator:
35 yield WikiqPage(page, collapse_user=self.collapse_user)
41 return next(self._pages)
44 __slots__ = ('id', 'title', 'namespace', 'redirect',
45 'restrictions', 'mwpage', '__revisions',
48 def __init__(self, page, collapse_user=False):
50 self.title = page.title
51 self.namespace = page.namespace
52 self.redirect = page.redirect
53 self.restrictions = page.restrictions
55 self.collapse_user = collapse_user
57 self.__revisions = self.rev_list()
60 # Outline for how we want to handle collapse_user=True
61 # iteration rev.user prev_rev.user add prev_rev?
68 for i, rev in enumerate(self.mwpage):
69 # never yield the first time
71 if self.collapse_user:
73 rev.collapsed_revs = collapsed_revs
76 if self.collapse_user:
77 # yield if this is the last edit in a seq by a user and reset
78 if not rev.contributor.user_text == prev_rev.contributor.user_text:
81 rev.collapsed_revs = collapsed_revs
82 # otherwise, add one to the counter
85 rev.collapsed_revs = collapsed_revs
86 # if collapse_user is false, we always yield
91 # also yield the final time
95 return self.__revisions
98 return next(self.__revisions)
101 def __init__(self, input_file, output_file, collapse_user=False, persist=False):
102 self.input_file = input_file
103 self.output_file = output_file
104 self.collapse_user = collapse_user
105 self.persist = persist
106 self.printed_header = False
109 def __get_namespace_from_title(self, title):
112 for ns in self.namespaces:
113 # skip if the namespace is not defined
115 default_ns = self.namespaces[ns]
118 if title.startswith(ns + ":"):
119 return self.namespaces[ns]
121 # if we've made it this far with no matches, we return the default namespace
125 print("Processing file: %s" % self.input_file.name, file=sys.stderr)
127 # create a regex that creates the output filename
128 # output_filename = re.sub(r'^.*/(enwiki\-\d+)\-.*p(\d+)p.*$',
129 # r'output/wikiq-\1-\2.tsv',
132 # Construct dump file iterator
133 dump = WikiqIterator(self.input_file, collapse_user=self.collapse_user)
135 # extract list of namspaces
136 self.namespaces = {ns.name : ns.id for ns in dump.mwiterator.namespaces}
140 # Iterate through pages
143 state = persistence.State()
144 window = deque(maxlen=PERSISTENCE_RADIUS)
146 rev_detector = reverts.Detector()
148 # Iterate through a page's revisions
151 rev_data = {'revid' : rev.id,
152 'date_time' : rev.timestamp.strftime('%Y-%m-%d %H:%M:%S'),
153 'articleid' : page.id,
154 'editor_id' : "" if rev.contributor.id == None else rev.contributor.id,
155 'title' : '"' + page.title + '"',
156 'namespace' : page.namespace if page.namespace else self.__get_namespace_from_title(page.title),
157 'deleted' : "TRUE" if rev.text.deleted else "FALSE" }
159 # if revisions are deleted, /many/ things will be missing
161 rev_data['text_chars'] = ""
162 rev_data['sha1'] = ""
163 rev_data['revert'] = ""
164 rev_data['reverteds'] = ""
167 # if text exists, we'll check for a sha1 and generate one otherwise
171 text_sha1 = sha1(bytes(rev.text, "utf8")).hexdigest()
173 rev_data['sha1'] = text_sha1
175 # TODO rev.bytes doesn't work.. looks like a bug
176 rev_data['text_chars'] = len(rev.text)
178 # generate revert data
179 revert = rev_detector.process(text_sha1, rev.id)
181 rev_data['revert'] = "TRUE"
182 rev_data['reverteds'] = '"' + ",".join([str(x) for x in revert.reverteds]) + '"'
184 rev_data['revert'] = "FALSE"
185 rev_data['reverteds'] = ""
187 # if the fact that the edit was minor can be hidden, this might be an issue
188 rev_data['minor'] = "TRUE" if rev.minor else "FALSE"
190 if rev.contributor.user_text:
191 # wrap user-defined editors in quotes for fread
192 rev_data['editor'] = '"' + rev.contributor.user_text + '"'
193 rev_data['anon'] = "TRUE" if rev.contributor.id == None else "FALSE"
196 rev_data['anon'] = ""
197 rev_data['editor'] = ""
199 #if re.match(r'^#redirect \[\[.*\]\]', rev.text, re.I):
204 #TODO missing: additions_size deletions_size
206 # if collapse user was on, lets run that
207 if self.collapse_user:
208 rev_data['collapsed_revs'] = rev.collapsed_revs
212 for k in ["token_revs", "tokens_added", "tokens_removed", "tokens_window"]:
213 old_rev_data[k] = None
215 _, tokens_added, tokens_removed = state.process(rev.text, rev.id, text_sha1)
216 window.append((rev.id, rev_data, tokens_added, tokens_removed))
218 if len(window) == PERSISTENCE_RADIUS:
219 old_rev_id, old_rev_data, old_tokens_added, old_tokens_removed = window[0]
221 num_token_revs, num_tokens = calculate_persistence(old_tokens_added)
223 old_rev_data["token_revs"] = num_token_revs
224 old_rev_data["tokens_added"] = num_tokens
225 old_rev_data["tokens_removed"] = len(old_tokens_removed)
226 old_rev_data["tokens_window"] = PERSISTENCE_RADIUS-1
228 self.print_rev_data(old_rev_data)
231 self.print_rev_data(rev_data)
236 # print out metadata for the last RADIUS revisions
237 for i, item in enumerate(window):
238 # if the window was full, we've already printed item 0
239 if len(window) == PERSISTENCE_RADIUS and i == 0:
242 rev_id, rev_data, tokens_added, tokens_removed = item
243 num_token_revs, num_tokens = calculate_persistence(tokens_added)
245 rev_data["token_revs"] = num_token_revs
246 rev_data["tokens_added"] = num_tokens
247 rev_data["tokens_removed"] = len(tokens_removed)
248 rev_data["tokens_window"] = len(window)-(i+1)
250 self.print_rev_data(rev_data)
254 print("Done: %s revisions and %s pages." % (rev_count, page_count),
257 def print_rev_data(self, rev_data):
258 # if it's the first time through, print the header
259 if not self.printed_header:
260 print("\t".join([str(k) for k in sorted(rev_data.keys())]), file=self.output_file)
261 self.printed_header = True
263 print("\t".join([str(v) for k, v in sorted(rev_data.items())]), file=self.output_file)
266 def open_input_file(input_filename):
267 if re.match(r'.*\.7z', input_filename):
268 cmd = ["7za", "x", "-so", input_filename]
269 elif re.match(r'.*\.gz', input_filename):
270 cmd = ["zcat", input_filename]
271 elif re.match(r'.*\.bz2', input_filename):
272 cmd = ["bzcat", input_filename]
275 input_file = Popen(cmd, stdout=PIPE).stdout
277 input_file = open(input_filename, 'r')
281 def open_output_file(input_filename):
282 # create a regex that creates the output filename
283 output_filename = re.sub(r'\.xml(\.(7z|gz|bz2))?$', '', input_filename)
284 output_filename = output_filename + ".tsv"
285 output_file = open(output_filename, "w")
289 parser = argparse.ArgumentParser(description='Parse MediaWiki XML database dumps into tab delimitted data.')
291 # arguments for the input direction
292 parser.add_argument('dumpfiles', metavar="DUMPFILE", nargs="*", type=str,
293 help="Filename of the compressed or uncompressed XML database dump. If absent, we'll look for content on stdin and output on stdout.")
295 parser.add_argument('-o', '--output-dir', metavar='DIR', dest='output_dir', type=str, nargs=1,
296 help="Directory for output files.")
298 parser.add_argument('-s', '--stdout', dest="stdout", action="store_true",
299 help="Write output to standard out (do not create dump file)")
301 parser.add_argument('--collapse-user', dest="collapse_user", action="store_true",
302 help="Operate only on the final revision made by user a user within all sequences of consecutive edits made by a user. This can be useful for addressing issues with text persistence measures.")
304 parser.add_argument('-p', '--persistence', dest="persist", action="store_true",
305 help="Compute and report measures of content persistent: (1) persistent token revisions, (2) tokens added, and (3) number of revision used in computing the first measure.")
307 args = parser.parse_args()
309 if len(args.dumpfiles) > 0:
310 for filename in args.dumpfiles:
311 input_file = open_input_file(filename)
313 # open file for output
315 output_file = sys.stdout
318 output_dir = args.output_dir[0]
322 filename = os.path.join(output_dir, os.path.basename(filename))
323 output_file = open_output_file(filename)
325 wikiq = WikiqParser(input_file, output_file,
326 collapse_user=args.collapse_user,
327 persist=args.persist)
334 wikiq = WikiqParser(sys.stdin, sys.stdout,
335 collapse_user=args.collapse_user,
336 persist=args.persist)
339 # stop_words = "a,able,about,across,after,all,almost,also,am,among,an,and,any,are,as,at,be,because,been,but,by,can,cannot,could,dear,did,do,does,either,else,ever,every,for,from,get,got,had,has,have,he,her,hers,him,his,how,however,i,if,in,into,is,it,its,just,least,let,like,likely,may,me,might,most,must,my,neither,no,nor,not,of,off,often,on,only,or,other,our,own,rather,said,say,says,she,should,since,so,some,than,that,the,their,them,then,there,these,they,this,tis,to,too,twas,us,wants,was,we,were,what,when,where,which,while,who,whom,why,will,with,would,yet,you,your"
340 # stop_words = stop_words.split(",")