]> code.communitydata.science - mediawiki_dump_tools.git/blobdiff - wikiq
refactor regex matching in a tidier object oriented style
[mediawiki_dump_tools.git] / wikiq
diff --git a/wikiq b/wikiq
index ae6b40f07970b08975afbd545827c09b5509d08a..7e75eda22228cd5c44c248c264c8435b3268a948 100755 (executable)
--- a/wikiq
+++ b/wikiq
@@ -34,6 +34,7 @@ def calculate_persistence(tokens_added):
     return(sum([(len(x.revisions)-1) for x in tokens_added]),
            len(tokens_added))
 
+
 class WikiqIterator():
     def __init__(self, fh, collapse_user=False):
         self.fh = fh
@@ -126,14 +127,70 @@ class WikiqPage():
     def __next__(self):
         return next(self.__revisions)
 
+
+class RegexPair(object):
+    def __init__(self, pattern, label):
+        self.pattern = re.compile(pattern)
+        self.label = label
+        self.has_groups = bool(self.pattern.groupindex)
+        if self.has_groups:
+            self.capture_groups = list(self.pattern.groupindex.keys())
+            
+    def _make_key(self, cap_group):
+        return ("{}_{}".format(self.label, cap_group))
+
+    def matchmake(self, content, rev_data):
+        
+        temp_dict = {}
+        # if there are named capture groups in the regex
+        if self.has_groups:
+            # initialize the {capture_group_name:list} for each capture group
+            # if there are matches of some sort in this revision content, fill the lists for each cap_group
+            if self.pattern.search(content) is not None:
+                m = self.pattern.finditer(content)
+                matchobjects = list(m)
+
+                for cap_group in self.capture_groups:
+                    key = self._make_key(cap_group)
+                    temp_list = []
+                    for match in matchobjects:
+                        # we only want to add the match for the capture group if the match is not None
+                        if match.group(cap_group) != None:
+                            temp_list.append(match.group(cap_group))
+
+                    # if temp_list of matches is empty just make that column None
+                    if len(temp_list)==0:
+                        temp_dict[key] = None
+                    # else we put in the list we made in the for-loop above
+                    else:
+                        temp_dict[key] = ', '.join(temp_list)
+
+            # there are no matches at all in this revision content, we default values to None
+            else:
+                for cap_group in self.capture_groups:
+                    key = self._make_key(cap_group)
+                    temp_dict[key] = None
+
+        # there are no capture groups, we just search for all the matches of the regex
+        else:
+            #given that there are matches to be made
+            if self.pattern.search(content) is not None:
+                m = self.pattern.findall(content)
+                temp_dict[self.label] = ', '.join(m)
+            else:
+                temp_dict[self.label] = None    
+        # update rev_data with our new columns
+        rev_data.update(temp_dict)
+        return rev_data
+
+
+        
 class WikiqParser():
-    
-    def __init__(self, input_file, output_file, collapse_user=False, persist=None, urlencode=False, namespaces = None):
+    def __init__(self, input_file, output_file, regex_match_revision, regex_match_comment, regex_revision_label, regex_comment_label, collapse_user=False, persist=None, urlencode=False, namespaces = None, revert_radius=15):
         """ 
         Parameters:
            persist : what persistence method to use. Takes a PersistMethod value
         """
-
         self.input_file = input_file
         self.output_file = output_file
         self.collapse_user = collapse_user
@@ -141,11 +198,42 @@ class WikiqParser():
         self.printed_header = False
         self.namespaces = []
         self.urlencode = urlencode
+        self.revert_radius = revert_radius
+
         if namespaces is not None:
             self.namespace_filter = set(namespaces)
         else:
             self.namespace_filter = None
 
+        self.regex_revision_pairs = self.make_matchmake_pairs(regex_match_revision, regex_revision_label)
+        self.regex_comment_pairs = self.make_matchmake_pairs(regex_match_comment, regex_comment_label)
+        
+
+    def make_matchmake_pairs(self, patterns, labels):
+        if (patterns is not None and labels is not None) and \
+           (len(patterns) == len(labels)):
+            return [RegexPair(pattern, label) for pattern, label in zip(patterns, labels)]
+        elif (patterns is None and labels is None):
+            return []
+        else:
+            sys.exit('Each regular expression *must* come with a corresponding label and vice versa.')
+
+    def matchmake(self, rev, rev_data):
+        rev_data = self.matchmake_revision(rev.text, rev_data)
+        rev_data = self.matchmake_comment(rev.comment, rev_data)
+        return rev_data
+
+    def matchmake_revision(self, text, rev_data):
+        return self.matchmake_pairs(text, rev_data, self.regex_revision_pairs)
+
+    def matchmake_comment(self, comment, rev_data):
+        return self.matchmake_pairs(comment, rev_data, self.regex_comment_pairs)
+
+    def matchmake_pairs(self, text, rev_data, pairs):
+        for pair in pairs:
+            rev_data = pair.matchmake(text, rev_data)
+        return rev_data
+
     def __get_namespace_from_title(self, title):
         default_ns = None
 
@@ -161,6 +249,7 @@ class WikiqParser():
         # if we've made it this far with no matches, we return the default namespace
         return default_ns
 
+
     def process(self):
 
         # create a regex that creates the output filename
@@ -187,7 +276,7 @@ class WikiqParser():
                 if namespace not in self.namespace_filter:
                     continue
 
-            rev_detector = mwreverts.Detector()
+            rev_detector = mwreverts.Detector(radius = self.revert_radius)
 
             if self.persist != PersistMethod.none:
                 window = deque(maxlen=PERSISTENCE_RADIUS)
@@ -207,14 +296,19 @@ class WikiqParser():
 
             # Iterate through a page's revisions
             for rev in page:
-
-                rev_data = {'revid' : rev.id,
-                            'date_time' : rev.timestamp.strftime('%Y-%m-%d %H:%M:%S'),
-                            'articleid' : page.id,
-                            'editor_id' : "" if rev.deleted.user == True or rev.user.id is None else rev.user.id,
-                            'title' : '"' + page.title + '"',
-                            'namespace' : namespace,
-                            'deleted' : "TRUE" if rev.deleted.text else "FALSE" } 
+                
+                # initialize rev_data
+                rev_data = {
+                    'revid':rev.id,
+                    'date_time' : rev.timestamp.strftime('%Y-%m-%d %H:%M:%S'),
+                    'articleid' : page.id,
+                    'editor_id' : "" if rev.deleted.user == True or rev.user.id is None else rev.user.id,
+                    'title' : '"' + page.title + '"',
+                    'namespace' : namespace,
+                    'deleted' : "TRUE" if rev.deleted.text else "FALSE"
+                }
+
+                rev_data = self.matchmake(rev, rev_data)
 
                 # if revisions are deleted, /many/ things will be missing
                 if rev.deleted.text:
@@ -239,7 +333,7 @@ class WikiqParser():
 
                     # TODO rev.bytes doesn't work.. looks like a bug
                     rev_data['text_chars'] = len(rev.text)
-               
+
                     # generate revert data
                     revert = rev_detector.process(text_sha1, rev.id)
                     
@@ -379,7 +473,7 @@ parser.add_argument('--collapse-user', dest="collapse_user", action="store_true"
                     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.")
 
 parser.add_argument('-p', '--persistence', dest="persist", default=None, const='', type=str, choices = ['','segment','sequence','legacy'], nargs='?',
-                    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. This may by slow.  Use -p=segment for advanced persistence calculation method that is robust to content moves. This might be very slow. Use -p=legacy for legacy behavior.")
+                    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. This may by slow.  The defualt is -p=sequence, which uses the same algorithm as in the past, but with improvements to wikitext parsing. Use -p=legacy for old behavior used in older research projects. Use -p=segment for advanced persistence calculation method that is robust to content moves, but prone to bugs, and slower.")
 
 parser.add_argument('-u', '--url-encode', dest="urlencode", action="store_true",
                     help="Output url encoded text strings. This works around some data issues like newlines in editor names. In the future it may be used to output other text data.")
@@ -387,7 +481,25 @@ parser.add_argument('-u', '--url-encode', dest="urlencode", action="store_true",
 parser.add_argument('-n', '--namespace-include', dest="namespace_filter", type=int, action='append',
                     help="Id number of namspace to include. Can be specified more than once.")
 
+parser.add_argument('-rr',
+                    '--revert-radius',
+                    dest="revert_radius",
+                    type=int,
+                    action='store',
+                    default=15,
+                    help="Number of edits to check when looking for reverts (default: 15)")
+
+parser.add_argument('-RP', '--revision-pattern', dest="regex_match_revision", default=None, type=str, action='append',
+                    help="The regular expression to search for in revision text. The regex must be surrounded by quotes.")
+
+parser.add_argument('-RPl', '--revision-pattern-label', dest="regex_revision_label", default=None, type=str, action='append',
+                    help="The label for the outputted column based on matching the regex in revision text.")
+
+parser.add_argument('-CP', '--comment-pattern', dest="regex_match_comment", default=None, type=str, action='append',
+                    help="The regular expression to search for in comments of revisions.")
 
+parser.add_argument('-CPl', '--comment-pattern-label', dest="regex_comment_label", default=None, type=str, action='append',
+                    help="The label for the outputted column based on matching the regex in comments.")
 
 args = parser.parse_args()
 
@@ -425,11 +537,17 @@ if len(args.dumpfiles) > 0:
             filename = os.path.join(output_dir, os.path.basename(filename))
             output_file = open_output_file(filename)
 
-        wikiq = WikiqParser(input_file, output_file, 
+        wikiq = WikiqParser(input_file,
+                            output_file,
                             collapse_user=args.collapse_user,
                             persist=persist,
                             urlencode=args.urlencode,
-                            namespaces = namespaces)
+                            namespaces=namespaces,
+                            revert_radius=args.revert_radius,
+                            regex_match_revision = args.regex_match_revision,
+                            regex_revision_label = args.regex_revision_label,
+                            regex_match_comment = args.regex_match_comment,
+                            regex_comment_label = args.regex_comment_label)
 
         wikiq.process()
 
@@ -437,13 +555,20 @@ if len(args.dumpfiles) > 0:
         input_file.close()
         output_file.close()
 else:
-    wikiq = WikiqParser(sys.stdin, sys.stdout,
+    wikiq = WikiqParser(sys.stdin,
+                        sys.stdout,
                         collapse_user=args.collapse_user,
                         persist=persist,
-                        persist_legacy=args.persist_legacy,
+                        #persist_legacy=args.persist_legacy,
                         urlencode=args.urlencode,
-                        namespaces = namespaces)
-    wikiq.process()
+                        namespaces=namespaces,
+                        revert_radius=args.revert_radius,
+                        regex_match_revision = args.regex_match_revision,
+                        regex_revision_label = args.regex_revision_label,
+                        regex_match_comment = args.regex_match_comment,
+                        regex_comment_label = args.regex_comment_label)
+
+    wikiq.process() 
 
 # 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"
 # stop_words = stop_words.split(",")

Community Data Science Collective || Want to submit a patch?