Add parameter for selecting specific namespaces.

test/Wikiq_Unit_Test.py

@@ -51,6 +51,28 @@ class Test_Wikipedia(unittest.TestCase):
         baseline = pd.read_table(baseline_file)
         assert_frame_equal(test,baseline)
 
+    def test_WP_namespaces(self):
+        print(os.path.abspath('.'))
+        test_filename =  "namespaces_" + self.wikiq_out_name
+        test_file = os.path.join(self.test_output_dir, test_filename)
+        if os.path.exists(test_file):
+            os.remove(test_file)
+        
+        call = self.base_call.format(self.input_file, self.test_output_dir)
+        call = call + " -n 0 -n 1"
+        print(call)
+        proc = subprocess.Popen(call,stdout=subprocess.PIPE,shell=True)
+        proc.wait()
+        copyfile(self.call_output, test_file)
+        baseline_file = os.path.join(os.path.abspath("."), self.baseline_output_dir, test_filename)
+
+        # as a test let's make sure that we get equal data frames
+        test = pd.read_table(test_file)
+        num_wrong_ns = sum(~ test.namespace.isin({0,1}))
+        self.assertEquals(num_wrong_ns, 0)
+        baseline = pd.read_table(baseline_file)
+        assert_frame_equal(test,baseline)
+
 
 class Test_Basic(unittest.TestCase):
 
@@ -189,7 +211,6 @@ class Test_Basic(unittest.TestCase):
 
 
 class Test_Malformed(unittest.TestCase):
-
     def setUp(self):
         if not os.path.exists("test_output"):
             os.mkdir("test_output")

wikiq

@@ -3,7 +3,6 @@
 # original wikiq headers are: title articleid revid date_time anon
 # editor editor_id minor text_size text_entropy text_md5 reversion
 # additions_size deletions_size
-import pdb
 import argparse
 import sys
 import os, os.path
@@ -128,7 +127,7 @@ class WikiqPage():
 
 class WikiqParser():
     
-    def __init__(self, input_file, output_file, collapse_user=False, persist=None, urlencode=False):
+    def __init__(self, input_file, output_file, collapse_user=False, persist=None, urlencode=False, namespaces = None):
         """ 
         Parameters:
            persist : what persistence method to use. Takes a PersistMethod value
@@ -141,7 +140,11 @@ class WikiqParser():
         self.printed_header = False
         self.namespaces = []
         self.urlencode = urlencode
-        
+        if namespaces is not None:
+            self.namespace_filter = set(namespaces)
+        else:
+            self.namespace_filter = None
+
     def __get_namespace_from_title(self, title):
         default_ns = None
 
@@ -176,6 +179,13 @@ class WikiqParser():
 
         # Iterate through pages
         for page in dump:
+            namespace = page.namespace if page.namespace is not None else self.__get_namespace_from_title(page.title)
+
+            # skip namespaces not in the filter
+            if self.namespace_filter is not None:
+                if namespace not in self.namespace_filter:
+                    continue
+
             rev_detector = mwreverts.Detector()
 
             if self.persist != PersistMethod.none:
@@ -194,6 +204,8 @@ class WikiqParser():
                     from mw.lib import persistence
                     state = persistence.State()
 
+                
+
             # Iterate through a page's revisions
             for rev in page:
 
@@ -202,7 +214,7 @@ class WikiqParser():
                             'articleid' : page.id,
                             'editor_id' : "" if rev.deleted.user == True or rev.user.id is None else rev.user.id,
                             'title' : '"' + page.title + '"',
-                            'namespace' : page.namespace if page.namespace is not None else self.__get_namespace_from_title(page.title),
+                            'namespace' : namespace,
                             'deleted' : "TRUE" if rev.deleted.text else "FALSE" } 
 
                 # if revisions are deleted, /many/ things will be missing
@@ -373,12 +385,12 @@ parser.add_argument('-p', '--persistence', dest="persist", default=None, const='
 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.")
 
-parser.add_argument('-n', '--namespace-include', dest="namespace_filter", type=int, action='append',
+parser.add_argument('-n', '--namespace-include', dest="namespace_include", type=int, action='append',
                     help="Id number of namspace to include.")
 
 
 args = parser.parse_args()
-print(args)
+
 # set persistence method
 
 if args.persist is None:
@@ -390,6 +402,11 @@ elif args.persist == "legacy":
 else:
     persist = PersistMethod.sequence
 
+if args.namespace_include is not None:
+    namespaces = args.namespace_include
+else:
+    namespaces = None
+
 if len(args.dumpfiles) > 0:
     for filename in args.dumpfiles:
         input_file = open_input_file(filename)
@@ -408,10 +425,11 @@ 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)
+                            persist=persist,
+                            urlencode=args.urlencode,
+                            namespaces = namespaces)
 
         wikiq.process()