From 5e4c5e0361186dbf39f5be5ca60a1edfec6dcea1 Mon Sep 17 00:00:00 2001
From: Dan Howe <d.howe@wrl.unsw.edu.au>
Date: Thu, 1 Oct 2020 18:33:10 +0200
Subject: [PATCH] Parse xml file with mwxml package

---
 wiki_stats.py | 31 ++++++++++++++-----------------
 1 file changed, 14 insertions(+), 17 deletions(-)

diff --git a/wiki_stats.py b/wiki_stats.py
index 90af9fa..85796b3 100644
--- a/wiki_stats.py
+++ b/wiki_stats.py
@@ -3,6 +3,7 @@ import re
 import pandas as pd
 from lxml import etree
 import matplotlib.pyplot as plt
+import mwxml  # pip install mwxml
 
 OUTPUT_DIR = 'csv'
 PROJECTS_TEAM = [  # Oderered by first entry
@@ -68,25 +69,21 @@ PROJECTS_TEAM = [  # Oderered by first entry
 ]
 
 
-
 def main(xml_name):
-    ns = {'mediawiki': 'http://www.mediawiki.org/xml/export-0.3/'}
-    root = etree.parse(xml_name)
-    pages = root.xpath('//mediawiki:page', namespaces=ns)
-
     revisions = []
-    for page in pages:
-        title = page.xpath('./mediawiki:title', namespaces=ns)[0].text
-        for revision in page.xpath('./mediawiki:revision', namespaces=ns):
-            timestamp = revision.xpath('./mediawiki:timestamp',
-                                       namespaces=ns)[0].text
-            contributor = revision.getchildren()[2].getchildren()[0].text
-
-            revisions.append({
-                'page': title,
-                'user': contributor,
-                'date': str(timestamp)
-            })
+    with open(xml_name, 'rb') as f:
+        pages = mwxml.Dump.from_file(f)
+        for page in pages:
+            title = page.title
+            for revision in page:
+                timestamp = str(revision.timestamp)
+                contributor = revision.user.text
+
+                revisions.append({
+                    'page': title,
+                    'user': contributor,
+                    'date': timestamp
+                })
 
     df = pd.DataFrame(revisions)
     df['date'] = pd.to_datetime(df['date'])