commit 87013cceff0c24a9a95e7a9bb15edadedf2e62e4
parent 1628ed85af979939b9150cccad113ac12c3eda40
Author: Antoine Amarilli <a3nm@a3nm.net>
Date: Thu, 18 Aug 2016 01:18:40 +0200
Merge branch 'master' of a3nm.net:git/mybin
Diffstat:
| wikicateg | | | 43 | +++++++++++++++++++++++++++++++++++++++++++ |
1 file changed, 43 insertions(+), 0 deletions(-)
diff --git a/wikicateg b/wikicateg
@@ -0,0 +1,43 @@
+#!/usr/bin/python3
+
+"""Produce text list of article names in a wikipedia category"""
+
+# license: public domain
+
+import json
+import urllib.request,urllib.parse
+import sys
+import time
+from pprint import pprint
+
+if len(sys.argv) != 3:
+ print("Usage: %s WIKI CATEGORY" % sys.argv[0], file=sys.stderr)
+ print("WIKI should be like \"en.wikipedia.org\"", file=sys.stderr)
+ sys.exit(1)
+
+progname, wiki, categ = sys.argv
+
+def retrieve(cont=''):
+ args = {'action': 'query',
+ 'list': 'categorymembers',
+ 'cmtitle': 'Category:%s' % categ,
+ 'cmlimit': 500,
+ 'format': 'json'
+ }
+ if cont != '':
+ args['cmcontinue'] = cont
+ data = json.loads(urllib.request.urlopen(
+ "https://"+wiki+"/w/api.php?",
+ urllib.parse.urlencode(args).encode('utf8')
+ ).read().decode('utf8'))
+ if 'error' in data.keys():
+ print("Error: %s" % (data['error']['info']), file=sys.stderr)
+ sys.exit(1)
+ for a in data['query']['categorymembers']:
+ print(a['title'])
+ if 'continue' in data.keys():
+ time.sleep(2)
+ retrieve(data['continue']['cmcontinue'])
+
+retrieve()
+
