From 0c03a5a7b40b598b88f22f46b9e9086af6c59877 Mon Sep 17 00:00:00 2001
From: Sergey Poznyakoff <gray@gnu.org>
Date: Fri, 17 Aug 2018 17:05:32 +0300
Subject: Improve bin/wikitrans

* bin/wikitrans (getwiki): Take options as second argument and
modify it directly. Deduce options.lang, options.itype, and
options.kwdict['html_base'] from the URL, when possible.
* README.rst: Update description of bin/wikitrans
---
 README.rst    |  4 +++-
 bin/wikitrans | 24 ++++++++++++------------
 2 files changed, 15 insertions(+), 13 deletions(-)

diff --git a/README.rst b/README.rst
index b6863bc..d189c98 100644
--- a/README.rst
+++ b/README.rst
@@ -170,5 +170,7 @@ Options are:
 ``-b URL``, ``--base-url=URL``
   Set base url. 
 
-
+Note: when using ``--base-url`` or passing URL as an argument (2nd and 3rd
+use cases above), if the URL is in 'wikipedia.org' or 'wiktionary.org'
+domain, the options ``--input-type``, and ``--lang`` are set automatically.
 
diff --git a/bin/wikitrans b/bin/wikitrans
index e9ab81f..caaa885 100755
--- a/bin/wikitrans
+++ b/bin/wikitrans
@@ -70,7 +70,7 @@ def setkw(option, opt, value, parser):
     if val:
         parser.values.kwdict[kw] = val
 
-def getwiki(url):
+def getwiki(url, options):
     tmp = tempfile.NamedTemporaryFile()            
     if sys.version_info[0] > 2:
         import urllib.request
@@ -87,7 +87,15 @@ def getwiki(url):
     if text is None:
         print("no page/revision/text element in the downloaded page")
         exit(0)
-    return text.text.encode()
+
+    m = re.match('(?P<url>(?:.+://)(?P<lang>.+?)\.(?P<root>wik(?:ipedia|tionary))\.org)', url)
+    if m:
+        options.lang = m.group('lang')
+        options.kwdict['html_base'] = m.group('url') + '/wiki/'
+        if m.group('root') == 'wiktionary':
+            options.itype = 'wiktionary'
+        
+    options.kwdict['text'] = text.text.encode()
         
 def main():
     usage = '%prog [OPTIONS] ARG'
@@ -141,19 +149,11 @@ Otherwise, ARG is name of the file to read wiki material from.
     (options, args) = parser.parse_args()
     if len(args) == 1:
         if options.base_url:
-            options.kwdict['text'] = getwiki(options.base_url
-                                             + '/wiki/Special:Export/' + args[0])
-            m = re.match('(?:.+://)(.+?)\.(wik(?:ipedia|tionary))\.org',
-                         options.base_url)
-            if m:
-                options.lang = m.group(1)
-                options.kwdict['html_base'] = options.base_url + '/wiki/'
-                if m.group(2) == 'wiktionary':
-                    options.itype = 'wiktionary'
+            getwiki(options.base_url + '/wiki/Special:Export/' + args[0], options)
         elif args[0] == '-':
             options.kwdict['file'] = sys.stdin
         elif re.match('^(http|ftp)s?://',args[0]):
-            options.kwdict['text'] = getwiki(args[0])
+            getwiki(args[0], options)
         else:
             options.kwdict['filename'] = args[0]
     else:
-- 
cgit v1.2.1