jenkins-bot submitted this change.

View Change

[bugfix] Do not rely on self.current_page.site

It is set to the item whenever a suitable value
to import is found. This causes page existence
checks againts Wikidata, not the original site,
and imports of wrong references.

Change-Id: I99e6914d996d9a7ae540d01cde073e52a4afd286
---
M pywikibot/bot.py
M scripts/harvest_template.py
M tests/harvest_templates_tests.py
3 files changed, 17 insertions(+), 10 deletions(-)

diff --git a/pywikibot/bot.py b/pywikibot/bot.py
index 6ca5c61..3fd0700 100644
--- a/pywikibot/bot.py
+++ b/pywikibot/bot.py
@@ -2218,6 +2218,9 @@
           page save will be reported and ignored (default: False)
         :return: whether the item was saved successfully
 
+        .. note:: calling this method sets the current_page property
+           to the item which changes the site property
+
         .. note:: calling this method with the 'source' argument modifies
            the provided claim object in place
         """
@@ -2266,6 +2269,9 @@
         :param logger_callback: function logging the output of the method
         :return: whether the claim could be added
 
+        .. note:: calling this method may change the current_page property
+           to the item which will also change the site property
+
         .. note:: calling this method with the 'source' argument modifies
            the provided claim object in place
         """
diff --git a/scripts/harvest_template.py b/scripts/harvest_template.py
index 2f28dd5..6dd4677 100755
--- a/scripts/harvest_template.py
+++ b/scripts/harvest_template.py
@@ -218,6 +218,7 @@
 
     def template_link_target(self,
                              item: pywikibot.ItemPage,
+                             site: pywikibot.site.BaseSite,
                              link_text: str) -> Optional[pywikibot.ItemPage]:
         """Find the ItemPage target for a given link text.
 
@@ -225,7 +226,7 @@
            Only follow the redirect target if redirect page has no
            wikibase item.
         """
-        linked_page = pywikibot.Page(self.current_page.site, link_text)
+        linked_page = pywikibot.Page(site, link_text)
         try:
             exists = linked_page.exists()
         except (InvalidTitleError, InvalidPageError):
@@ -296,10 +297,11 @@
 
             # We found the template we were looking for
             for field_item in fielddict.items():
-                self.treat_field(item, field_item)
+                self.treat_field(item, page.site, field_item)
 
     def treat_field(self,
                     item: pywikibot.page.ItemPage,
+                    site: pywikibot.site.BaseSite,
                     field_item: Tuple[str, str]) -> None:
         """Process a single field of template fielddict.
 
@@ -310,8 +312,6 @@
         if not field or field not in self.fields:
             return
 
-        site = self.current_page.site
-
         # todo: extend the list of tags to ignore
         value = textlib.removeDisabledParts(
             # todo: eventually we may want to import the references
@@ -334,7 +334,7 @@
         do_multi = self._get_option_with_fallback(options, 'multi')
         inverse_prop = self._get_option_with_fallback(options, 'inverse')
 
-        for target in handler(value, item, field):
+        for target in handler(value, site, item, field):
             claim = ppage.newClaim()
             claim.setTarget(target)
             # A generator might yield pages from multiple sites
@@ -362,6 +362,7 @@
                 exists_arg.add('p')
 
     def handle_wikibase_item(self, value: str,
+                             site: pywikibot.site.BaseSite,
                              item: pywikibot.page.ItemPage,
                              field: str) -> Iterator[pywikibot.ItemPage]:
         """Handle 'wikibase-item' claim type.
@@ -375,7 +376,7 @@
         for match in pywikibot.link_regex.finditer(value):
             matched = True
             link_text = match.group(1)
-            linked_item = self.template_link_target(item, link_text)
+            linked_item = self.template_link_target(item, site, link_text)
             if linked_item:
                 yield linked_item
 
@@ -388,7 +389,7 @@
                 .format(prop, field, value))
             return
 
-        linked_item = self.template_link_target(item, value)
+        linked_item = self.template_link_target(item, site, value)
         if linked_item:
             yield linked_item
 
@@ -410,13 +411,13 @@
         for match in self.linkR.finditer(value):
             yield match.group('url')
 
-    def handle_commonsmedia(self, value, *args
+    def handle_commonsmedia(self, value, site, *args
                             ) -> Iterator[pywikibot.FilePage]:
         """Handle 'commonsMedia' claim type.
 
         .. versionadded:: 7.5
         """
-        repo = self.current_page.site.image_repository()
+        repo = site.image_repository()
         image = pywikibot.FilePage(repo, value)
         if image.isRedirectPage():
             image = pywikibot.FilePage(image.getRedirectTarget())
diff --git a/tests/harvest_templates_tests.py b/tests/harvest_templates_tests.py
index b0804ee..3909727 100644
--- a/tests/harvest_templates_tests.py
+++ b/tests/harvest_templates_tests.py
@@ -44,7 +44,7 @@
             with self.subTest(link=link, item=item):
                 dummy_item = ItemPage(self.site.data_repository(), 'Q1')
                 target = HarvestRobot.template_link_target(
-                    DummyBot(self.site), dummy_item, link)
+                    DummyBot(self.site), dummy_item, self.site, link)
                 self.assertIsInstance(target, ItemPage)
                 self.assertEqual(target.title(), item)
 

To view, visit change 815234. To unsubscribe, or for help writing mail filters, visit settings.