jenkins-bot has submitted this change. (
https://gerrit.wikimedia.org/r/c/pywikibot/core/+/617118 )
Change subject: [cleanup] Code improvements to archivebot.py
......................................................................
[cleanup] Code improvements to archivebot.py
Add more typehints and optimize.
Change-Id: I9fa6119379d1317558a8bc156bcf9a7abea07eee
---
M scripts/archivebot.py
1 file changed, 24 insertions(+), 18 deletions(-)
Approvals:
Xqt: Looks good to me, approved
jenkins-bot: Verified
diff --git a/scripts/archivebot.py b/scripts/archivebot.py
index eda4978..a9d4537 100755
--- a/scripts/archivebot.py
+++ b/scripts/archivebot.py
@@ -164,7 +164,7 @@
"""
-def str2localized_duration(site, string) -> str:
+def str2localized_duration(site, string: str) -> str:
"""
Localise a shorthand duration.
@@ -181,7 +181,7 @@
return to_local_digits(string, site.code)
-def str2time(string, timestamp=None) -> datetime.timedelta:
+def str2time(string: str, timestamp=None) -> datetime.timedelta:
"""
Return a timedelta for a shorthand duration.
@@ -221,7 +221,7 @@
return datetime.timedelta(days=days)
-def checkstr(string) -> Tuple[str, str]:
+def checkstr(string: str) -> Tuple[str, str]:
"""
Return the key and duration extracted from the string.
@@ -246,7 +246,7 @@
return key, duration
-def str2size(string) -> Size:
+def str2size(string: str) -> Size:
"""
Return a size for a shorthand size.
@@ -271,7 +271,7 @@
return val, unit
-def template_title_regex(tpl_page) -> Pattern:
+def template_title_regex(tpl_page: pywikibot.Page) -> Pattern:
"""
Return a regex that matches to variations of the template title.
@@ -338,7 +338,7 @@
:Reply, etc. ~~~~
"""
- def __init__(self, title, _now=None, timestripper=None) -> None:
+ def __init__(self, title: str, _now=None, timestripper=None) -> None:
"""Initializer."""
if _now is not None:
issue_deprecation_warning(
@@ -357,7 +357,7 @@
return '{}("{}",{} bytes)'.format(self.__class__.__name__,
self.title,
len(self.content.encode('utf-8')))
- def feed_line(self, line) -> None:
+ def feed_line(self, line: str) -> None:
"""Add a line to the content and find the newest
timestamp."""
if not self.content and not line:
return
@@ -404,7 +404,7 @@
def __init__(self, source, archiver, params=None) -> None:
"""Initializer."""
- super(DiscussionPage, self).__init__(source)
+ super().__init__(source)
self.threads = []
self.full = False
self.archiver = archiver
@@ -451,8 +451,8 @@
for thread_heading, thread_content in threads:
cur_thread = DiscussionThread(
thread_heading.strip('= '), timestripper=self.timestripper)
- lines = thread_content.replace(marker, '').splitlines()
- lines = lines[1:] # remove heading line
+ # remove heading line
+ _, *lines = thread_content.replace(marker, '').splitlines()
for line in lines:
cur_thread.feed_line(line)
self.threads.append(cur_thread)
@@ -481,7 +481,13 @@
return self.is_full(max_archive_size)
def size(self) -> int:
- """Return size of talk page threads."""
+ """
+ Return size of talk page threads.
+
+ Note that this method counts bytes, rather than codepoints
+ (characters). This corresponds to MediaWiki's definition
+ of page size.
+ """
return len(self.header.encode('utf-8')) + sum(t.size()
for t in self.threads)
@@ -589,7 +595,8 @@
if not self.get_attr('archive', ''):
raise MissingConfigError('Missing argument "archive" in
template')
- def should_archive_thread(self, thread) -> Optional[ShouldArchive]:
+ def should_archive_thread(self, thread: DiscussionThread
+ ) -> Optional[ShouldArchive]:
"""
Check whether a thread has to be archived.
@@ -831,17 +838,16 @@
else:
ns = []
pywikibot.output('Fetching template transclusions...')
- for pg in tmpl.getReferences(only_template_inclusion=True,
- follow_redirects=False,
- namespaces=ns):
- pagelist.append(pg)
+ pagelist.extend(tmpl.getReferences(only_template_inclusion=True,
+ follow_redirects=False,
+ namespaces=ns))
if filename:
for pg in open(filename, 'r').readlines():
pagelist.append(pywikibot.Page(site, pg, ns=10))
if pagename:
pagelist.append(pywikibot.Page(site, pagename, ns=3))
- pagelist = sorted(pagelist)
- for pg in iter(pagelist):
+ pagelist.sort()
+ for pg in pagelist:
pywikibot.output('Processing {}'.format(pg))
# Catching exceptions, so that errors in one page do not bail out
# the entire process
--
To view, visit
https://gerrit.wikimedia.org/r/c/pywikibot/core/+/617118
To unsubscribe, or for help writing mail filters, visit
https://gerrit.wikimedia.org/r/settings
Gerrit-Project: pywikibot/core
Gerrit-Branch: master
Gerrit-Change-Id: I9fa6119379d1317558a8bc156bcf9a7abea07eee
Gerrit-Change-Number: 617118
Gerrit-PatchSet: 1
Gerrit-Owner: Matěj Suchánek <matejsuchanek97(a)gmail.com>
Gerrit-Reviewer: D3r1ck01 <xsavitar.wiki(a)aol.com>
Gerrit-Reviewer: Xqt <info(a)gno.de>
Gerrit-Reviewer: jenkins-bot
Gerrit-MessageType: merged