jenkins-bot has submitted this change and it was merged.
Change subject: pep8-ified scripts/data_ingestion.py
......................................................................
pep8-ified scripts/data_ingestion.py
Change-Id: I04c5b328e16db2594c6436a68c73cc66ee116ac5
---
M scripts/data_ingestion.py
1 file changed, 16 insertions(+), 11 deletions(-)
Approvals:
Ladsgroup: Looks good to me, approved
jenkins-bot: Verified
diff --git a/scripts/data_ingestion.py b/scripts/data_ingestion.py
index df43f9d..386ff91 100755
--- a/scripts/data_ingestion.py
+++ b/scripts/data_ingestion.py
@@ -5,10 +5,13 @@
'''
import pywikibot
-import posixpath, urlparse
+import posixpath
+import urlparse
import urllib
-import hashlib, base64
+import hashlib
+import base64
import StringIO
+
class Photo(object):
'''
@@ -36,11 +39,11 @@
TODO: Add exception handling
'''
if not self.contents:
- imageFile=urllib.urlopen(self.URL).read()
+ imageFile = urllib.urlopen(self.URL).read()
self.contents = StringIO.StringIO(imageFile)
return self.contents
- def findDuplicateImages(self, site = pywikibot.getSite(u'commons',
u'commons')):
+ def findDuplicateImages(self, site=pywikibot.getSite(u'commons',
u'commons')):
'''
Takes the photo, calculates the SHA1 hash and asks the mediawiki api for a list
of duplicates.
@@ -76,12 +79,14 @@
def _safeTemplateValue(self, value):
return value.replace("|", "{{!}}")
+
def CSVReader(fileobj, urlcolumn, *args, **kwargs):
import csv
reader = csv.DictReader(fileobj, *args, **kwargs)
for line in reader:
yield Photo(line[urlcolumn], line)
+
class DataIngestionBot:
def __init__(self, reader, titlefmt, pagefmt,
site=pywikibot.getSite(u'commons', u'commons')):
@@ -99,12 +104,12 @@
title = photo.getTitle(self.titlefmt)
description = photo.getDescription(self.pagefmt)
- bot = upload.UploadRobot(url = photo.URL,
- description = description,
- useFilename = title,
- keepFilename = True,
- verifyDescription = False,
- targetSite = self.site)
+ bot = upload.UploadRobot(url=photo.URL,
+ description=description,
+ useFilename=title,
+ keepFilename=True,
+ verifyDescription=False,
+ targetSite=self.site)
bot._contents = photo.downloadPhoto().getvalue()
bot._retrieved = True
bot.run()
@@ -118,7 +123,7 @@
for photo in self.reader:
self._doUpload(photo)
-if __name__=="__main__":
+if __name__ == "__main__":
reader = CSVReader(open('tests/data/csv_ingestion.csv'), 'url')
bot = DataIngestionBot(reader, "%(name)s - %(set)s.%(_ext)s",
":user:valhallasw/test_template", pywikibot.getSite('test',
'test'))
bot.run()
--
To view, visit
https://gerrit.wikimedia.org/r/80847
To unsubscribe, visit
https://gerrit.wikimedia.org/r/settings
Gerrit-MessageType: merged
Gerrit-Change-Id: I04c5b328e16db2594c6436a68c73cc66ee116ac5
Gerrit-PatchSet: 1
Gerrit-Project: pywikibot/core
Gerrit-Branch: master
Gerrit-Owner: Merlijn van Deen <valhallasw(a)arctus.nl>
Gerrit-Reviewer: Ladsgroup <ladsgroup(a)gmail.com>
Gerrit-Reviewer: Legoktm <legoktm.wikipedia(a)gmail.com>
Gerrit-Reviewer: jenkins-bot