Source code for scripts.freebasemappingupload

# -*- coding: utf-8 -*-
Script to upload the mappings of Freebase to Wikidata.

Can be easily adapted to upload other String identifiers as well

This bot needs the dump from

The script takes a single parameter:

-filename: the filename to read the freebase-wikidata mappings from;
           default: fb2w.nt.gz
# (C) Pywikibot team, 2013-2018
# Distributed under the terms of the MIT license.
from __future__ import absolute_import, division, unicode_literals

import gzip
import os
import sys

import pywikibot

[docs]class FreebaseMapperRobot(object): """Freebase Mapping bot."""
[docs] def __init__(self, filename): """Initializer.""" self.repo = pywikibot.Site('wikidata', 'wikidata').data_repository() self.filename = filename if not os.path.exists(self.filename): pywikibot.output('Cannot find %s. Try providing the absolute path.' % self.filename) sys.exit(1)
[docs] def run(self): """Run the bot.""" # Set up some items we will use a lot. self.claim = pywikibot.Claim(self.repo, 'P646') # freebase mapping # And sources! self.statedin = pywikibot.Claim(self.repo, 'P248') # stated in # Freebase data dump freebasedumpitem = pywikibot.ItemPage(self.repo, 'Q15241312') self.statedin.setTarget(freebasedumpitem) # date of publication self.dateofpub = pywikibot.Claim(self.repo, 'P577') oct28 = pywikibot.WbTime(site=self.repo, year=2013, month=10, day=28, precision='day') self.dateofpub.setTarget(oct28) for line in self.processLine(line.strip())
[docs] def processLine(self, line): """Process a single line.""" if not line or line.startswith('#'): return mid, sameas, qid, dot = line.split() if sameas != '<>': return if dot != '.': return if not mid.startswith('<'): return mid = '/m/' + mid[30:-1] if not qid.startswith('<'): return qid = 'Q' + qid[33:-1] data = pywikibot.ItemPage(self.repo, qid) data.get() if not data.labels: label = '' elif 'en' in data.labels: label = data.labels['en'] else: # Just pick up the first label label = list(data.labels.values())[0] pywikibot.output('Parsed: {} <--> {}'.format(qid, mid)) pywikibot.output('{} is {}'.format(data.getID(), label)) if and 'P646' in # We assume that there is only one claim. # If there are multiple ones, our logs might be wrong # but the constraint value reports will catch them if mid !=['P646'][0].getTarget(): pywikibot.output('Mismatch: expected {}, has {} instead' .format(mid,['P646'][0].getTarget())) else: pywikibot.output('Already has mid set, is consistent.') else: # No claim set, lets add it. pywikibot.output('Going to add a new claim.') self.claim.setTarget(mid) data.addClaim(self.claim) self.claim.addSources([self.statedin, self.dateofpub]) pywikibot.output('Claim added!')
[docs]def main(*args): """ Process command line arguments and invoke bot. If args is an empty list, sys.argv is used. @param args: command line arguments @type args: str """ filename = 'fb2w.nt.gz' # Default filename for arg in pywikibot.handle_args(args): if arg.startswith('-filename'): filename = arg[11:] bot = FreebaseMapperRobot(filename)
if __name__ == '__main__': main()