summaryrefslogtreecommitdiff
path: root/utils/fetch_wikidata_units.py
diff options
context:
space:
mode:
authorAlexandre Flament <alex@al-f.net>2021-03-05 09:43:39 +0100
committerGitHub <noreply@github.com>2021-03-05 09:43:39 +0100
commitaaae9a209e4d6518965064e8b515e411fb8dd0d6 (patch)
tree922a8875536ac058d501bb207c04219744eee538 /utils/fetch_wikidata_units.py
parent1d10ae175c0929d383d268f56bfadb304365ccf2 (diff)
parentb8cd3264644208d7afa1a239f829222d45226334 (diff)
Merge pull request #2600 from dalf/searx-extra
Add searx_extra package
Diffstat (limited to 'utils/fetch_wikidata_units.py')
-rw-r--r--utils/fetch_wikidata_units.py56
1 files changed, 0 insertions, 56 deletions
diff --git a/utils/fetch_wikidata_units.py b/utils/fetch_wikidata_units.py
deleted file mode 100644
index 69ae8ab27..000000000
--- a/utils/fetch_wikidata_units.py
+++ /dev/null
@@ -1,56 +0,0 @@
-#!/usr/bin/env python
-
-import json
-import collections
-
-# set path
-from sys import path
-from os.path import realpath, dirname, join
-path.append(realpath(dirname(realpath(__file__)) + '/../'))
-
-from searx import searx_dir
-from searx.engines.wikidata import send_wikidata_query
-
-
-# the response contains duplicate ?item with the different ?symbol
-# "ORDER BY ?item DESC(?rank) ?symbol" provides a deterministic result
-# even if a ?item has different ?symbol of the same rank.
-# A deterministic result
-# see:
-# * https://www.wikidata.org/wiki/Help:Ranking
-# * https://www.mediawiki.org/wiki/Wikibase/Indexing/RDF_Dump_Format ("Statement representation" section)
-# * https://w.wiki/32BT
-# see the result for https://www.wikidata.org/wiki/Q11582
-# there are multiple symbols the same rank
-SARQL_REQUEST = """
-SELECT DISTINCT ?item ?symbol
-WHERE
-{
- ?item wdt:P31/wdt:P279 wd:Q47574 .
- ?item p:P5061 ?symbolP .
- ?symbolP ps:P5061 ?symbol ;
- wikibase:rank ?rank .
- FILTER(LANG(?symbol) = "en").
-}
-ORDER BY ?item DESC(?rank) ?symbol
-"""
-
-
-def get_data():
- results = collections.OrderedDict()
- response = send_wikidata_query(SARQL_REQUEST)
- for unit in response['results']['bindings']:
- name = unit['item']['value'].replace('http://www.wikidata.org/entity/', '')
- unit = unit['symbol']['value']
- if name not in results:
- # ignore duplicate: always use the first one
- results[name] = unit
- return results
-
-
-def get_wikidata_units_filename():
- return join(join(searx_dir, "data"), "wikidata_units.json")
-
-
-with open(get_wikidata_units_filename(), 'w') as f:
- json.dump(get_data(), f, indent=4, ensure_ascii=False)