From 1bb82a6b54e53d683c3041a1576be64ae234abee Mon Sep 17 00:00:00 2001
From: Alexandre Flament <alex@al-f.net>
Date: Sat, 2 Oct 2021 17:30:39 +0200
Subject: SearXNG: searxng_extra

---
 searxng_extra/update/update_wikidata_units.py | 55 +++++++++++++++++++++++++++
 1 file changed, 55 insertions(+)
 create mode 100755 searxng_extra/update/update_wikidata_units.py

(limited to 'searxng_extra/update/update_wikidata_units.py')

diff --git a/searxng_extra/update/update_wikidata_units.py b/searxng_extra/update/update_wikidata_units.py
new file mode 100755
index 000000000..ddde4c135
--- /dev/null
+++ b/searxng_extra/update/update_wikidata_units.py
@@ -0,0 +1,55 @@
+#!/usr/bin/env python
+
+import json
+import collections
+
+# set path
+from os.path import join
+
+from searx import searx_dir
+from searx.engines import wikidata, set_loggers
+
+set_loggers(wikidata, 'wikidata')
+
+# 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 = wikidata.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)
-- 
cgit v1.2.3


From 955eab8240b4dd903e431b85f6daf35c8672d39c Mon Sep 17 00:00:00 2001
From: Markus Heiser <markus.heiser@darmarit.de>
Date: Sun, 3 Oct 2021 15:12:09 +0200
Subject: [mod] searxng_extras - minor improvements

- fix docs/searxng_extra/standalone_searx.py.rst
- add SPDX tag
- pylint standalone_searx.py and update_wikidata_units.py

Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>
---
 searxng_extra/update/update_wikidata_units.py | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

(limited to 'searxng_extra/update/update_wikidata_units.py')

diff --git a/searxng_extra/update/update_wikidata_units.py b/searxng_extra/update/update_wikidata_units.py
index ddde4c135..43a872b1b 100755
--- a/searxng_extra/update/update_wikidata_units.py
+++ b/searxng_extra/update/update_wikidata_units.py
@@ -1,4 +1,7 @@
 #!/usr/bin/env python
+# SPDX-License-Identifier: AGPL-3.0-or-later
+# lint: pylint
+# pylint: disable=missing-module-docstring
 
 import json
 import collections
@@ -51,5 +54,5 @@ def get_wikidata_units_filename():
     return join(join(searx_dir, "data"), "wikidata_units.json")
 
 
-with open(get_wikidata_units_filename(), 'w') as f:
+with open(get_wikidata_units_filename(), 'w', encoding="utf8") as f:
     json.dump(get_data(), f, indent=4, ensure_ascii=False)
-- 
cgit v1.2.3