Add searx_extra package

Split the utils directory into:
* searx_extra contains update scripts, standalone_searx.py
* utils contains the files to build and setup searx.

1 files changed, 54 insertions, 0 deletions

diff --git a/searx_extra/update/update_wikidata_units.py b/searx_extra/update/update_wikidata_units.py
new file mode 100755
index 000000000..1e6b8b9ca
--- /dev/null
+++ b/searx_extra/update/update_wikidata_units.py
@@ -0,0 +1,54 @@
+#!/usr/bin/env python
+
+import json
+import collections
+
+# set path
+from os.path import join
+
+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)