[feat] plugins: new unit converter plugin

1 files changed, 19 insertions, 4 deletions

diff --git a/searxng_extra/update/update_wikidata_units.py b/searxng_extra/update/update_wikidata_units.py
index a1a3731fc..f384df749 100755
--- a/searxng_extra/update/update_wikidata_units.py
+++ b/searxng_extra/update/update_wikidata_units.py
@@ -29,31 +29,46 @@ set_loggers(wikidata, 'wikidata')
 # * https://www.wikidata.org/wiki/Help:Ranking
 # * https://www.mediawiki.org/wiki/Wikibase/Indexing/RDF_Dump_Format ("Statement representation" section)
 # * https://w.wiki/32BT
+# * https://en.wikibooks.org/wiki/SPARQL/WIKIDATA_Precision,_Units_and_Coordinates#Quantities
 #   see the result for https://www.wikidata.org/wiki/Q11582
 #   there are multiple symbols the same rank
 SARQL_REQUEST = """
-SELECT DISTINCT ?item ?symbol
+SELECT DISTINCT ?item ?symbol ?tosi ?tosiUnit
 WHERE
 {
   ?item wdt:P31/wdt:P279 wd:Q47574 .
   ?item p:P5061 ?symbolP .
   ?symbolP ps:P5061 ?symbol ;
            wikibase:rank ?rank .
+  OPTIONAL {
+    ?item p:P2370 ?tosistmt .
+    ?tosistmt psv:P2370 ?tosinode .
+    ?tosinode wikibase:quantityAmount ?tosi .
+    ?tosinode wikibase:quantityUnit ?tosiUnit .
+  }
   FILTER(LANG(?symbol) = "en").
 }
 ORDER BY ?item DESC(?rank) ?symbol
 """
 
+_wikidata_url = "https://www.wikidata.org/entity/"
+
 
 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']
+        name = unit['item']['value'].replace(_wikidata_url, '')
+        symbol = unit['symbol']['value']
+        si_name = unit.get('tosiUnit', {}).get('value', '').replace(_wikidata_url, '')
+        to_si_factor = unit.get('tosi', {}).get('value', '')
         if name not in results:
             # ignore duplicate: always use the first one
-            results[name] = unit
+            results[name] = {
+                'symbol': symbol,
+                'si_name': si_name if si_name else None,
+                'to_si_factor': float(to_si_factor) if to_si_factor else None,
+            }
     return results