From 46efb2f36d327e618c5cd6af702b96fa9143fc27 Mon Sep 17 00:00:00 2001
From: Bnyro <bnyro@tutanota.com>
Date: Sun, 7 Apr 2024 16:17:11 +0200
Subject: [PATCH] [feat] plugins: new unit converter plugin

---
 searx/engines/duckduckgo_definitions.py       |  5 +-
 searx/plugins/unit_converter.py               | 76 +++++++++++++++++++
 searxng_extra/update/update_wikidata_units.py | 23 +++++-
 3 files changed, 99 insertions(+), 5 deletions(-)
 create mode 100644 searx/plugins/unit_converter.py

diff --git a/searx/engines/duckduckgo_definitions.py b/searx/engines/duckduckgo_definitions.py
index e8146cc3e..59caed8ce 100644
--- a/searx/engines/duckduckgo_definitions.py
+++ b/searx/engines/duckduckgo_definitions.py
@@ -238,7 +238,10 @@ def unit_to_str(unit):
     for prefix in WIKIDATA_PREFIX:
         if unit.startswith(prefix):
             wikidata_entity = unit[len(prefix) :]
-            return WIKIDATA_UNITS.get(wikidata_entity, unit)
+            real_unit = WIKIDATA_UNITS.get(wikidata_entity)
+            if real_unit is None:
+                return unit
+            return real_unit['symbol']
     return unit
 
 
diff --git a/searx/plugins/unit_converter.py b/searx/plugins/unit_converter.py
new file mode 100644
index 000000000..dd515aa72
--- /dev/null
+++ b/searx/plugins/unit_converter.py
@@ -0,0 +1,76 @@
+# SPDX-License-Identifier: AGPL-3.0-or-later
+"""Calculate mathematical expressions using ack#eval
+"""
+
+from flask_babel import gettext
+
+from searx.data import WIKIDATA_UNITS
+
+name = "Unit converter plugin"
+description = gettext("Convert between units")
+default_on = True
+
+CONVERT_KEYWORDS = ["in", "to", "as"]
+
+
+def _convert(from_value, source_si_factor, target_si_factor):
+    return from_value * source_si_factor / target_si_factor
+
+
+def _parse_text_and_convert(search, splitted_query):
+    if len(splitted_query) != 2 or splitted_query[0].strip() == "" or splitted_query[1].strip() == "":
+        return
+
+    from_value = ""
+    from_unit_key = ""
+
+    # only parse digits as value that belong together
+    read_alpha = False
+    for c in splitted_query[0]:
+        if not read_alpha and (c in ("-", ".") or str.isdigit(c)):
+            from_value += c
+            read_alpha = True
+        elif c != " ":
+            from_unit_key += c
+
+    to_unit_key = splitted_query[1].strip()
+
+    from_unit = None
+    to_unit = None
+
+    for unit in WIKIDATA_UNITS.values():
+        if unit['symbol'] == from_unit_key:
+            from_unit = unit
+
+        if unit['symbol'] == to_unit_key:
+            to_unit = unit
+
+        if from_unit and to_unit:
+            break
+
+    if from_unit is None or to_unit is None or to_unit.get('si_name') != from_unit.get('si_name'):
+        return
+
+    result = _convert(float(from_value), from_unit['to_si_factor'], to_unit['to_si_factor'])
+    search.result_container.answers['conversion'] = {'answer': f"{result:g} {to_unit['symbol']}"}
+
+
+def post_search(_request, search):
+    # only convert between units on the first page
+    if search.search_query.pageno > 1:
+        return True
+
+    query = search.search_query.query
+    query_parts = query.split(" ")
+
+    if len(query_parts) < 3:
+        return True
+
+    for query_part in query_parts:
+        for keyword in CONVERT_KEYWORDS:
+            if query_part == keyword:
+                keyword_split = query.split(keyword, 1)
+                _parse_text_and_convert(search, keyword_split)
+                return True
+
+    return True
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