[mod] lower memory footprint by lazy loading JSON data

This patch implements lazy loading of the JSON data.

Motivation: in most requests not all JSON data is needed, but loaded.  By
example these four JSON files:

- currencies.json ~550KB
- engine_descriptions.json ~1,3MB
- external_bangs.json ~1,3MB
- osm_keys_tags.json ~ 2,2MB

most often not used and consume a lot of memory and BTW they also extend the
time required to instantiate a walker.

Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>

searx/plugins/ahmia_filter.py

@@ -2,7 +2,7 @@
 # pylint: disable=missing-module-docstring
 
 from hashlib import md5
-from searx.data import ahmia_blacklist_loader
+from searx import data
 
 name = "Ahmia blacklist"
 description = "Filter out onion results that appear in Ahmia's blacklist. (See https://ahmia.fi/blacklist)"
@@ -24,5 +24,5 @@ def init(_app, settings):
     if not settings['outgoing']['using_tor_proxy']:
         # disable the plugin
         return False
-    ahmia_blacklist = ahmia_blacklist_loader()
+    ahmia_blacklist = data.ahmia_blacklist_loader()
     return True

searx/plugins/unit_converter.py

@@ -4,7 +4,7 @@
 
 from flask_babel import gettext
 
-from searx.data import WIKIDATA_UNITS
+from searx import data
 
 name = "Unit converter plugin"
 description = gettext("Convert between units")
@@ -38,7 +38,7 @@ def _parse_text_and_convert(search, splitted_query):
     from_unit = None
     to_unit = None
 
-    for unit in WIKIDATA_UNITS.values():
+    for unit in data.WIKIDATA_UNITS.values():
         if unit['symbol'] == from_unit_key:
             from_unit = unit