data: currencies: use SQLite instead of JSON

searx/data/__init__.py

@@ -7,7 +7,6 @@
 
 __all__ = [
     'ENGINE_TRAITS',
-    'CURRENCIES',
     'USER_AGENTS',
     'EXTERNAL_URLS',
     'WIKIDATA_UNITS',
@@ -16,11 +15,16 @@ __all__ = [
     'LOCALES',
     'ahmia_blacklist_loader',
     'fetch_engine_descriptions',
+    'fetch_iso4217_from_user',
+    'fetch_name_from_iso4217',
 ]
 
+import re
+import unicodedata
 import json
 import sqlite3
-from typing import Dict, List
+from typing import Dict, List, Optional
+from functools import lru_cache
 from threading import local
 from pathlib import Path
 
@@ -44,7 +48,7 @@ def _get_connection(filename: str) -> sqlite3.Connection:
     if connection is not None:
         return connection
 
-    data_filename = str(data_dir / 'engine_descriptions.db')
+    data_filename = str(data_dir / filename)
     # open database in read only mode
     data_connection = sqlite3.connect(f'file:{data_filename}?mode=ro', uri=True)
     # https://phiresky.github.io/blog/2020/sqlite-performance-tuning/
@@ -67,6 +71,44 @@ def fetch_engine_descriptions(language) -> Dict[str, List[str]]:
     return {result[0]: [result[1], result[2]] for result in res.fetchall()}
 
 
+def _normalize_name(name):
+    name = name.lower().replace('-', ' ').rstrip('s')
+    name = re.sub(' +', ' ', name)
+    return unicodedata.normalize('NFKD', name).lower()
+
+
+@lru_cache(10)
+def fetch_iso4217_from_user(name: str) -> Optional[str]:
+    connection = _get_connection("currencies.db")
+
+    # try the iso4217
+    res = connection.execute("SELECT iso4217 FROM currencies WHERE lower(iso4217)=? LIMIT 1", (name.lower(),))
+    result = res.fetchone()
+    if result:
+        return result[0]
+
+    # try the currency names
+    name = _normalize_name(name)
+    res = connection.execute("SELECT iso4217 FROM currencies WHERE name=?", (name,))
+    result = list(set(result[0] for result in res.fetchall()))
+    if len(result) == 1:
+        return result[0]
+
+    # ambiguity --> return nothing
+    return None
+
+
+@lru_cache(10)
+def fetch_name_from_iso4217(iso4217: str, language: str) -> Optional[str]:
+    res = _get_connection("currencies.db").execute(
+        "SELECT name FROM currencies WHERE iso4217=? AND language=?", (iso4217, language)
+    )
+    result = [result[0] for result in res.fetchall()]
+    if len(result) == 1:
+        return result[0]
+    return None
+
+
 def ahmia_blacklist_loader():
     """Load data from `ahmia_blacklist.txt` and return a list of MD5 values of onion
     names.  The MD5 values are fetched by::
@@ -80,7 +122,6 @@ def ahmia_blacklist_loader():
         return f.read().split()
 
 
-CURRENCIES = _load('currencies.json')
 USER_AGENTS = _load('useragents.json')
 EXTERNAL_URLS = _load('external_urls.json')
 WIKIDATA_UNITS = _load('wikidata_units.json')

searx/engines/openstreetmap.py

@@ -10,7 +10,7 @@ from functools import partial
 
 from flask_babel import gettext
 
-from searx.data import OSM_KEYS_TAGS, CURRENCIES
+from searx.data import OSM_KEYS_TAGS, fetch_name_from_iso4217
 from searx.utils import searx_useragent
 from searx.external_urls import get_external_url
 from searx.engines.wikidata import send_wikidata_query, sparql_string_escape, get_thumbnail
@@ -449,9 +449,9 @@ def get_key_label(key_name, lang):
         # https://taginfo.openstreetmap.org/keys/currency#values
         currency = key_name.split(':')
         if len(currency) > 1:
-            o = CURRENCIES['iso4217'].get(currency[1])
-            if o:
-                return get_label(o, lang).lower()
+            label = fetch_name_from_iso4217(currency[1], lang)
+            if label:
+                return label
             return currency[1]
 
     labels = OSM_KEYS_TAGS['keys']

searx/search/processors/online_currency.py

@@ -3,33 +3,14 @@
 
 """
 
-import unicodedata
 import re
 
-from searx.data import CURRENCIES
+from searx.data import fetch_iso4217_from_user, fetch_name_from_iso4217
 from .online import OnlineProcessor
 
 parser_re = re.compile('.*?(\\d+(?:\\.\\d+)?) ([^.0-9]+) (?:in|to) ([^.0-9]+)', re.I)
 
 
-def normalize_name(name):
-    name = name.lower().replace('-', ' ').rstrip('s')
-    name = re.sub(' +', ' ', name)
-    return unicodedata.normalize('NFKD', name).lower()
-
-
-def name_to_iso4217(name):
-    name = normalize_name(name)
-    currency = CURRENCIES['names'].get(name, [name])
-    if isinstance(currency, str):
-        return currency
-    return currency[0]
-
-
-def iso4217_to_name(iso4217, language):
-    return CURRENCIES['iso4217'].get(iso4217, {}).get(language, iso4217)
-
-
 class OnlineCurrencyProcessor(OnlineProcessor):
     """Processor class used by ``online_currency`` engines."""
 
@@ -52,14 +33,17 @@ class OnlineCurrencyProcessor(OnlineProcessor):
             amount = float(amount_str)
         except ValueError:
             return None
-        from_currency = name_to_iso4217(from_currency.strip())
-        to_currency = name_to_iso4217(to_currency.strip())
+        from_currency = fetch_iso4217_from_user(from_currency.strip())
+        to_currency = fetch_iso4217_from_user(to_currency.strip())
+
+        if from_currency is None or to_currency is None:
+            return None
 
         params['amount'] = amount
         params['from'] = from_currency
         params['to'] = to_currency
-        params['from_name'] = iso4217_to_name(from_currency, 'en')
-        params['to_name'] = iso4217_to_name(to_currency, 'en')
+        params['from_name'] = fetch_name_from_iso4217(from_currency, 'en')
+        params['to_name'] = fetch_name_from_iso4217(to_currency, 'en')
         return params
 
     def get_default_tests(self):

searxng_extra/update/update_currencies.py

@@ -9,15 +9,20 @@ Output file: :origin:`searx/data/currencies.json` (:origin:`CI Update data ...
 
 # pylint: disable=invalid-name
 
+import csv
 import re
 import unicodedata
-import json
+import sqlite3
+from pathlib import Path
 
+from searx.network import set_timeout_for_thread
 from searx.locales import LOCALE_NAMES, locales_initialize
 from searx.engines import wikidata, set_loggers
 from searx.data import data_dir
 
-DATA_FILE = data_dir / 'currencies.json'
+DATABASE_FILE = data_dir / 'currencies.db'
+CSV_FILE = data_dir / 'dumps' / 'currencies.csv'
+
 
 set_loggers(wikidata, 'wikidata')
 locales_initialize()
@@ -75,57 +80,45 @@ def _normalize_name(name):
     return name
 
 
-def add_currency_name(db, name, iso4217, normalize_name=True):
-    db_names = db['names']
-
+def add_entry(db, language, iso4217, name, normalize_name=True):
     if normalize_name:
         name = _normalize_name(name)
 
-    iso4217_set = db_names.setdefault(name, [])
-    if iso4217 not in iso4217_set:
-        iso4217_set.insert(0, iso4217)
-
-
-def add_currency_label(db, label, iso4217, language):
-    labels = db['iso4217'].setdefault(iso4217, {})
-    labels[language] = label
+    entry = (language, iso4217, name)
+    db.add(entry)
 
 
 def wikidata_request_result_iterator(request):
+    set_timeout_for_thread(60)
     result = wikidata.send_wikidata_query(request.replace('%LANGUAGES_SPARQL%', LANGUAGES_SPARQL))
     if result is not None:
         yield from result['results']['bindings']
 
 
 def fetch_db():
-    db = {
-        'names': {},
-        'iso4217': {},
-    }
+    db = set()
 
     for r in wikidata_request_result_iterator(SPARQL_WIKIPEDIA_NAMES_REQUEST):
         iso4217 = r['iso4217']['value']
         article_name = r['article_name']['value']
         article_lang = r['article_name']['xml:lang']
-        add_currency_name(db, article_name, iso4217)
-        add_currency_label(db, article_name, iso4217, article_lang)
+        add_entry(db, article_lang, iso4217, article_name)
 
     for r in wikidata_request_result_iterator(SARQL_REQUEST):
         iso4217 = r['iso4217']['value']
         if 'label' in r:
             label = r['label']['value']
             label_lang = r['label']['xml:lang']
-            add_currency_name(db, label, iso4217)
-            add_currency_label(db, label, iso4217, label_lang)
+            add_entry(db, label_lang, iso4217, label)
 
         if 'alias' in r:
-            add_currency_name(db, r['alias']['value'], iso4217)
+            add_entry(db, "", iso4217, r['alias']['value'])
 
         if 'unicode' in r:
-            add_currency_name(db, r['unicode']['value'], iso4217, normalize_name=False)
+            add_entry(db, "", iso4217, r['unicode']['value'], normalize_name=False)
 
         if 'unit' in r:
-            add_currency_name(db, r['unit']['value'], iso4217, normalize_name=False)
+            add_entry(db, "", iso4217, r['unit']['value'], normalize_name=False)
 
     return db
 
@@ -135,22 +128,33 @@ def main():
     db = fetch_db()
 
     # static
-    add_currency_name(db, "euro", 'EUR')
-    add_currency_name(db, "euros", 'EUR')
-    add_currency_name(db, "dollar", 'USD')
-    add_currency_name(db, "dollars", 'USD')
-    add_currency_name(db, "peso", 'MXN')
-    add_currency_name(db, "pesos", 'MXN')
+    add_entry(db, "", 'EUR', "euro")
+    add_entry(db, "", 'EUR', "euros")
+    add_entry(db, "", 'USD', "dollar")
+    add_entry(db, "", 'USD', "dollars")
+    add_entry(
+        db,
+        "",
+        'MXN',
+        "peso",
+    )
+    add_entry(db, "", 'MXN', "pesos")
 
-    # reduce memory usage:
-    # replace lists with one item by the item.  see
-    # searx.search.processors.online_currency.name_to_iso4217
-    for name in db['names']:
-        if len(db['names'][name]) == 1:
-            db['names'][name] = db['names'][name][0]
-
-    with DATA_FILE.open('w', encoding='utf8') as f:
-        json.dump(db, f, indent=4, sort_keys=True, ensure_ascii=False)
+    db = list(db)
+    db.sort(key=lambda entry: (entry[0], entry[1], entry[2]))
+    Path(DATABASE_FILE).unlink(missing_ok=True)
+    with sqlite3.connect(DATABASE_FILE) as con:
+        cur = con.cursor()
+        cur.execute("CREATE TABLE currencies(language, iso4217, name)")
+        cur.executemany("INSERT INTO currencies VALUES(?, ?, ?)", db)
+        cur.execute("CREATE INDEX index_currencies_iso4217 ON currencies('iso4217')")
+        cur.execute("CREATE INDEX index_currencies_name ON currencies('name')")
+        con.commit()
+    with CSV_FILE.open('w', encoding='utf8') as f:
+        w = csv.writer(f, quoting=csv.QUOTE_NONNUMERIC)
+        w.writerow(["language", "iso4217", "name"])
+        for row in db:
+            w.writerow(row)
 
 
 if __name__ == '__main__':