data: osm_keys_tags: use SQLite instead of JSON

searx/data/__init__.py

@@ -11,12 +11,12 @@ __all__ = [
     'EXTERNAL_URLS',
     'WIKIDATA_UNITS',
     'EXTERNAL_BANGS',
-    'OSM_KEYS_TAGS',
     'LOCALES',
     'ahmia_blacklist_loader',
     'fetch_engine_descriptions',
     'fetch_iso4217_from_user',
     'fetch_name_from_iso4217',
+    'fetch_osm_key_label',
 ]
 
 import re
@@ -102,6 +102,43 @@ def fetch_name_from_iso4217(iso4217: str, language: str) -> Optional[str]:
     return None
 
 
+@lru_cache(100)
+def fetch_osm_key_label(key_name: str, language: str) -> Optional[str]:
+    if key_name.startswith('currency:'):
+        # currency:EUR --> get the name from the CURRENCIES variable
+        # see https://wiki.openstreetmap.org/wiki/Key%3Acurrency
+        # and for example https://taginfo.openstreetmap.org/keys/currency:EUR#values
+        # but there is also currency=EUR (currently not handled)
+        # https://taginfo.openstreetmap.org/keys/currency#values
+        currency = key_name.split(':')
+        if len(currency) > 1:
+            label = fetch_name_from_iso4217(currency[1], language)
+            if label:
+                return label
+            return currency[1]
+
+    language = language.lower()
+    language_short = language.split('-')[0]
+    res = _get_connection("osm_keys_tags.db").execute(
+        "SELECT language, label FROM osm_keys WHERE name=? AND language in (?, ?, 'en')",
+        (key_name, language, language_short),
+    )
+    result = {result[0]: result[1] for result in res.fetchall()}
+    return result.get(language) or result.get(language_short) or result.get('en')
+
+
+@lru_cache(100)
+def fetch_osm_tag_label(tag_key: str, tag_value: str, language: str) -> Optional[str]:
+    language = language.lower()
+    language_short = language.split('-')[0]
+    res = _get_connection("osm_keys_tags.db").execute(
+        "SELECT language, label FROM osm_tags WHERE tag_key=? AND tag_value=? AND language in (?, ?, 'en')",
+        (tag_key, tag_value, language, language_short),
+    )
+    result = {result[0]: result[1] for result in res.fetchall()}
+    return result.get(language) or result.get(language_short) or result.get('en')
+
+
 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::
@@ -119,6 +156,5 @@ USER_AGENTS = _load('useragents.json')
 EXTERNAL_URLS = _load('external_urls.json')
 WIKIDATA_UNITS = _load('wikidata_units.json')
 EXTERNAL_BANGS = _load('external_bangs.json')
-OSM_KEYS_TAGS = _load('osm_keys_tags.json')
 ENGINE_TRAITS = _load('engine_traits.json')
 LOCALES = _load('locales.json')

searx/engines/apple_maps.py

@@ -6,7 +6,7 @@ from time import time
 from urllib.parse import urlencode
 
 from searx.network import get as http_get
-from searx.engines.openstreetmap import get_key_label
+from searx.data import fetch_osm_key_label
 
 about = {
     "website": 'https://www.apple.com/maps/',
@@ -72,7 +72,7 @@ def response(resp):
             telephone = result['telephone']
             links.append(
                 {
-                    'label': get_key_label('phone', user_language),
+                    'label': fetch_osm_key_label('phone', user_language),
                     'url': 'tel:' + telephone,
                     'url_label': telephone,
                 }
@@ -81,7 +81,7 @@ def response(resp):
             url = result['urls'][0]
             links.append(
                 {
-                    'label': get_key_label('website', user_language),
+                    'label': fetch_osm_key_label('website', user_language),
                     'url': url,
                     'url_label': url,
                 }

searx/engines/openstreetmap.py

@@ -10,7 +10,7 @@ from functools import partial
 
 from flask_babel import gettext
 
-from searx.data import OSM_KEYS_TAGS, fetch_name_from_iso4217
+from searx.data import fetch_osm_tag_label, fetch_osm_key_label
 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
@@ -187,14 +187,14 @@ def response(resp):
                 'template': 'map.html',
                 'title': title,
                 'address': address,
-                'address_label': get_key_label('addr', user_language),
+                'address_label': fetch_osm_key_label('addr', user_language),
                 'url': url,
                 'osm': osm,
                 'geojson': geojson,
                 'img_src': img_src,
                 'links': links,
                 'data': data,
-                'type': get_tag_label(result.get('category'), result.get('type', ''), user_language),
+                'type': fetch_osm_tag_label(result.get('category'), result.get('type', ''), user_language),
                 'type_icon': result.get('icon'),
                 'content': '',
                 'longitude': result['lon'],
@@ -367,7 +367,7 @@ def get_links(result, user_language):
             url_label = result.get('wikidata', {}).get('itemLabel') or url_label
         links.append(
             {
-                'label': get_key_label(k, user_language),
+                'label': fetch_osm_key_label(k, user_language),
                 'url': url,
                 'url_label': url_label,
             }
@@ -389,7 +389,7 @@ def get_data(result, user_language, ignore_keys):
             continue
         if get_key_rank(k) is None:
             continue
-        k_label = get_key_label(k, user_language)
+        k_label = fetch_osm_key_label(k, user_language)
         if k_label:
             data.append(
                 {
@@ -412,51 +412,3 @@ def get_key_rank(k):
         # "payment:*" in KEY_ORDER matches "payment:cash", "payment:debit card", etc...
         key_rank = KEY_RANKS.get(k.split(':')[0] + ':*')
     return key_rank
-
-
-def get_label(labels, lang):
-    """Get label from labels in OSM_KEYS_TAGS
-
-    in OSM_KEYS_TAGS, labels have key == '*'
-    """
-    tag_label = labels.get(lang.lower())
-    if tag_label is None:
-        # example: if 'zh-hk' is not found, check 'zh'
-        tag_label = labels.get(lang.split('-')[0])
-    if tag_label is None and lang != 'en':
-        # example: if 'zh' is not found, check 'en'
-        tag_label = labels.get('en')
-    if tag_label is None and len(labels.values()) > 0:
-        # example: if still not found, use the first entry
-        tag_label = labels.values()[0]
-    return tag_label
-
-
-def get_tag_label(tag_category, tag_name, lang):
-    """Get tag label from OSM_KEYS_TAGS"""
-    tag_name = '' if tag_name is None else tag_name
-    tag_labels = OSM_KEYS_TAGS['tags'].get(tag_category, {}).get(tag_name, {})
-    return get_label(tag_labels, lang)
-
-
-def get_key_label(key_name, lang):
-    """Get key label from OSM_KEYS_TAGS"""
-    if key_name.startswith('currency:'):
-        # currency:EUR --> get the name from the CURRENCIES variable
-        # see https://wiki.openstreetmap.org/wiki/Key%3Acurrency
-        # and for example https://taginfo.openstreetmap.org/keys/currency:EUR#values
-        # but there is also currency=EUR (currently not handled)
-        # https://taginfo.openstreetmap.org/keys/currency#values
-        currency = key_name.split(':')
-        if len(currency) > 1:
-            label = fetch_name_from_iso4217(currency[1], lang)
-            if label:
-                return label
-            return currency[1]
-
-    labels = OSM_KEYS_TAGS['keys']
-    for k in key_name.split(':') + ['*']:
-        labels = labels.get(k)
-        if labels is None:
-            return None
-    return get_label(labels, lang)

searxng_extra/update/update_osm_keys_tags.py

@@ -42,8 +42,9 @@ Output file: :origin:`searx/data/osm_keys_tags` (:origin:`CI Update data ...
 
 """
 
-import json
-import collections
+import csv
+import sqlite3
+from pathlib import Path
 
 from searx.network import set_timeout_for_thread
 from searx.engines import wikidata, set_loggers
@@ -51,7 +52,9 @@ from searx.sxng_locales import sxng_locales
 from searx.engines.openstreetmap import get_key_rank, VALUE_TO_LINK
 from searx.data import data_dir
 
-DATA_FILE = data_dir / 'osm_keys_tags.json'
+DATABASE_FILE = data_dir / 'osm_keys_tags.db'
+CSV_KEYS_FILE = data_dir / 'dumps' / 'osm_keys.csv'
+CSV_TAGS_FILE = data_dir / 'dumps' / 'osm_tags.csv'
 
 set_loggers(wikidata, 'wikidata')
 
@@ -78,42 +81,39 @@ ORDER BY ?key ?item ?itemLabel
 
 LANGUAGES = [l[0].lower() for l in sxng_locales]
 
-PRESET_KEYS = {
-    ('wikidata',): {'en': 'Wikidata'},
-    ('wikipedia',): {'en': 'Wikipedia'},
-    ('email',): {'en': 'Email'},
-    ('facebook',): {'en': 'Facebook'},
-    ('fax',): {'en': 'Fax'},
-    ('internet_access', 'ssid'): {'en': 'Wi-Fi'},
-}
+PRESET_KEYS = [
+    ["wikidata", "en", "Wikidata"],
+    ["wikipedia", "en", "Wikipedia"],
+    ["email", "en", "email"],
+    ["facebook", "en", "facebook"],
+    ["fax", "en", "Fax"],
+    ["internet_access:ssid", "en", "Wi-Fi"],
+]
 
 INCLUDED_KEYS = {('addr',)}
 
 
-def get_preset_keys():
-    results = collections.OrderedDict()
-    for keys, value in PRESET_KEYS.items():
-        r = results
-        for k in keys:
-            r = r.setdefault(k, {})
-        r.setdefault('*', value)
-    return results
-
-
 def get_keys():
-    results = get_preset_keys()
+    result_keys = set()
+    results = PRESET_KEYS.copy()
     response = wikidata.send_wikidata_query(SPARQL_KEYS_REQUEST)
 
     for key in response['results']['bindings']:
         keys = key['key']['value'].split(':')[1:]
+        label = key['itemLabel']['value'].lower()
+        lang = key['itemLabel']['xml:lang']
+
+        if lang not in LANGUAGES:
+            continue
+
         if keys[0] == 'currency' and len(keys) > 1:
             # special case in openstreetmap.py
             continue
         if keys[0] == 'contact' and len(keys) > 1:
-            # label for the key "contact.email" is "Email"
-            # whatever the language
-            r = results.setdefault('contact', {})
-            r[keys[1]] = {'*': {'en': keys[1]}}
+            if lang == "en":
+                # label for the key "contact.email" is "Email"
+                # whatever the language
+                results.append((":".join(keys), "en", keys[1]))
             continue
         if tuple(keys) in PRESET_KEYS:
             # skip presets (already set above)
@@ -125,40 +125,46 @@ def get_keys():
         ):
             # keep only keys that will be displayed by openstreetmap.py
             continue
-        label = key['itemLabel']['value'].lower()
-        lang = key['itemLabel']['xml:lang']
-        r = results
-        for k in keys:
-            r = r.setdefault(k, {})
-        r = r.setdefault('*', {})
-        if lang in LANGUAGES:
-            r.setdefault(lang, label)
 
-    # special cases
-    results['delivery']['covid19']['*'].clear()
-    for k, v in results['delivery']['*'].items():
-        results['delivery']['covid19']['*'][k] = v + ' (COVID19)'
+        entry = (":".join(keys), lang, label)
+        entry_key = (entry[0], entry[1])
+        if entry_key not in result_keys:
+            results.append(entry)
+            result_keys.add(entry_key)
 
-    results['opening_hours']['covid19']['*'].clear()
-    for k, v in results['opening_hours']['*'].items():
-        results['opening_hours']['covid19']['*'][k] = v + ' (COVID19)'
+    # special cases
+    results = [entry for entry in results if entry[0] != 'delivery:covid19']
+    results.extend(
+        [['delivery:covid19', entry[1], entry[2] + ' (COVID19)'] for entry in results if entry[0] == 'delivery']
+    )
+
+    results = [entry for entry in results if entry[0] != 'opening_hours:covid19']
+    results.extend(
+        [
+            ['opening_hours:covid19', entry[1], entry[2] + ' (COVID19)']
+            for entry in results
+            if entry[0] == 'opening_hours'
+        ]
+    )
 
     return results
 
 
 def get_tags():
-    results = collections.OrderedDict()
+    results = []
     response = wikidata.send_wikidata_query(SPARQL_TAGS_REQUEST)
     for tag in response['results']['bindings']:
-        tag_names = tag['tag']['value'].split(':')[1].split('=')
-        if len(tag_names) == 2:
-            tag_category, tag_type = tag_names
-        else:
-            tag_category, tag_type = tag_names[0], ''
+        try:
+            tag_key, tag_value = tag['tag']['value'].split('=')
+            if tag_key.startswith("Tag:"):
+                tag_key = tag_key[4:]
+        except ValueError:
+            print("ignore tag", tag['tag']['value'])
+            continue
         label = tag['itemLabel']['value'].lower()
         lang = tag['itemLabel']['xml:lang']
         if lang in LANGUAGES:
-            results.setdefault(tag_category, {}).setdefault(tag_type, {}).setdefault(lang, label)
+            results.append((tag_key, tag_value, lang, label))
     return results
 
 
@@ -206,9 +212,30 @@ def optimize_keys(data):
 if __name__ == '__main__':
 
     set_timeout_for_thread(60)
-    result = {
-        'keys': optimize_keys(get_keys()),
-        'tags': optimize_tags(get_tags()),
-    }
-    with DATA_FILE.open('w', encoding="utf8") as f:
-        json.dump(result, f, indent=4, sort_keys=True, ensure_ascii=False)
+    osm_keys = get_keys()
+    osm_tags = get_tags()
+
+    osm_keys.sort(key=lambda item: (item[0], item[1]))
+    osm_tags.sort(key=lambda item: (item[0], item[1]))
+
+    Path(DATABASE_FILE).unlink(missing_ok=True)
+    with sqlite3.connect(DATABASE_FILE) as con:
+        cur = con.cursor()
+        cur.execute("CREATE TABLE osm_keys(name, language, label)")
+        cur.executemany("INSERT INTO osm_keys VALUES(?, ?, ?)", osm_keys)
+        cur.execute("CREATE INDEX index_osm_keys ON osm_keys('name', 'language')")
+        cur.execute("CREATE TABLE osm_tags(tag_key, tag_value, language, label)")
+        cur.executemany("INSERT INTO osm_tags VALUES(?, ?, ?, ?)", osm_tags)
+        cur.execute("CREATE INDEX index_osm_tags ON osm_tags('tag_key', 'tag_value', 'language')")
+        con.commit()
+
+    with CSV_KEYS_FILE.open('w', encoding="utf8") as f:
+        w = csv.writer(f, quoting=csv.QUOTE_NONNUMERIC)
+        w.writerow(["name", "language", "label"])
+        for row in osm_keys:
+            w.writerow(row)
+    with CSV_TAGS_FILE.open('w', encoding="utf8") as f:
+        w = csv.writer(f, quoting=csv.QUOTE_NONNUMERIC)
+        w.writerow(["tag_key", "tag_value", "language", "label"])
+        for row in osm_tags:
+            w.writerow(row)