mirror of
https://github.com/searxng/searxng.git
synced 2024-06-11 01:39:35 +00:00
242 lines
8 KiB
Python
Executable file
242 lines
8 KiB
Python
Executable file
#!/usr/bin/env python
|
|
# SPDX-License-Identifier: AGPL-3.0-or-later
|
|
"""Fetch OSM keys and tags.
|
|
|
|
To get the i18n names, the scripts uses `Wikidata Query Service`_ instead of for
|
|
example `OSM tags API`_ (side note: the actual change log from
|
|
map.atownsend.org.uk_ might be useful to normalize OSM tags).
|
|
|
|
Output file: :origin:`searx/data/osm_keys_tags` (:origin:`CI Update data ...
|
|
<.github/workflows/data-update.yml>`).
|
|
|
|
.. _Wikidata Query Service: https://query.wikidata.org/
|
|
.. _OSM tags API: https://taginfo.openstreetmap.org/taginfo/apidoc
|
|
.. _map.atownsend.org.uk: https://map.atownsend.org.uk/maps/map/changelog.html
|
|
|
|
:py:obj:`SPARQL_TAGS_REQUEST` :
|
|
Wikidata SPARQL query that returns *type-categories* and *types*. The
|
|
returned tag is ``Tag:{category}={type}`` (see :py:func:`get_tags`).
|
|
Example:
|
|
|
|
- https://taginfo.openstreetmap.org/tags/building=house#overview
|
|
- https://wiki.openstreetmap.org/wiki/Tag:building%3Dhouse
|
|
at the bottom of the infobox (right side), there is a link to wikidata:
|
|
https://www.wikidata.org/wiki/Q3947
|
|
see property "OpenStreetMap tag or key" (P1282)
|
|
- https://wiki.openstreetmap.org/wiki/Tag%3Abuilding%3Dbungalow
|
|
https://www.wikidata.org/wiki/Q850107
|
|
|
|
:py:obj:`SPARQL_KEYS_REQUEST` :
|
|
Wikidata SPARQL query that returns *keys*. Example with "payment":
|
|
|
|
- https://wiki.openstreetmap.org/wiki/Key%3Apayment
|
|
at the bottom of infobox (right side), there is a link to wikidata:
|
|
https://www.wikidata.org/wiki/Q1148747
|
|
link made using the "OpenStreetMap tag or key" property (P1282)
|
|
to be confirm: there is a one wiki page per key ?
|
|
- https://taginfo.openstreetmap.org/keys/payment#values
|
|
- https://taginfo.openstreetmap.org/keys/payment:cash#values
|
|
|
|
``rdfs:label`` get all the labels without language selection
|
|
(as opposed to SERVICE ``wikibase:label``).
|
|
|
|
"""
|
|
|
|
import csv
|
|
import sqlite3
|
|
from pathlib import Path
|
|
|
|
from searx.network import set_timeout_for_thread
|
|
from searx.engines import wikidata, set_loggers
|
|
from searx.sxng_locales import sxng_locales
|
|
from searx.engines.openstreetmap import get_key_rank, VALUE_TO_LINK
|
|
from searx.data import data_dir
|
|
|
|
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')
|
|
|
|
|
|
SPARQL_TAGS_REQUEST = """
|
|
SELECT ?tag ?item ?itemLabel WHERE {
|
|
?item wdt:P1282 ?tag .
|
|
?item rdfs:label ?itemLabel .
|
|
FILTER(STRSTARTS(?tag, 'Tag'))
|
|
}
|
|
GROUP BY ?tag ?item ?itemLabel
|
|
ORDER BY ?tag ?item ?itemLabel
|
|
"""
|
|
|
|
SPARQL_KEYS_REQUEST = """
|
|
SELECT ?key ?item ?itemLabel WHERE {
|
|
?item wdt:P1282 ?key .
|
|
?item rdfs:label ?itemLabel .
|
|
FILTER(STRSTARTS(?key, 'Key'))
|
|
}
|
|
GROUP BY ?key ?item ?itemLabel
|
|
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"],
|
|
]
|
|
|
|
INCLUDED_KEYS = {('addr',)}
|
|
|
|
|
|
def get_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:
|
|
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)
|
|
continue
|
|
if (
|
|
get_key_rank(':'.join(keys)) is None
|
|
and ':'.join(keys) not in VALUE_TO_LINK
|
|
and tuple(keys) not in INCLUDED_KEYS
|
|
):
|
|
# keep only keys that will be displayed by openstreetmap.py
|
|
continue
|
|
|
|
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)
|
|
|
|
# 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 = []
|
|
response = wikidata.send_wikidata_query(SPARQL_TAGS_REQUEST)
|
|
for tag in response['results']['bindings']:
|
|
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.append((tag_key, tag_value, lang, label))
|
|
return results
|
|
|
|
|
|
def optimize_data_lang(translations):
|
|
language_to_delete = []
|
|
# remove "zh-hk" entry if the value is the same as "zh"
|
|
# same for "en-ca" / "en" etc...
|
|
for language in translations:
|
|
if '-' in language:
|
|
base_language = language.split('-')[0]
|
|
if translations.get(base_language) == translations.get(language):
|
|
language_to_delete.append(language)
|
|
|
|
for language in language_to_delete:
|
|
del translations[language]
|
|
language_to_delete = []
|
|
|
|
# remove entries that have the same value than the "en" entry
|
|
value_en = translations.get('en')
|
|
if value_en:
|
|
for language, value in translations.items():
|
|
if language != 'en' and value == value_en:
|
|
language_to_delete.append(language)
|
|
|
|
for language in language_to_delete:
|
|
del translations[language]
|
|
|
|
|
|
def optimize_tags(data):
|
|
for v in data.values():
|
|
for translations in v.values():
|
|
optimize_data_lang(translations)
|
|
return data
|
|
|
|
|
|
def optimize_keys(data):
|
|
for k, v in data.items():
|
|
if k == '*':
|
|
optimize_data_lang(v)
|
|
elif isinstance(v, dict):
|
|
optimize_keys(v)
|
|
return data
|
|
|
|
|
|
if __name__ == '__main__':
|
|
|
|
set_timeout_for_thread(60)
|
|
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)
|