mirror of
				https://github.com/searxng/searxng
				synced 2024-01-01 19:24:07 +01:00 
			
		
		
		
	[fix] searx_extra/update scripts: set_loggers(wikidata, 'wikidata') (#331)
* [fix] searx_extra/update scripts: set_loggers(wikidata, 'wikidata')
To test use::
    ./manage pyenv.cmd searx_extra/update/update_currencies.py
    ./manage pyenv.cmd searx_extra/update/update_osm_keys_tags.py
    ./manage pyenv.cmd searx_extra/update/update_wikidata_units.py
The script `update_engine_descriptions.py` seems to have some issues not related
to this patch.
    ./manage pyenv.cmd python -m pip install -U pycld3
    ./manage pyenv.cmd searx_extra/update/update_engine_descriptions.py
Closes: https://github.com/searxng/searxng/issues/328
Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>
			
			
This commit is contained in:
		
							parent
							
								
									d95c46d646
								
							
						
					
					
						commit
						55b30b5b49
					
				
					 4 changed files with 28 additions and 15 deletions
				
			
		| 
						 | 
				
			
			@ -10,8 +10,9 @@ from os.path import realpath, dirname, join
 | 
			
		|||
 | 
			
		||||
from searx import searx_dir
 | 
			
		||||
from searx.locales import LOCALE_NAMES
 | 
			
		||||
from searx.engines.wikidata import send_wikidata_query
 | 
			
		||||
from searx.engines import wikidata, set_loggers
 | 
			
		||||
 | 
			
		||||
set_loggers(wikidata, 'wikidata')
 | 
			
		||||
 | 
			
		||||
# ORDER BY (with all the query fields) is important to keep a deterministic result order
 | 
			
		||||
# so multiple invokation of this script doesn't change currencies.json
 | 
			
		||||
| 
						 | 
				
			
			@ -83,7 +84,9 @@ def add_currency_label(db, label, iso4217, language):
 | 
			
		|||
 | 
			
		||||
 | 
			
		||||
def wikidata_request_result_iterator(request):
 | 
			
		||||
    result = send_wikidata_query(request.replace('%LANGUAGES_SPARQL%', LANGUAGES_SPARQL))
 | 
			
		||||
    result = wikidata.send_wikidata_query(
 | 
			
		||||
        request.replace('%LANGUAGES_SPARQL%', LANGUAGES_SPARQL)
 | 
			
		||||
    )
 | 
			
		||||
    if result is not None:
 | 
			
		||||
        for r in result['results']['bindings']:
 | 
			
		||||
            yield r
 | 
			
		||||
| 
						 | 
				
			
			
 | 
			
		|||
| 
						 | 
				
			
			@ -6,13 +6,15 @@ from urllib.parse import quote, urlparse
 | 
			
		|||
import detect_language
 | 
			
		||||
from lxml.html import fromstring
 | 
			
		||||
 | 
			
		||||
from searx.engines.wikidata import send_wikidata_query
 | 
			
		||||
from searx.engines import wikidata, set_loggers
 | 
			
		||||
from searx.utils import extract_text
 | 
			
		||||
from searx.locales import LOCALE_NAMES
 | 
			
		||||
import searx
 | 
			
		||||
import searx.search
 | 
			
		||||
import searx.network
 | 
			
		||||
 | 
			
		||||
set_loggers(wikidata, 'wikidata')
 | 
			
		||||
 | 
			
		||||
SPARQL_WIKIPEDIA_ARTICLE = """
 | 
			
		||||
SELECT DISTINCT ?item ?name
 | 
			
		||||
WHERE {
 | 
			
		||||
| 
						 | 
				
			
			@ -128,9 +130,11 @@ def initialize():
 | 
			
		|||
 | 
			
		||||
def fetch_wikidata_descriptions():
 | 
			
		||||
    global IDS
 | 
			
		||||
    result = send_wikidata_query(SPARQL_DESCRIPTION
 | 
			
		||||
                                 .replace('%IDS%', IDS)
 | 
			
		||||
                                 .replace('%LANGUAGES_SPARQL%', LANGUAGES_SPARQL))
 | 
			
		||||
    result = wikidata.send_wikidata_query(
 | 
			
		||||
        SPARQL_DESCRIPTION
 | 
			
		||||
        .replace('%IDS%', IDS)
 | 
			
		||||
        .replace('%LANGUAGES_SPARQL%', LANGUAGES_SPARQL)
 | 
			
		||||
    )
 | 
			
		||||
    if result is not None:
 | 
			
		||||
        for binding in result['results']['bindings']:
 | 
			
		||||
            wikidata_id = binding['item']['value'].replace('http://www.wikidata.org/entity/', '')
 | 
			
		||||
| 
						 | 
				
			
			@ -143,9 +147,11 @@ def fetch_wikidata_descriptions():
 | 
			
		|||
 | 
			
		||||
def fetch_wikipedia_descriptions():
 | 
			
		||||
    global IDS
 | 
			
		||||
    result = send_wikidata_query(SPARQL_WIKIPEDIA_ARTICLE
 | 
			
		||||
                                 .replace('%IDS%', IDS)
 | 
			
		||||
                                 .replace('%LANGUAGES_SPARQL%', LANGUAGES_SPARQL))
 | 
			
		||||
    result = wikidata.send_wikidata_query(
 | 
			
		||||
        SPARQL_WIKIPEDIA_ARTICLE
 | 
			
		||||
        .replace('%IDS%', IDS)
 | 
			
		||||
        .replace('%LANGUAGES_SPARQL%', LANGUAGES_SPARQL)
 | 
			
		||||
    )
 | 
			
		||||
    if result is not None:
 | 
			
		||||
        for binding in result['results']['bindings']:
 | 
			
		||||
            wikidata_id = binding['item']['value'].replace('http://www.wikidata.org/entity/', '')
 | 
			
		||||
| 
						 | 
				
			
			
 | 
			
		|||
| 
						 | 
				
			
			@ -46,10 +46,13 @@ from pathlib import Path
 | 
			
		|||
 | 
			
		||||
from searx import searx_dir
 | 
			
		||||
from searx.network import set_timeout_for_thread
 | 
			
		||||
from searx.engines.wikidata import send_wikidata_query
 | 
			
		||||
from searx.engines import wikidata, set_loggers
 | 
			
		||||
from searx.languages import language_codes
 | 
			
		||||
from searx.engines.openstreetmap import get_key_rank, VALUE_TO_LINK
 | 
			
		||||
 | 
			
		||||
set_loggers(wikidata, 'wikidata')
 | 
			
		||||
 | 
			
		||||
 | 
			
		||||
SPARQL_TAGS_REQUEST = """
 | 
			
		||||
SELECT ?tag ?item ?itemLabel WHERE {
 | 
			
		||||
  ?item wdt:P1282 ?tag .
 | 
			
		||||
| 
						 | 
				
			
			@ -96,7 +99,7 @@ def get_preset_keys():
 | 
			
		|||
 | 
			
		||||
def get_keys():
 | 
			
		||||
    results = get_preset_keys()
 | 
			
		||||
    response = send_wikidata_query(SPARQL_KEYS_REQUEST)
 | 
			
		||||
    response = wikidata.send_wikidata_query(SPARQL_KEYS_REQUEST)
 | 
			
		||||
 | 
			
		||||
    for key in response['results']['bindings']:
 | 
			
		||||
        keys = key['key']['value'].split(':')[1:]
 | 
			
		||||
| 
						 | 
				
			
			@ -144,7 +147,7 @@ def get_keys():
 | 
			
		|||
 | 
			
		||||
def get_tags():
 | 
			
		||||
    results = collections.OrderedDict()
 | 
			
		||||
    response = send_wikidata_query(SPARQL_TAGS_REQUEST)
 | 
			
		||||
    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:
 | 
			
		||||
| 
						 | 
				
			
			
 | 
			
		|||
| 
						 | 
				
			
			@ -7,13 +7,14 @@ import collections
 | 
			
		|||
from os.path import join
 | 
			
		||||
 | 
			
		||||
from searx import searx_dir
 | 
			
		||||
from searx.engines.wikidata import send_wikidata_query
 | 
			
		||||
from searx.engines import wikidata, set_loggers
 | 
			
		||||
 | 
			
		||||
set_loggers(wikidata, 'wikidata')
 | 
			
		||||
 | 
			
		||||
# the response contains duplicate ?item with the different ?symbol
 | 
			
		||||
# "ORDER BY ?item DESC(?rank) ?symbol" provides a deterministic result
 | 
			
		||||
# even if a ?item has different ?symbol of the same rank.
 | 
			
		||||
# A deterministic result 
 | 
			
		||||
# A deterministic result
 | 
			
		||||
# see:
 | 
			
		||||
# * https://www.wikidata.org/wiki/Help:Ranking
 | 
			
		||||
# * https://www.mediawiki.org/wiki/Wikibase/Indexing/RDF_Dump_Format ("Statement representation" section)
 | 
			
		||||
| 
						 | 
				
			
			@ -36,7 +37,7 @@ ORDER BY ?item DESC(?rank) ?symbol
 | 
			
		|||
 | 
			
		||||
def get_data():
 | 
			
		||||
    results = collections.OrderedDict()
 | 
			
		||||
    response = send_wikidata_query(SARQL_REQUEST)
 | 
			
		||||
    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']
 | 
			
		||||
| 
						 | 
				
			
			
 | 
			
		|||
		Loading…
	
	Add table
		
		Reference in a new issue