on master help sign in
Merge pull request #2275 from dalf/mod-ahmia-filter
d3d50efAdam Tauber
searx/data/__init__.py+6−1
@@ -2,7 +2,7 @@ import json
2from pathlib import Path
2from pathlib import Path
3
3
4
4
5−__init__ = ['ENGINES_LANGUGAGES', 'CURRENCIES', 'USER_AGENTS', 'bangs_loader']
5+__init__ = ['ENGINES_LANGUGAGES', 'CURRENCIES', 'USER_AGENTS', 'bangs_loader', 'ahmia_blacklist_loader']
6data_dir = Path(__file__).parent
6data_dir = Path(__file__).parent
7
7
8
8
@@ -16,6 +16,11 @@ def bangs_loader():
16 return load('bangs.json')
16 return load('bangs.json')
17
17
18
18
19+def ahmia_blacklist_loader():
20+ with open(str(data_dir / 'ahmia_blacklist.txt'), encoding='utf-8') as fd:
21+ return fd.read().split()
22+
23+
19ENGINES_LANGUAGES = load('engines_languages.json')
24ENGINES_LANGUAGES = load('engines_languages.json')
20CURRENCIES = load('currencies.json')
25CURRENCIES = load('currencies.json')
21USER_AGENTS = load('useragents.json')
26USER_AGENTS = load('useragents.json')
searx/plugins/ahmia_filter.py+4−7
@@ -3,9 +3,7 @@
3'''
3'''
4
4
5from hashlib import md5
5from hashlib import md5
6−from os.path import join
6+from searx.data import ahmia_blacklist_loader
7−from urllib.parse import urlparse
8−from searx import searx_dir
9
7
10name = "Ahmia blacklist"
8name = "Ahmia blacklist"
11description = "Filter out onion results that appear in Ahmia's blacklist. (See https://ahmia.fi/blacklist)"
9description = "Filter out onion results that appear in Ahmia's blacklist. (See https://ahmia.fi/blacklist)"
@@ -18,15 +16,14 @@ ahmia_blacklist = None
18def get_ahmia_blacklist():
16def get_ahmia_blacklist():
19 global ahmia_blacklist
17 global ahmia_blacklist
20 if not ahmia_blacklist:
18 if not ahmia_blacklist:
21− with open(join(join(searx_dir, "data"), "ahmia_blacklist.txt"), 'r') as f:
22− ahmia_blacklist = f.read().split()
19+ ahmia_blacklist = ahmia_blacklist_loader()
23 return ahmia_blacklist
20 return ahmia_blacklist
24
21
25
22
26def not_blacklisted(result):
23def not_blacklisted(result):
27− if not result.get('is_onion'):
24+ if not result.get('is_onion') or not result.get('parsed_url'):
28 return True
25 return True
29− result_hash = md5(urlparse(result.get('url')).hostname.encode()).hexdigest()
26+ result_hash = md5(result['parsed_url'].hostname.encode()).hexdigest()
30 return result_hash not in get_ahmia_blacklist()
27 return result_hash not in get_ahmia_blacklist()
31
28
32
29