Merge pull request #839 from dalf/docs_searx_utils

[mod] add documentation about searx.utils
This commit is contained in:
Markus Heiser 2022-02-04 21:28:27 +00:00 committed by GitHub
commit 24e2acf88e
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
3 changed files with 118 additions and 104 deletions

8
docs/src/searx.utils.rst Normal file
View file

@ -0,0 +1,8 @@
.. _searx.utils:
=================================
Utility functions for the engines
=================================
.. automodule:: searx.utils
:members:

View file

@ -1,7 +1,15 @@
# -*- coding: utf-8 -*-
# SPDX-License-Identifier: AGPL-3.0-or-later
# lint: pylint
# pyright: basic
"""Utility functions for the engines
"""
import re
import importlib
import importlib.util
import types
from typing import Optional, Union, Any, Set, List, Dict, MutableMapping, Tuple, Callable
from numbers import Number
from os.path import splitext, join
from random import choice
@ -23,42 +31,56 @@ from searx import logger
logger = logger.getChild('utils')
blocked_tags = ('script', 'style')
XPathSpecType = Union[str, XPath]
ecma_unescape4_re = re.compile(r'%u([0-9a-fA-F]{4})', re.UNICODE)
ecma_unescape2_re = re.compile(r'%([0-9a-fA-F]{2})', re.UNICODE)
_BLOCKED_TAGS = ('script', 'style')
xpath_cache = dict()
lang_to_lc_cache = dict()
_ECMA_UNESCAPE4_RE = re.compile(r'%u([0-9a-fA-F]{4})', re.UNICODE)
_ECMA_UNESCAPE2_RE = re.compile(r'%([0-9a-fA-F]{2})', re.UNICODE)
_STORAGE_UNIT_VALUE: Dict[str, int] = {
'TB': 1024 * 1024 * 1024 * 1024,
'GB': 1024 * 1024 * 1024,
'MB': 1024 * 1024,
'TiB': 1000 * 1000 * 1000 * 1000,
'MiB': 1000 * 1000,
'KiB': 1000,
}
_XPATH_CACHE: Dict[str, XPath] = {}
_LANG_TO_LC_CACHE: Dict[str, Dict[str, str]] = {}
class NotSetClass:
pass
class _NotSetClass: # pylint: disable=too-few-public-methods
"""Internal class for this module, do not create instance of this class.
Replace the None value, allow explicitly pass None as a function argument"""
NOTSET = NotSetClass()
_NOTSET = _NotSetClass()
def searx_useragent():
def searx_useragent() -> str:
"""Return the searx User Agent"""
return 'searx/{searx_version} {suffix}'.format(
searx_version=VERSION_TAG, suffix=settings['outgoing']['useragent_suffix']
).strip()
def gen_useragent(os=None):
def gen_useragent(os_string: str = None) -> str:
"""Return a random browser User Agent
See searx/data/useragents.json
"""
return str(USER_AGENTS['ua'].format(os=os or choice(USER_AGENTS['os']), version=choice(USER_AGENTS['versions'])))
return USER_AGENTS['ua'].format(os=os_string or choice(USER_AGENTS['os']), version=choice(USER_AGENTS['versions']))
class HTMLTextExtractorException(Exception):
pass
class _HTMLTextExtractorException(Exception):
"""Internal exception raised when the HTML is invalid"""
class HTMLTextExtractor(HTMLParser): # pylint: disable=W0223 # (see https://bugs.python.org/issue31844)
class _HTMLTextExtractor(HTMLParser): # pylint: disable=W0223 # (see https://bugs.python.org/issue31844)
"""Internal class to extract text from HTML"""
def __init__(self):
HTMLParser.__init__(self)
self.result = []
@ -72,12 +94,12 @@ class HTMLTextExtractor(HTMLParser): # pylint: disable=W0223 # (see https://bu
return
if tag != self.tags[-1]:
raise HTMLTextExtractorException()
raise _HTMLTextExtractorException()
self.tags.pop()
def is_valid_tag(self):
return not self.tags or self.tags[-1] not in blocked_tags
return not self.tags or self.tags[-1] not in _BLOCKED_TAGS
def handle_data(self, data):
if not self.is_valid_tag():
@ -104,7 +126,7 @@ class HTMLTextExtractor(HTMLParser): # pylint: disable=W0223 # (see https://bu
return ''.join(self.result).strip()
def html_to_text(html_str):
def html_to_text(html_str: str) -> str:
"""Extract text from a HTML string
Args:
@ -122,15 +144,15 @@ def html_to_text(html_str):
"""
html_str = html_str.replace('\n', ' ')
html_str = ' '.join(html_str.split())
s = HTMLTextExtractor()
s = _HTMLTextExtractor()
try:
s.feed(html_str)
except HTMLTextExtractorException:
except _HTMLTextExtractorException:
logger.debug("HTMLTextExtractor: invalid HTML\n%s", html_str)
return s.get_text()
def extract_text(xpath_results, allow_none=False):
def extract_text(xpath_results, allow_none: bool = False) -> Optional[str]:
"""Extract text from a lxml result
* if xpath_results is list, extract the text from each result and concat the list
@ -142,24 +164,23 @@ def extract_text(xpath_results, allow_none=False):
# it's list of result : concat everything using recursive call
result = ''
for e in xpath_results:
result = result + extract_text(e)
result = result + (extract_text(e) or '')
return result.strip()
elif isinstance(xpath_results, ElementBase):
if isinstance(xpath_results, ElementBase):
# it's a element
text = html.tostring(xpath_results, encoding='unicode', method='text', with_tail=False)
text: str = html.tostring(xpath_results, encoding='unicode', method='text', with_tail=False)
text = text.strip().replace('\n', ' ')
return ' '.join(text.split())
elif isinstance(xpath_results, (_ElementStringResult, _ElementUnicodeResult, str, Number, bool)):
if isinstance(xpath_results, (_ElementStringResult, _ElementUnicodeResult, str, Number, bool)):
return str(xpath_results)
elif xpath_results is None and allow_none:
if xpath_results is None and allow_none:
return None
elif xpath_results is None and not allow_none:
if xpath_results is None and not allow_none:
raise ValueError('extract_text(None, allow_none=False)')
else:
raise ValueError('unsupported type')
raise ValueError('unsupported type')
def normalize_url(url, base_url):
def normalize_url(url: str, base_url: str) -> str:
"""Normalize URL: add protocol, join URL with base_url, add trailing slash if there is no path
Args:
@ -209,7 +230,7 @@ def normalize_url(url, base_url):
return url
def extract_url(xpath_results, base_url):
def extract_url(xpath_results, base_url) -> str:
"""Extract and normalize URL from lxml Element
Args:
@ -245,10 +266,12 @@ def extract_url(xpath_results, base_url):
raise ValueError('Empty url resultset')
url = extract_text(xpath_results)
return normalize_url(url, base_url)
if url:
return normalize_url(url, base_url)
raise ValueError('URL not found')
def dict_subset(d, properties):
def dict_subset(dictionnary: MutableMapping, properties: Set[str]) -> Dict:
"""Extract a subset of a dict
Examples:
@ -257,10 +280,10 @@ def dict_subset(d, properties):
>>> >> dict_subset({'A': 'a', 'B': 'b', 'C': 'c'}, ['A', 'D'])
{'A': 'a'}
"""
return {k: d[k] for k in properties if k in d}
return {k: dictionnary[k] for k in properties if k in dictionnary}
def get_torrent_size(filesize, filesize_multiplier):
def get_torrent_size(filesize: str, filesize_multiplier: str) -> Optional[int]:
"""
Args:
@ -277,39 +300,20 @@ def get_torrent_size(filesize, filesize_multiplier):
3140000
"""
try:
filesize = float(filesize)
if filesize_multiplier == 'TB':
filesize = int(filesize * 1024 * 1024 * 1024 * 1024)
elif filesize_multiplier == 'GB':
filesize = int(filesize * 1024 * 1024 * 1024)
elif filesize_multiplier == 'MB':
filesize = int(filesize * 1024 * 1024)
elif filesize_multiplier == 'KB':
filesize = int(filesize * 1024)
elif filesize_multiplier == 'TiB':
filesize = int(filesize * 1000 * 1000 * 1000 * 1000)
elif filesize_multiplier == 'GiB':
filesize = int(filesize * 1000 * 1000 * 1000)
elif filesize_multiplier == 'MiB':
filesize = int(filesize * 1000 * 1000)
elif filesize_multiplier == 'KiB':
filesize = int(filesize * 1000)
multiplier = _STORAGE_UNIT_VALUE.get(filesize_multiplier, 1)
return int(float(filesize) * multiplier)
except ValueError:
filesize = None
return filesize
return None
def convert_str_to_int(number_str):
def convert_str_to_int(number_str: str) -> int:
"""Convert number_str to int or 0 if number_str is not a number."""
if number_str.isdigit():
return int(number_str)
else:
return 0
return 0
def int_or_zero(num):
def int_or_zero(num: Union[List[str], str]) -> int:
"""Convert num to int or 0. num can be either a str or a list.
If num is a list, the first element is converted to int (or return 0 if the list is empty).
If num is a str, see convert_str_to_int
@ -321,12 +325,12 @@ def int_or_zero(num):
return convert_str_to_int(num)
def is_valid_lang(lang):
def is_valid_lang(lang) -> Optional[Tuple[bool, str, str]]:
"""Return language code and name if lang describe a language.
Examples:
>>> is_valid_lang('zz')
False
None
>>> is_valid_lang('uk')
(True, 'uk', 'ukrainian')
>>> is_valid_lang(b'uk')
@ -346,28 +350,27 @@ def is_valid_lang(lang):
for l in language_codes:
if l[0][:2] == lang:
return (True, l[0][:2], l[3].lower())
return False
else:
for l in language_codes:
if l[1].lower() == lang or l[3].lower() == lang:
return (True, l[0][:2], l[3].lower())
return False
return None
for l in language_codes:
if l[1].lower() == lang or l[3].lower() == lang:
return (True, l[0][:2], l[3].lower())
return None
def _get_lang_to_lc_dict(lang_list):
def _get_lang_to_lc_dict(lang_list: List[str]) -> Dict[str, str]:
key = str(lang_list)
value = lang_to_lc_cache.get(key, None)
value = _LANG_TO_LC_CACHE.get(key, None)
if value is None:
value = dict()
for lc in lang_list:
value.setdefault(lc.split('-')[0], lc)
lang_to_lc_cache[key] = value
value = {}
for lang in lang_list:
value.setdefault(lang.split('-')[0], lang)
_LANG_TO_LC_CACHE[key] = value
return value
# babel's get_global contains all sorts of miscellaneous locale and territory related data
# see get_global in: https://github.com/python-babel/babel/blob/master/babel/core.py
def _get_from_babel(lang_code, key):
def _get_from_babel(lang_code: str, key: str):
match = get_global(key).get(lang_code.replace('-', '_'))
# for some keys, such as territory_aliases, match may be a list
if isinstance(match, str):
@ -375,7 +378,7 @@ def _get_from_babel(lang_code, key):
return match
def _match_language(lang_code, lang_list=[], custom_aliases={}): # pylint: disable=W0102
def _match_language(lang_code: str, lang_list=[], custom_aliases={}) -> Optional[str]: # pylint: disable=W0102
"""auxiliary function to match lang_code in lang_list"""
# replace language code with a custom alias if necessary
if lang_code in custom_aliases:
@ -397,10 +400,12 @@ def _match_language(lang_code, lang_list=[], custom_aliases={}): # pylint: disa
return new_code
# try to get the any supported country for this language
return _get_lang_to_lc_dict(lang_list).get(lang_code, None)
return _get_lang_to_lc_dict(lang_list).get(lang_code)
def match_language(locale_code, lang_list=[], custom_aliases={}, fallback='en-US'): # pylint: disable=W0102
def match_language( # pylint: disable=W0102
locale_code, lang_list=[], custom_aliases={}, fallback: Optional[str] = 'en-US'
) -> Optional[str]:
"""get the language code from lang_list that best matches locale_code"""
# try to get language from given locale_code
language = _match_language(locale_code, lang_list, custom_aliases)
@ -438,29 +443,30 @@ def match_language(locale_code, lang_list=[], custom_aliases={}, fallback='en-US
return language or fallback
def load_module(filename, module_dir):
def load_module(filename: str, module_dir: str) -> types.ModuleType:
modname = splitext(filename)[0]
filepath = join(module_dir, filename)
modpath = join(module_dir, filename)
# and https://docs.python.org/3/library/importlib.html#importing-a-source-file-directly
spec = importlib.util.spec_from_file_location(modname, filepath)
spec = importlib.util.spec_from_file_location(modname, modpath)
if not spec:
raise ValueError(f"Error loading '{modpath}' module")
module = importlib.util.module_from_spec(spec)
if not spec.loader:
raise ValueError(f"Error loading '{modpath}' module")
spec.loader.exec_module(module)
return module
def to_string(obj):
def to_string(obj: Any) -> str:
"""Convert obj to its string representation."""
if isinstance(obj, str):
return obj
if isinstance(obj, Number):
return str(obj)
if hasattr(obj, '__str__'):
return obj.__str__()
if hasattr(obj, '__repr__'):
return obj.__repr__()
return repr(obj)
def ecma_unescape(s):
def ecma_unescape(string: str) -> str:
"""Python implementation of the unescape javascript function
https://www.ecma-international.org/ecma-262/6.0/#sec-unescape-string
@ -475,23 +481,23 @@ def ecma_unescape(s):
'ó'
"""
# "%u5409" becomes "吉"
s = ecma_unescape4_re.sub(lambda e: chr(int(e.group(1), 16)), s)
string = _ECMA_UNESCAPE4_RE.sub(lambda e: chr(int(e.group(1), 16)), string)
# "%20" becomes " ", "%F3" becomes "ó"
s = ecma_unescape2_re.sub(lambda e: chr(int(e.group(1), 16)), s)
return s
string = _ECMA_UNESCAPE2_RE.sub(lambda e: chr(int(e.group(1), 16)), string)
return string
def get_string_replaces_function(replaces):
def get_string_replaces_function(replaces: Dict[str, str]) -> Callable[[str], str]:
rep = {re.escape(k): v for k, v in replaces.items()}
pattern = re.compile("|".join(rep.keys()))
def f(text):
def func(text):
return pattern.sub(lambda m: rep[re.escape(m.group(0))], text)
return f
return func
def get_engine_from_settings(name):
def get_engine_from_settings(name: str) -> Dict:
"""Return engine configuration from settings.yml of a given engine name"""
if 'engines' not in settings:
@ -506,7 +512,7 @@ def get_engine_from_settings(name):
return {}
def get_xpath(xpath_spec):
def get_xpath(xpath_spec: XPathSpecType) -> XPath:
"""Return cached compiled XPath
There is no thread lock.
@ -523,13 +529,13 @@ def get_xpath(xpath_spec):
* SearxXPathSyntaxException: Raise when there is a syntax error in the XPath
"""
if isinstance(xpath_spec, str):
result = xpath_cache.get(xpath_spec, None)
result = _XPATH_CACHE.get(xpath_spec, None)
if result is None:
try:
result = XPath(xpath_spec)
except XPathSyntaxError as e:
raise SearxXPathSyntaxException(xpath_spec, str(e.msg)) from e
xpath_cache[xpath_spec] = result
_XPATH_CACHE[xpath_spec] = result
return result
if isinstance(xpath_spec, XPath):
@ -538,7 +544,7 @@ def get_xpath(xpath_spec):
raise TypeError('xpath_spec must be either a str or a lxml.etree.XPath')
def eval_xpath(element, xpath_spec):
def eval_xpath(element: ElementBase, xpath_spec: XPathSpecType):
"""Equivalent of element.xpath(xpath_str) but compile xpath_str once for all.
See https://lxml.de/xpathxslt.html#xpath-return-values
@ -562,7 +568,7 @@ def eval_xpath(element, xpath_spec):
raise SearxEngineXPathException(xpath_spec, arg) from e
def eval_xpath_list(element, xpath_spec, min_len=None):
def eval_xpath_list(element: ElementBase, xpath_spec: XPathSpecType, min_len: int = None):
"""Same as eval_xpath, check if the result is a list
Args:
@ -586,7 +592,7 @@ def eval_xpath_list(element, xpath_spec, min_len=None):
return result
def eval_xpath_getindex(elements, xpath_spec, index, default=NOTSET):
def eval_xpath_getindex(elements: ElementBase, xpath_spec: XPathSpecType, index: int, default=_NOTSET):
"""Call eval_xpath_list then get one element using the index parameter.
If the index does not exist, either aise an exception is default is not set,
other return the default value (can be None).
@ -606,9 +612,9 @@ def eval_xpath_getindex(elements, xpath_spec, index, default=NOTSET):
* result (bool, float, list, str): Results.
"""
result = eval_xpath_list(elements, xpath_spec)
if index >= -len(result) and index < len(result):
if -len(result) <= index < len(result):
return result[index]
if default == NOTSET:
if default == _NOTSET:
# raise an SearxEngineXPathException instead of IndexError
# to record xpath_spec
raise SearxEngineXPathException(xpath_spec, 'index ' + str(index) + ' not found')

View file

@ -128,7 +128,7 @@ class TestUtils(SearxTestCase):
class TestHTMLTextExtractor(SearxTestCase):
def setUp(self):
self.html_text_extractor = utils.HTMLTextExtractor()
self.html_text_extractor = utils._HTMLTextExtractor()
def test__init__(self):
self.assertEqual(self.html_text_extractor.result, [])
@ -149,7 +149,7 @@ class TestHTMLTextExtractor(SearxTestCase):
def test_invalid_html(self):
text = '<p><b>Lorem ipsum</i>dolor sit amet</p>'
with self.assertRaises(utils.HTMLTextExtractorException):
with self.assertRaises(utils._HTMLTextExtractorException):
self.html_text_extractor.feed(text)