[enh] replace requests by httpx

requirements.txt

@@ -8,5 +8,10 @@ lxml==4.6.3
 pygments==2.8.0
 python-dateutil==2.8.1
 pyyaml==5.4.1
-requests[socks]==2.25.1
+httpx[http2]==0.17.1
+Brotli==1.0.9
+uvloop==0.15.2; python_version >= '3.7'
+uvloop==0.14.0; python_version < '3.7'
+httpx-socks[asyncio]==0.3.1
 langdetect==1.0.8
+setproctitle==1.2.2

searx/autocomplete.py

@@ -20,7 +20,8 @@ from lxml import etree
 from json import loads
 from urllib.parse import urlencode
 
-from requests import RequestException
+from httpx import HTTPError
+
 
 from searx import settings
 from searx.poolrequests import get as http_get
@@ -136,5 +137,5 @@ def search_autocomplete(backend_name, query, lang):
 
     try:
         return backend(query, lang)
-    except (RequestException, SearxEngineResponseException):
+    except (HTTPError, SearxEngineResponseException):
         return []

searx/engines/dictzone.py

@@ -52,7 +52,7 @@ def response(resp):
                 to_results.append(to_result.text_content())
 
         results.append({
-            'url': urljoin(resp.url, '?%d' % k),
+            'url': urljoin(str(resp.url), '?%d' % k),
             'title': from_result.text_content(),
             'content': '; '.join(to_results)
         })

searx/engines/elasticsearch.py

@@ -4,7 +4,6 @@
 """
 
 from json import loads, dumps
-from requests.auth import HTTPBasicAuth
 from searx.exceptions import SearxEngineAPIException
 
 
@@ -32,7 +31,7 @@ def request(query, params):
         return params
 
     if username and password:
-        params['auth'] = HTTPBasicAuth(username, password)
+        params['auth'] = (username, password)
 
     params['url'] = search_url
     params['method'] = 'GET'

searx/engines/google.py

@@ -10,7 +10,7 @@ Definitions`_.
 
 # pylint: disable=invalid-name, missing-function-docstring
 
-from urllib.parse import urlencode, urlparse
+from urllib.parse import urlencode
 from lxml import html
 from searx import logger
 from searx.utils import match_language, extract_text, eval_xpath, eval_xpath_list, eval_xpath_getindex
@@ -186,8 +186,7 @@ def get_lang_info(params, lang_list, custom_aliases):
     return ret_val
 
 def detect_google_sorry(resp):
-    resp_url = urlparse(resp.url)
+    if resp.url.host == 'sorry.google.com' or resp.url.path.startswith('/sorry'):
-    if resp_url.netloc == 'sorry.google.com' or resp_url.path.startswith('/sorry'):
         raise SearxEngineCaptchaException()
 
 

searx/engines/seznam.py

@@ -3,7 +3,7 @@
  Seznam
 """
 
-from urllib.parse import urlencode, urlparse
+from urllib.parse import urlencode
 from lxml import html
 from searx.poolrequests import get
 from searx.exceptions import SearxEngineAccessDeniedException
@@ -46,8 +46,7 @@ def request(query, params):
 
 
 def response(resp):
-    resp_url = urlparse(resp.url)
+    if resp.url.path.startswith('/verify'):
-    if resp_url.path.startswith('/verify'):
         raise SearxEngineAccessDeniedException()
 
     results = []

searx/engines/spotify.py

@@ -5,7 +5,7 @@
 
 from json import loads
 from urllib.parse import urlencode
-import requests
+import searx.poolrequests as requests
 import base64
 
 # about

searx/engines/stackoverflow.py

@@ -3,7 +3,7 @@
  Stackoverflow (IT)
 """
 
-from urllib.parse import urlencode, urljoin, urlparse
+from urllib.parse import urlencode, urljoin
 from lxml import html
 from searx.utils import extract_text
 from searx.exceptions import SearxEngineCaptchaException
@@ -41,8 +41,7 @@ def request(query, params):
 
 # get response from search-request
 def response(resp):
-    resp_url = urlparse(resp.url)
+    if resp.url.path.startswith('/nocaptcha'):
-    if resp_url.path.startswith('/nocaptcha'):
         raise SearxEngineCaptchaException()
 
     results = []

searx/engines/yacy.py

@@ -7,7 +7,7 @@ from json import loads
 from dateutil import parser
 from urllib.parse import urlencode
 
-from requests.auth import HTTPDigestAuth
+from httpx import DigestAuth
 
 from searx.utils import html_to_text
 
@@ -56,7 +56,7 @@ def request(query, params):
                           search_type=search_type)
 
     if http_digest_auth_user and http_digest_auth_pass:
-        params['auth'] = HTTPDigestAuth(http_digest_auth_user, http_digest_auth_pass)
+        params['auth'] = DigestAuth(http_digest_auth_user, http_digest_auth_pass)
 
     # add language tag if specified
     if params['language'] != 'all':

searx/metrology/error_recorder.py

@@ -3,7 +3,7 @@ import inspect
 import logging
 from json import JSONDecodeError
 from urllib.parse import urlparse
-from requests.exceptions import RequestException
+from httpx import HTTPError, HTTPStatusError
 from searx.exceptions import (SearxXPathSyntaxException, SearxEngineXPathException, SearxEngineAPIException,
                               SearxEngineAccessDeniedException)
 from searx import logger
@@ -60,28 +60,28 @@ def get_trace(traces):
     return traces[-1]
 
 
-def get_hostname(exc: RequestException) -> typing.Optional[None]:
+def get_hostname(exc: HTTPError) -> typing.Optional[None]:
     url = exc.request.url
     if url is None and exc.response is not None:
         url = exc.response.url
     return urlparse(url).netloc
 
 
-def get_request_exception_messages(exc: RequestException)\
+def get_request_exception_messages(exc: HTTPError)\
         -> typing.Tuple[typing.Optional[str], typing.Optional[str], typing.Optional[str]]:
     url = None
     status_code = None
     reason = None
     hostname = None
-    if exc.request is not None:
+    if hasattr(exc, 'request') and exc.request is not None:
         url = exc.request.url
-    if url is None and exc.response is not None:
+    if url is None and hasattr(exc, 'response') and exc.respones is not None:
         url = exc.response.url
     if url is not None:
-        hostname = str(urlparse(url).netloc)
+        hostname = url.host
-    if exc.response is not None:
+    if isinstance(exc, HTTPStatusError):
         status_code = str(exc.response.status_code)
-        reason = exc.response.reason
+        reason = exc.response.reason_phrase
     return (status_code, reason, hostname)
 
 
@@ -92,7 +92,7 @@ def get_messages(exc, filename) -> typing.Tuple:
         return (str(exc), )
     if isinstance(exc, ValueError) and 'lxml' in filename:
         return (str(exc), )
-    if isinstance(exc, RequestException):
+    if isinstance(exc, HTTPError):
         return get_request_exception_messages(exc)
     if isinstance(exc, SearxXPathSyntaxException):
         return (exc.xpath_str, exc.message)

searx/poolrequests.py

@@ -1,14 +1,54 @@
+import atexit
 import sys
+import threading
+import asyncio
+import logging
+import concurrent.futures
 from time import time
 from itertools import cycle
-from threading import local
 
-import requests
+import httpcore
+import httpx
+import h2.exceptions
+from httpx_socks import AsyncProxyTransport
+from python_socks import parse_proxy_url
+import python_socks._errors
 
 from searx import settings
 from searx import logger
 from searx.raise_for_httperror import raise_for_httperror
 
+# Optional uvloop (support Python 3.6)
+try:
+    import uvloop
+except ImportError:
+    pass
+else:
+    uvloop.install()
+
+# queue.SimpleQueue: Support Python 3.6
+try:
+    from queue import SimpleQueue
+except ImportError:
+    from queue import Empty
+    from collections import deque
+
+    class SimpleQueue:
+        """Minimal backport of queue.SimpleQueue"""
+
+        def __init__(self):
+            self._queue = deque()
+            self._count = threading.Semaphore(0)
+
+        def put(self, item):
+            self._queue.append(item)
+            self._count.release()
+
+        def get(self, timeout=None):
+            if not self._count.acquire(True, timeout):
+                raise Empty
+            return self._queue.popleft()
+
 
 logger = logger.getChild('poolrequests')
 
@@ -31,99 +71,63 @@ if not getattr(ssl, "HAS_SNI", False):
         sys.exit(1)
 
 
-class HTTPAdapterWithConnParams(requests.adapters.HTTPAdapter):
+LOOP = None
-
+CLIENTS = dict()
-    def __init__(self, pool_connections=requests.adapters.DEFAULT_POOLSIZE,
+THREADLOCAL = threading.local()
-                 pool_maxsize=requests.adapters.DEFAULT_POOLSIZE,
+LIMITS = httpx.Limits(
-                 max_retries=requests.adapters.DEFAULT_RETRIES,
+    # Magic number kept from previous code
-                 pool_block=requests.adapters.DEFAULT_POOLBLOCK,
+    max_connections=settings['outgoing'].get('pool_connections', 100),
-                 **conn_params):
+    # Picked from constructor
-        if max_retries == requests.adapters.DEFAULT_RETRIES:
+    max_keepalive_connections=settings['outgoing'].get('pool_maxsize', 10),
-            self.max_retries = requests.adapters.Retry(0, read=False)
+    #
-        else:
+    keepalive_expiry=settings['outgoing'].get('keepalive_expiry', 5.0)
-            self.max_retries = requests.adapters.Retry.from_int(max_retries)
+)
-        self.config = {}
+# default parameters for AsyncHTTPTransport
-        self.proxy_manager = {}
+# see https://github.com/encode/httpx/blob/e05a5372eb6172287458b37447c30f650047e1b8/httpx/_transports/default.py#L108-L121  # noqa
-
+TRANSPORT_KWARGS = {
-        super().__init__()
+    'http2': settings['outgoing'].get('http2', False),
-
+    'retries': 0,
-        self._pool_connections = pool_connections
+    'trust_env': False,
-        self._pool_maxsize = pool_maxsize
+    'backend': 'asyncio'
-        self._pool_block = pool_block
+}
-        self._conn_params = conn_params
+# requests compatibility when reading proxy settings from settings.yml
-
+PROXY_PATTERN_MAPPING = {
-        self.init_poolmanager(pool_connections, pool_maxsize, block=pool_block, **conn_params)
+    'http': 'https://',
-
+    'https:': 'https://'
-    def __setstate__(self, state):
+}
-        # Can't handle by adding 'proxy_manager' to self.__attrs__ because
+# default maximum redirect
-        # because self.poolmanager uses a lambda function, which isn't pickleable.
+# from https://github.com/psf/requests/blob/8c211a96cdbe9fe320d63d9e1ae15c5c07e179f8/requests/models.py#L55
-        self.proxy_manager = {}
+DEFAULT_REDIRECT_LIMIT = 30
-        self.config = {}
-
-        for attr, value in state.items():
-            setattr(self, attr, value)
-
-        self.init_poolmanager(self._pool_connections, self._pool_maxsize,
-                              block=self._pool_block, **self._conn_params)
 
 
-threadLocal = local()
-connect = settings['outgoing'].get('pool_connections', 100)  # Magic number kept from previous code
-maxsize = settings['outgoing'].get('pool_maxsize', requests.adapters.DEFAULT_POOLSIZE)  # Picked from constructor
 if settings['outgoing'].get('source_ips'):
-    http_adapters = cycle(HTTPAdapterWithConnParams(pool_connections=connect, pool_maxsize=maxsize,
+    LOCAL_ADDRESS_CYCLE = cycle(settings['outgoing'].get('source_ips'))
-                                                    source_address=(source_ip, 0))
-                          for source_ip in settings['outgoing']['source_ips'])
-    https_adapters = cycle(HTTPAdapterWithConnParams(pool_connections=connect, pool_maxsize=maxsize,
-                                                     source_address=(source_ip, 0))
-                           for source_ip in settings['outgoing']['source_ips'])
 else:
-    http_adapters = cycle((HTTPAdapterWithConnParams(pool_connections=connect, pool_maxsize=maxsize), ))
+    LOCAL_ADDRESS_CYCLE = cycle((None, ))
-    https_adapters = cycle((HTTPAdapterWithConnParams(pool_connections=connect, pool_maxsize=maxsize), ))
-
-
-class SessionSinglePool(requests.Session):
-
-    def __init__(self):
-        super().__init__()
-
-        # reuse the same adapters
-        self.adapters.clear()
-
-        https_adapter = threadLocal.__dict__.setdefault('https_adapter', next(https_adapters))
-        self.mount('https://', https_adapter)
-        if get_enable_http_protocol():
-            http_adapter = threadLocal.__dict__.setdefault('http_adapter', next(http_adapters))
-            self.mount('http://', http_adapter)
-
-    def close(self):
-        """Call super, but clear adapters since there are managed globaly"""
-        self.adapters.clear()
-        super().close()
 
 
 def set_timeout_for_thread(timeout, start_time=None):
-    threadLocal.timeout = timeout
+    THREADLOCAL.timeout = timeout
-    threadLocal.start_time = start_time
+    THREADLOCAL.start_time = start_time
 
 
 def set_enable_http_protocol(enable_http):
-    threadLocal.enable_http = enable_http
+    THREADLOCAL.enable_http = enable_http
 
 
 def get_enable_http_protocol():
     try:
-        return threadLocal.enable_http
+        return THREADLOCAL.enable_http
     except AttributeError:
         return False
 
 
 def reset_time_for_thread():
-    threadLocal.total_time = 0
+    THREADLOCAL.total_time = 0
 
 
 def get_time_for_thread():
-    return threadLocal.total_time
+    return THREADLOCAL.total_time
 
 
 def get_proxy_cycles(proxy_settings):
@@ -152,22 +156,197 @@ def get_global_proxies():
     return get_proxies(GLOBAL_PROXY_CYCLES)
 
 
+async def close_connections_for_url(connection_pool: httpcore.AsyncConnectionPool, url: httpcore._utils.URL):
+    origin = httpcore._utils.url_to_origin(url)
+    logger.debug('Drop connections for %r', origin)
+    connections_to_close = connection_pool._connections_for_origin(origin)
+    for connection in connections_to_close:
+        await connection_pool._remove_from_pool(connection)
+        try:
+            await connection.aclose()
+        except httpcore.NetworkError as e:
+            logger.warning('Error closing an existing connection', exc_info=e)
+
+
+class AsyncHTTPTransportNoHttp(httpcore.AsyncHTTPTransport):
+    """Block HTTP request"""
+
+    async def arequest(self, method, url, headers=None, stream=None, ext=None):
+        raise httpcore.UnsupportedProtocol("HTTP protocol is disabled")
+
+
+class AsyncProxyTransportFixed(AsyncProxyTransport):
+    """Fix httpx_socks.AsyncProxyTransport
+
+    Map python_socks exceptions to httpcore.ProxyError
+
+    Map socket.gaierror to httpcore.ConnectError
+
+    Note: keepalive_expiry is ignored, AsyncProxyTransport should call:
+    * self._keepalive_sweep()
+    * self._response_closed(self, connection)
+
+    Note: AsyncProxyTransport inherit from AsyncConnectionPool
+
+    Note: the API is going to change on httpx 0.18.0
+    see https://github.com/encode/httpx/pull/1522
+    """
+
+    async def arequest(self, method, url, headers=None, stream=None, ext=None):
+        retry = 2
+        while retry > 0:
+            retry -= 1
+            try:
+                return await super().arequest(method, url, headers, stream, ext)
+            except (python_socks._errors.ProxyConnectionError,
+                    python_socks._errors.ProxyTimeoutError,
+                    python_socks._errors.ProxyError) as e:
+                raise httpcore.ProxyError(e)
+            except OSError as e:
+                # socket.gaierror when DNS resolution fails
+                raise httpcore.NetworkError(e)
+            except httpcore.RemoteProtocolError as e:
+                # in case of httpcore.RemoteProtocolError: Server disconnected
+                await close_connections_for_url(self, url)
+                logger.warning('httpcore.RemoteProtocolError: retry', exc_info=e)
+                # retry
+            except (httpcore.NetworkError, httpcore.ProtocolError) as e:
+                # httpcore.WriteError on HTTP/2 connection leaves a new opened stream
+                # then each new request creates a new stream and raise the same WriteError
+                await close_connections_for_url(self, url)
+                raise e
+
+
+class AsyncHTTPTransportFixed(httpx.AsyncHTTPTransport):
+    """Fix httpx.AsyncHTTPTransport"""
+
+    async def arequest(self, method, url, headers=None, stream=None, ext=None):
+        retry = 2
+        while retry > 0:
+            retry -= 1
+            try:
+                return await super().arequest(method, url, headers, stream, ext)
+            except OSError as e:
+                # socket.gaierror when DNS resolution fails
+                raise httpcore.ConnectError(e)
+            except httpcore.CloseError as e:
+                # httpcore.CloseError: [Errno 104] Connection reset by peer
+                # raised by _keepalive_sweep()
+                #   from https://github.com/encode/httpcore/blob/4b662b5c42378a61e54d673b4c949420102379f5/httpcore/_backends/asyncio.py#L198  # noqa
+                await close_connections_for_url(self._pool, url)
+                logger.warning('httpcore.CloseError: retry', exc_info=e)
+                # retry
+            except httpcore.RemoteProtocolError as e:
+                # in case of httpcore.RemoteProtocolError: Server disconnected
+                await close_connections_for_url(self._pool, url)
+                logger.warning('httpcore.RemoteProtocolError: retry', exc_info=e)
+                # retry
+            except (httpcore.ProtocolError, httpcore.NetworkError) as e:
+                await close_connections_for_url(self._pool, url)
+                raise e
+
+
+def get_transport_for_socks_proxy(verify, local_address, proxy_url):
+    global LOOP, LIMITS, TRANSPORT_KWARGS
+    # support socks5h (requests compatibility):
+    # https://requests.readthedocs.io/en/master/user/advanced/#socks
+    # socks5://   hostname is resolved on client side
+    # socks5h://  hostname is resolved on proxy side
+    rdns = False
+    socks5h = 'socks5h://'
+    if proxy_url.startswith(socks5h):
+        proxy_url = 'socks5://' + proxy_url[len(socks5h):]
+        rdns = True
+
+    proxy_type, proxy_host, proxy_port, proxy_username, proxy_password = parse_proxy_url(proxy_url)
+
+    return AsyncProxyTransportFixed(proxy_type=proxy_type, proxy_host=proxy_host, proxy_port=proxy_port,
+                                    username=proxy_username, password=proxy_password,
+                                    rdns=rdns,
+                                    loop=LOOP,
+                                    verify=verify,
+                                    local_address=local_address,
+                                    max_connections=LIMITS.max_connections,
+                                    max_keepalive_connections=LIMITS.max_keepalive_connections,
+                                    keepalive_expiry=LIMITS.keepalive_expiry,
+                                    **TRANSPORT_KWARGS)
+
+
+def get_transport(verify, local_address, proxy_url):
+    global LIMITS
+    return AsyncHTTPTransportFixed(verify=verify,
+                                   local_address=local_address,
+                                   limits=LIMITS,
+                                   proxy=httpx._config.Proxy(proxy_url) if proxy_url else None,
+                                   **TRANSPORT_KWARGS)
+
+
+def iter_proxies(proxies):
+    # https://www.python-httpx.org/compatibility/#proxy-keys
+    if isinstance(proxies, str):
+        yield 'all://', proxies
+    elif isinstance(proxies, dict):
+        for pattern, proxy_url in proxies.items():
+            pattern = PROXY_PATTERN_MAPPING.get(pattern, pattern)
+            yield pattern, proxy_url
+
+
+def new_client(verify, local_address, proxies, max_redirects, enable_http):
+    # See https://www.python-httpx.org/advanced/#routing
+    mounts = {}
+    for pattern, proxy_url in iter_proxies(proxies):
+        if not enable_http and (pattern == 'http' or pattern.startswith('http://')):
+            continue
+        if proxy_url.startswith('socks4://') \
+           or proxy_url.startswith('socks5://') \
+           or proxy_url.startswith('socks5h://'):
+            mounts[pattern] = get_transport_for_socks_proxy(verify, local_address, proxy_url)
+        else:
+            mounts[pattern] = get_transport(verify, local_address, proxy_url)
+
+    if not enable_http:
+        mounts['http://'] = AsyncHTTPTransportNoHttp()
+
+    transport = get_transport(verify, local_address, None)
+    return httpx.AsyncClient(transport=transport, mounts=mounts, max_redirects=max_redirects)
+
+
+def get_client(verify, local_address, proxies, max_redirects, allow_http):
+    global CLIENTS
+    key = (verify, local_address, repr(proxies), max_redirects, allow_http)
+    if key not in CLIENTS:
+        CLIENTS[key] = new_client(verify, local_address, proxies, max_redirects, allow_http)
+    return CLIENTS[key]
+
+
+async def send_request(method, url, enable_http, kwargs):
+    if isinstance(url, bytes):
+        url = url.decode()
+
+    verify = kwargs.pop('verify', True)
+    local_address = next(LOCAL_ADDRESS_CYCLE)
+    proxies = kwargs.pop('proxies', None) or get_global_proxies()
+    max_redirects = kwargs.pop('max_redirects', DEFAULT_REDIRECT_LIMIT)
+
+    client = get_client(verify, local_address, proxies, max_redirects, enable_http)
+    response = await client.request(method.upper(), url, **kwargs)
+
+    # requests compatibility
+    # see also https://www.python-httpx.org/compatibility/#checking-for-4xx5xx-responses
+    response.ok = not response.is_error
+
+    return response
+
+
 def request(method, url, **kwargs):
     """same as requests/requests/api.py request(...)"""
     time_before_request = time()
 
-    # session start
-    session = SessionSinglePool()
-
-    # proxies
-    if not kwargs.get('proxies'):
-        kwargs['proxies'] = get_global_proxies()
-
     # timeout
     if 'timeout' in kwargs:
         timeout = kwargs['timeout']
     else:
-        timeout = getattr(threadLocal, 'timeout', None)
+        timeout = getattr(THREADLOCAL, 'timeout', None)
         if timeout is not None:
             kwargs['timeout'] = timeout
 
@@ -178,24 +357,23 @@ def request(method, url, **kwargs):
         del kwargs['raise_for_httperror']
 
     # do request
-    response = session.request(method=method, url=url, **kwargs)
+    future = asyncio.run_coroutine_threadsafe(send_request(method, url, get_enable_http_protocol(), kwargs), LOOP)
+    try:
+        if timeout:
+            timeout += 0.2  # overhead
+            start_time = getattr(THREADLOCAL, 'start_time', time_before_request)
+            if start_time:
+                timeout -= time() - start_time
 
+        response = future.result(timeout or 120)
+    except concurrent.futures.TimeoutError as e:
+        raise httpx.TimeoutException('Timeout', request=None) from e
+
+    # update total_time.
+    # See get_time_for_thread() and reset_time_for_thread()
+    if hasattr(THREADLOCAL, 'total_time'):
         time_after_request = time()
-
+        THREADLOCAL.total_time += time_after_request - time_before_request
-    # is there a timeout for this engine ?
-    if timeout is not None:
-        timeout_overhead = 0.2  # seconds
-        # start_time = when the user request started
-        start_time = getattr(threadLocal, 'start_time', time_before_request)
-        search_duration = time_after_request - start_time
-        if search_duration > timeout + timeout_overhead:
-            raise requests.exceptions.Timeout(response=response)
-
-    # session end
-    session.close()
-
-    if hasattr(threadLocal, 'total_time'):
-        threadLocal.total_time += time_after_request - time_before_request
 
     # raise an exception
     if check_for_httperror:
@@ -204,6 +382,49 @@ def request(method, url, **kwargs):
     return response
 
 
+async def stream_chunk_to_queue(method, url, q, **kwargs):
+    verify = kwargs.pop('verify', True)
+    local_address = next(LOCAL_ADDRESS_CYCLE)
+    proxies = kwargs.pop('proxies', None) or get_global_proxies()
+    # "30" from requests:
+    # https://github.com/psf/requests/blob/8c211a96cdbe9fe320d63d9e1ae15c5c07e179f8/requests/models.py#L55
+    max_redirects = kwargs.pop('max_redirects', 30)
+    client = get_client(verify, local_address, proxies, max_redirects, True)
+    try:
+        async with client.stream(method, url, **kwargs) as response:
+            q.put(response)
+            async for chunk in response.aiter_bytes(65536):
+                if len(chunk) > 0:
+                    q.put(chunk)
+    except (httpx.HTTPError, OSError, h2.exceptions.ProtocolError) as e:
+        q.put(e)
+    finally:
+        q.put(None)
+
+
+def stream(method, url, **kwargs):
+    """Replace httpx.stream.
+
+    Usage:
+    stream = poolrequests.stream(...)
+    response = next(stream)
+    for chunk in stream:
+        ...
+
+    httpx.Client.stream requires to write the httpx.HTTPTransport version of the
+    the httpx.AsyncHTTPTransport declared above.
+    """
+    q = SimpleQueue()
+    future = asyncio.run_coroutine_threadsafe(stream_chunk_to_queue(method, url, q, **kwargs), LOOP)
+    chunk_or_exception = q.get(timeout=60)
+    while chunk_or_exception is not None:
+        if isinstance(chunk_or_exception, Exception):
+            raise chunk_or_exception
+        yield chunk_or_exception
+        chunk_or_exception = q.get(timeout=60)
+    return future.result()
+
+
 def get(url, **kwargs):
     kwargs.setdefault('allow_redirects', True)
     return request('get', url, **kwargs)
@@ -233,3 +454,97 @@ def patch(url, data=None, **kwargs):
 
 def delete(url, **kwargs):
     return request('delete', url, **kwargs)
+
+
+def init():
+    # log
+    for logger_name in ('hpack.hpack', 'hpack.table'):
+        logging.getLogger(logger_name).setLevel(logging.WARNING)
+
+    # loop
+    def loop_thread():
+        global LOOP
+        LOOP = asyncio.new_event_loop()
+        LOOP.run_forever()
+
+    th = threading.Thread(
+        target=loop_thread,
+        name='asyncio_loop',
+        daemon=True,
+    )
+    th.start()
+
+
+@atexit.register
+def done():
+    """Close all HTTP client
+
+    Avoid a warning at exit
+    see https://github.com/encode/httpx/blob/1a6e254f72d9fd5694a1c10a28927e193ab4f76b/httpx/_client.py#L1785
+    """
+    global LOOP
+
+    async def close_client(client):
+        try:
+            await client.aclose()
+        except httpx.HTTPError:
+            pass
+
+    async def close_clients():
+        await asyncio.gather(*[close_client(client) for client in CLIENTS.values()], return_exceptions=False)
+    future = asyncio.run_coroutine_threadsafe(close_clients(), LOOP)
+    # wait 3 seconds to close the HTTP clients
+    future.result(3)
+
+
+init()
+
+
+# ## TEMPORARY DEBUG ##
+
+
+def debug_connection(connection):
+    now = LOOP.time()
+    expired = (connection.state == httpcore._async.base.ConnectionState.IDLE
+               and connection.expires_at is not None
+               and now >= connection.expires_at)
+    return connection.info()\
+        + (', connect_failed' if connection.connect_failed else '')\
+        + (', expired' if expired else '')
+
+
+def debug_origin(origin):
+    return origin[0].decode() + '://' + origin[1].decode() + ':' + str(origin[2])
+
+
+def debug_transport(transport):
+    result = {
+        '__class__': str(transport.__class__.__name__)
+    }
+    if isinstance(transport, (httpx.AsyncHTTPTransport, AsyncHTTPTransportFixed)):
+        pool = transport._pool
+        result['__pool_class__'] = str(pool.__class__.__name__)
+        if isinstance(pool, httpcore.AsyncConnectionPool):
+            for origin, connections in pool._connections.items():
+                result[debug_origin(origin)] = [debug_connection(connection) for connection in connections]
+            return result
+    elif isinstance(transport, AsyncProxyTransportFixed):
+        for origin, connections in transport._connections.items():
+            result[debug_origin(origin)] = [debug_connection(connection) for connection in connections]
+        return result
+    return result
+
+
+def debug_asyncclient(client, key=None):
+    result = {}
+    if key:
+        result['__key__'] = [k if isinstance(k, (str, int, float, bool, type(None))) else repr(k) for k in key]
+    result['__default__'] = debug_transport(client._transport)
+    for urlpattern, transport in client._mounts.items():
+        result[urlpattern.pattern] = debug_transport(transport)
+    return result
+
+
+def debug_asyncclients():
+    global CLIENTS
+    return [debug_asyncclient(client, key) for key, client in CLIENTS.items()]

searx/search/checker/impl.py

@@ -11,7 +11,7 @@ from urllib.parse import urlparse
 import re
 from langdetect import detect_langs
 from langdetect.lang_detect_exception import LangDetectException
-import requests.exceptions
+import httpx
 
 from searx import poolrequests, logger
 from searx.results import ResultContainer
@@ -90,10 +90,10 @@ def _is_url_image(image_url):
             if r.headers["content-type"].startswith('image/'):
                 return True
             return False
-        except requests.exceptions.Timeout:
+        except httpx.TimeoutException:
             logger.error('Timeout for %s: %i', image_url, int(time() - a))
             retry -= 1
-        except requests.exceptions.RequestException:
+        except httpx.HTTPError:
             logger.exception('Exception for %s', image_url)
             return False
 

searx/search/processors/online.py

@@ -1,10 +1,10 @@
 # SPDX-License-Identifier: AGPL-3.0-or-later
 
-from urllib.parse import urlparse
 from time import time
 import threading
+import asyncio
 
-import requests.exceptions
+import httpx
 
 import searx.poolrequests as poolrequests
 from searx.engines import settings
@@ -99,8 +99,8 @@ class OnlineProcessor(EngineProcessor):
             # unexpected redirect : record an error
             # but the engine might still return valid results.
             status_code = str(response.status_code or '')
-            reason = response.reason or ''
+            reason = response.reason_phrase or ''
-            hostname = str(urlparse(response.url or '').netloc)
+            hostname = response.url.host
             record_error(self.engine_name,
                          '{} redirects, maximum: {}'.format(len(response.history), soft_max_redirects),
                          (status_code, reason, hostname))
@@ -135,7 +135,7 @@ class OnlineProcessor(EngineProcessor):
         poolrequests.set_enable_http_protocol(self.engine.enable_http)
 
         # suppose everything will be alright
-        requests_exception = False
+        http_exception = False
         suspended_time = None
 
         try:
@@ -169,20 +169,20 @@ class OnlineProcessor(EngineProcessor):
             with threading.RLock():
                 self.engine.stats['errors'] += 1
 
-            if (issubclass(e.__class__, requests.exceptions.Timeout)):
+            if (issubclass(e.__class__, (httpx.TimeoutException, asyncio.TimeoutError))):
                 result_container.add_unresponsive_engine(self.engine_name, 'HTTP timeout')
                 # requests timeout (connect or read)
                 logger.error("engine {0} : HTTP requests timeout"
                              "(search duration : {1} s, timeout: {2} s) : {3}"
                              .format(self.engine_name, engine_time, timeout_limit, e.__class__.__name__))
-                requests_exception = True
+                http_exception = True
-            elif (issubclass(e.__class__, requests.exceptions.RequestException)):
+            elif (issubclass(e.__class__, (httpx.HTTPError, httpx.StreamError))):
                 result_container.add_unresponsive_engine(self.engine_name, 'HTTP error')
                 # other requests exception
                 logger.exception("engine {0} : requests exception"
                                  "(search duration : {1} s, timeout: {2} s) : {3}"
                                  .format(self.engine_name, engine_time, timeout_limit, e))
-                requests_exception = True
+                http_exception = True
             elif (issubclass(e.__class__, SearxEngineCaptchaException)):
                 result_container.add_unresponsive_engine(self.engine_name, 'CAPTCHA required')
                 logger.exception('engine {0} : CAPTCHA'.format(self.engine_name))
@@ -206,7 +206,7 @@ class OnlineProcessor(EngineProcessor):
         # suspend the engine if there is an HTTP error
         # or suspended_time is defined
         with threading.RLock():
-            if requests_exception or suspended_time:
+            if http_exception or suspended_time:
                 # update continuous_errors / suspend_end_time
                 self.engine.continuous_errors += 1
                 if suspended_time is None:

searx/settings.yml

@@ -67,11 +67,13 @@ ui:
 #    key : !!binary "your_morty_proxy_key"
 
 outgoing: # communication with search engines
-    request_timeout : 2.0 # default timeout in seconds, can be override by engine
+    request_timeout : 3.0 # default timeout in seconds, can be override by engine
     # max_request_timeout: 10.0 # the maximum timeout in seconds
     useragent_suffix : "" # suffix of searx_useragent, could contain informations like an email address to the administrator
-    pool_connections : 100 # Number of different hosts
+    pool_connections : 100 # The maximum number of concurrent connections that may be established.
-    pool_maxsize : 10 # Number of simultaneous requests by host
+    pool_maxsize : 20 # Allow the connection pool to maintain keep-alive connections below this point.
+    keepalive_expiry: 30.0 # Number of seconds to keep a connection in the pool
+    http2: True  # Enable HTTP/2 (experimental)
 # uncomment below section if you want to use a proxy
 # see https://2.python-requests.org/en/latest/user/advanced/#proxies
 # SOCKS proxies are also supported: see https://2.python-requests.org/en/latest/user/advanced/#socks

searx/utils.py

@@ -45,7 +45,7 @@ def searx_useragent():
     """Return the searx User Agent"""
     return 'searx/{searx_version} {suffix}'.format(
            searx_version=VERSION_STRING,
-           suffix=settings['outgoing'].get('useragent_suffix', ''))
+           suffix=settings['outgoing'].get('useragent_suffix', '')).strip()
 
 
 def gen_useragent(os=None):

searx/webapp.py

@@ -26,12 +26,26 @@ if __name__ == '__main__':
     from os.path import realpath, dirname
     sys.path.append(realpath(dirname(realpath(__file__)) + '/../'))
 
+# set Unix thread name
+try:
+    import setproctitle
+except ImportError:
+    pass
+else:
+    import threading
+    old_thread_init = threading.Thread.__init__
+
+    def new_thread_init(self, *args, **kwargs):
+        old_thread_init(self, *args, **kwargs)
+        setproctitle.setthreadtitle(self._name)
+    threading.Thread.__init__ = new_thread_init
+
 import hashlib
 import hmac
 import json
 import os
 
-import requests
+import httpx
 
 from searx import logger
 logger = logger.getChild('webapp')
@@ -79,7 +93,7 @@ from searx.plugins import plugins
 from searx.plugins.oa_doi_rewrite import get_doi_resolver
 from searx.preferences import Preferences, ValidationException, LANGUAGE_CODES
 from searx.answerers import answerers
-from searx.poolrequests import get_global_proxies
+from searx import poolrequests
 from searx.answerers import ask
 from searx.metrology.error_recorder import errors_per_engines
 
@@ -890,24 +904,33 @@ def _is_selected_language_supported(engine, preferences):
 
 @app.route('/image_proxy', methods=['GET'])
 def image_proxy():
-    url = request.args.get('url').encode()
+    url = request.args.get('url')
 
     if not url:
         return '', 400
 
-    h = new_hmac(settings['server']['secret_key'], url)
+    h = new_hmac(settings['server']['secret_key'], url.encode())
 
     if h != request.args.get('h'):
         return '', 400
 
+    maximum_size = 5 * 1024 * 1024
+
+    try:
         headers = dict_subset(request.headers, {'If-Modified-Since', 'If-None-Match'})
         headers['User-Agent'] = gen_useragent()
-
+        stream = poolrequests.stream(
-    resp = requests.get(url,
+            method='GET',
-                        stream=True,
+            url=url,
-                        timeout=settings['outgoing']['request_timeout'],
             headers=headers,
-                        proxies=get_global_proxies())
+            timeout=settings['outgoing']['request_timeout'],
+            allow_redirects=True,
+            max_redirects=20)
+
+        resp = next(stream)
+        content_length = resp.headers.get('Content-Length')
+        if content_length and content_length.isdigit() and int(content_length) > maximum_size:
+            return 'Max size', 400
 
         if resp.status_code == 304:
             return '', resp.status_code
@@ -922,18 +945,21 @@ def image_proxy():
             logger.debug('image-proxy: wrong content-type: {0}'.format(resp.headers.get('content-type')))
             return '', 400
 
-    img = b''
-    chunk_counter = 0
-
-    for chunk in resp.iter_content(1024 * 1024):
-        chunk_counter += 1
-        if chunk_counter > 5:
-            return '', 502  # Bad gateway - file is too big (>5M)
-        img += chunk
-
         headers = dict_subset(resp.headers, {'Content-Length', 'Length', 'Date', 'Last-Modified', 'Expires', 'Etag'})
 
-    return Response(img, mimetype=resp.headers['content-type'], headers=headers)
+        total_length = 0
+
+        def forward_chunk():
+            nonlocal total_length
+            for chunk in stream:
+                total_length += len(chunk)
+                if total_length > maximum_size:
+                    break
+                yield chunk
+
+        return Response(forward_chunk(), mimetype=resp.headers['Content-Type'], headers=headers)
+    except httpx.HTTPError:
+        return '', 400
 
 
 @app.route('/stats', methods=['GET'])
@@ -1083,6 +1109,11 @@ def config():
     })
 
 
+@app.route('/config/http')
+def config_http():
+    return jsonify(poolrequests.debug_asyncclients())
+
+
 @app.errorhandler(404)
 def page_not_found(e):
     return render('404.html'), 404

searx_extra/update/update_external_bangs.py

@@ -17,7 +17,7 @@ import json
 import re
 from os.path import join
 
-import requests
+import httpx
 
 from searx import searx_dir  # pylint: disable=E0401 C0413
 
@@ -30,7 +30,7 @@ HTTP_COLON = 'http:'
 
 
 def get_bang_url():
-    response = requests.get(URL_BV1)
+    response = httpx.get(URL_BV1)
     response.raise_for_status()
 
     r = RE_BANG_VERSION.findall(response.text)
@@ -38,7 +38,7 @@ def get_bang_url():
 
 
 def fetch_ddg_bangs(url):
-    response = requests.get(url)
+    response = httpx.get(url)
     response.raise_for_status()
     return json.loads(response.content.decode())
 

tests/unit/test_poolrequests.py

@@ -1,9 +1,5 @@
-from unittest.mock import patch
-from requests.models import Response
-
 from searx.testing import SearxTestCase
 
-import searx.poolrequests
 from searx.poolrequests import get_proxy_cycles, get_proxies
 
 
@@ -64,26 +60,3 @@ class TestProxy(SearxTestCase):
             'http': 'http://localhost:9092',
             'https': 'http://localhost:9093'
         })
-
-    @patch('searx.poolrequests.get_global_proxies')
-    def test_request(self, mock_get_global_proxies):
-        method = 'GET'
-        url = 'http://localhost'
-        custom_proxies = {
-            'https': 'http://localhost:1080'
-        }
-        global_proxies = {
-            'http': 'http://localhost:9092',
-            'https': 'http://localhost:9093'
-        }
-        mock_get_global_proxies.return_value = global_proxies
-
-        # check the global proxies usage
-        with patch.object(searx.poolrequests.SessionSinglePool, 'request', return_value=Response()) as mock_method:
-            searx.poolrequests.request(method, url)
-        mock_method.assert_called_once_with(method=method, url=url, proxies=global_proxies)
-
-        # check if the proxies parameter overrides the global proxies
-        with patch.object(searx.poolrequests.SessionSinglePool, 'request', return_value=Response()) as mock_method:
-            searx.poolrequests.request(method, url, proxies=custom_proxies)
-        mock_method.assert_called_once_with(method=method, url=url, proxies=custom_proxies)