Merge pull request #28 from searxng/mod-processors-error-message

[mod] processors: show identical error messages on /search and /stats

searx/search/processors/__init__.py

@@ -1,37 +1,49 @@
 # SPDX-License-Identifier: AGPL-3.0-or-later
+# lint: pylint
+
+"""Implement request processores used by engine-types.
+
+"""
+
+__all__ = [
+    'EngineProcessor',
+    'OfflineProcessor',
+    'OnlineProcessor',
+    'OnlineDictionaryProcessor',
+    'OnlineCurrencyProcessor',
+    'processors',
+]
+
+from searx import logger
+import searx.engines as engines
 
 from .online import OnlineProcessor
 from .offline import OfflineProcessor
 from .online_dictionary import OnlineDictionaryProcessor
 from .online_currency import OnlineCurrencyProcessor
 from .abstract import EngineProcessor
-from searx import logger
-import searx.engines as engines
-
 
-__all__ = ['EngineProcessor', 'OfflineProcessor', 'OnlineProcessor',
-           'OnlineDictionaryProcessor', 'OnlineCurrencyProcessor', 'processors']
 logger = logger.getChild('search.processors')
 processors = {}
-
+"""Cache request processores, stored by *engine-name* (:py:func:`initialize`)"""
 
 def get_processor_class(engine_type):
+    """Return processor class according to the ``engine_type``"""
     for c in [OnlineProcessor, OfflineProcessor, OnlineDictionaryProcessor, OnlineCurrencyProcessor]:
         if c.engine_type == engine_type:
             return c
     return None
 
-
 def get_processor(engine, engine_name):
+    """Return processor instance that fits to ``engine.engine.type``)"""
     engine_type = getattr(engine, 'engine_type', 'online')
     processor_class = get_processor_class(engine_type)
     if processor_class:
         return processor_class(engine, engine_name)
-    else:
-        return None
-
+    return None
 
 def initialize(engine_list):
+    """Initialize all engines and store a processor for each engine in :py:obj:`processors`."""
     engines.initialize_engines(engine_list)
     for engine_name, engine in engines.engines.items():
         processor = get_processor(engine, engine_name)

searx/search/processors/abstract.py

@@ -1,4 +1,9 @@
 # SPDX-License-Identifier: AGPL-3.0-or-later
+# lint: pylint
+
+"""Abstract base classes for engine request processores.
+
+"""
 
 import threading
 from abc import abstractmethod, ABC
@@ -10,12 +15,13 @@ from searx.network import get_time_for_thread, get_network
 from searx.metrics import histogram_observe, counter_inc, count_exception, count_error
 from searx.exceptions import SearxEngineAccessDeniedException
 
-
 logger = logger.getChild('searx.search.processor')
 SUSPENDED_STATUS = {}
 
+# pylint: disable=missing-function-docstring
 
 class SuspendedStatus:
+    """Class to handle suspend state."""
 
     __slots__ = 'suspend_end_time', 'suspend_reason', 'continuous_errors', 'lock'
 
@@ -49,6 +55,7 @@ class SuspendedStatus:
 
 
 class EngineProcessor(ABC):
+    """Base classes used for all types of reqest processores."""
 
     __slots__ = 'engine', 'engine_name', 'lock', 'suspended_status'
 
@@ -59,22 +66,28 @@ class EngineProcessor(ABC):
         key = id(key) if key else self.engine_name
         self.suspended_status = SUSPENDED_STATUS.setdefault(key, SuspendedStatus())
 
-    def handle_exception(self, result_container, reason, exception, suspend=False, display_exception=True):
+    def handle_exception(self, result_container, exception_or_message, suspend=False):
         # update result_container
-        error_message = str(exception) if display_exception and exception else None
-        result_container.add_unresponsive_engine(self.engine_name, reason, error_message)
+        if isinstance(exception_or_message, BaseException):
+            exception_class = exception_or_message.__class__
+            module_name = getattr(exception_class, '__module__', 'builtins')
+            module_name = '' if module_name == 'builtins' else module_name + '.'
+            error_message = module_name + exception_class.__qualname__
+        else:
+            error_message = exception_or_message
+        result_container.add_unresponsive_engine(self.engine_name, error_message)
         # metrics
         counter_inc('engine', self.engine_name, 'search', 'count', 'error')
-        if exception:
-            count_exception(self.engine_name, exception)
+        if isinstance(exception_or_message, BaseException):
+            count_exception(self.engine_name, exception_or_message)
         else:
-            count_error(self.engine_name, reason)
+            count_error(self.engine_name, exception_or_message)
         # suspend the engine ?
         if suspend:
             suspended_time = None
-            if isinstance(exception, SearxEngineAccessDeniedException):
-                suspended_time = exception.suspended_time
-            self.suspended_status.suspend(suspended_time, reason)  # pylint: disable=no-member
+            if isinstance(exception_or_message, SearxEngineAccessDeniedException):
+                suspended_time = exception_or_message.suspended_time
+            self.suspended_status.suspend(suspended_time, error_message)  # pylint: disable=no-member
 
     def _extend_container_basic(self, result_container, start_time, search_results):
         # update result_container
@@ -91,7 +104,7 @@ class EngineProcessor(ABC):
     def extend_container(self, result_container, start_time, search_results):
         if getattr(threading.current_thread(), '_timeout', False):
             # the main thread is not waiting anymore
-            self.handle_exception(result_container, 'Timeout', None)
+            self.handle_exception(result_container, 'timeout', None)
         else:
             # check if the engine accepted the request
             if search_results is not None:
@@ -137,9 +150,7 @@ class EngineProcessor(ABC):
         if tests is None:
             tests = getattr(self.engine, 'additional_tests', {})
             tests.update(self.get_default_tests())
-            return tests
-        else:
-            return tests
+        return tests
 
-    def get_default_tests(self):
+    def get_default_tests(self):  # pylint: disable=no-self-use
         return {}

searx/search/processors/offline.py

@@ -1,13 +1,17 @@
 # SPDX-License-Identifier: AGPL-3.0-or-later
+# lint: pylint
 
-from searx import logger
-from searx.search.processors.abstract import EngineProcessor
+"""Processores for engine-type: ``offline``
 
+"""
 
-logger = logger.getChild('searx.search.processor.offline')
+from searx import logger
+from .abstract import EngineProcessor
 
+logger = logger.getChild('searx.search.processor.offline')
 
 class OfflineProcessor(EngineProcessor):
+    """Processor class used by ``offline`` engines"""
 
     engine_type = 'offline'
 
@@ -21,6 +25,6 @@ class OfflineProcessor(EngineProcessor):
         except ValueError as e:
             # do not record the error
             logger.exception('engine {0} : invalid input : {1}'.format(self.engine_name, e))
-        except Exception as e:
-            self.handle_exception(result_container, 'unexpected crash', e)
+        except Exception as e: # pylint: disable=broad-except
+            self.handle_exception(result_container, e)
             logger.exception('engine {0} : exception : {1}'.format(self.engine_name, e))

searx/search/processors/online.py

@@ -1,24 +1,29 @@
 # SPDX-License-Identifier: AGPL-3.0-or-later
+# lint: pylint
+
+"""Processores for engine-type: ``online``
+
+"""
 
 from time import time
 import asyncio
-
 import httpx
 
 import searx.network
 from searx import logger
 from searx.utils import gen_useragent
-from searx.exceptions import (SearxEngineAccessDeniedException, SearxEngineCaptchaException,
-                              SearxEngineTooManyRequestsException,)
+from searx.exceptions import (
+    SearxEngineAccessDeniedException,
+    SearxEngineCaptchaException,
+    SearxEngineTooManyRequestsException,
+)
 from searx.metrics.error_recorder import count_error
-
-from searx.search.processors.abstract import EngineProcessor
-
+from .abstract import EngineProcessor
 
 logger = logger.getChild('searx.search.processor.online')
 
-
 def default_request_params():
+    """Default request parameters for ``online`` engines."""
     return {
         'method': 'GET',
         'headers': {},
@@ -31,6 +36,7 @@ def default_request_params():
 
 
 class OnlineProcessor(EngineProcessor):
+    """Processor class for ``online`` engines."""
 
     engine_type = 'online'
 
@@ -130,7 +136,7 @@ class OnlineProcessor(EngineProcessor):
             self.extend_container(result_container, start_time, search_results)
         except (httpx.TimeoutException, asyncio.TimeoutError) as e:
             # requests timeout (connect or read)
-            self.handle_exception(result_container, 'HTTP timeout', e, suspend=True, display_exception=False)
+            self.handle_exception(result_container, e, suspend=True)
             logger.error("engine {0} : HTTP requests timeout"
                          "(search duration : {1} s, timeout: {2} s) : {3}"
                          .format(self.engine_name, time() - start_time,
@@ -138,23 +144,23 @@ class OnlineProcessor(EngineProcessor):
                                  e.__class__.__name__))
         except (httpx.HTTPError, httpx.StreamError) as e:
             # other requests exception
-            self.handle_exception(result_container, 'HTTP error', e, suspend=True, display_exception=False)
+            self.handle_exception(result_container, e, suspend=True)
             logger.exception("engine {0} : requests exception"
                              "(search duration : {1} s, timeout: {2} s) : {3}"
                              .format(self.engine_name, time() - start_time,
                                      timeout_limit,
                                      e))
         except SearxEngineCaptchaException as e:
-            self.handle_exception(result_container, 'CAPTCHA required', e, suspend=True, display_exception=False)
+            self.handle_exception(result_container, e, suspend=True)
             logger.exception('engine {0} : CAPTCHA'.format(self.engine_name))
         except SearxEngineTooManyRequestsException as e:
-            self.handle_exception(result_container, 'too many requests', e, suspend=True, display_exception=False)
+            self.handle_exception(result_container, e, suspend=True)
             logger.exception('engine {0} : Too many requests'.format(self.engine_name))
         except SearxEngineAccessDeniedException as e:
-            self.handle_exception(result_container, 'blocked', e, suspend=True, display_exception=False)
+            self.handle_exception(result_container, e, suspend=True)
             logger.exception('engine {0} : Searx is blocked'.format(self.engine_name))
-        except Exception as e:
-            self.handle_exception(result_container, 'unexpected crash', e, display_exception=False)
+        except Exception as e:  # pylint: disable=broad-except
+            self.handle_exception(result_container, e)
             logger.exception('engine {0} : exception : {1}'.format(self.engine_name, e))
 
     def get_default_tests(self):

searx/search/processors/online_currency.py

@@ -1,4 +1,8 @@
 # SPDX-License-Identifier: AGPL-3.0-or-later
+# lint: pylint
+"""Processores for engine-type: ``online_currency``
+
+"""
 
 import unicodedata
 import re
@@ -6,32 +10,31 @@ import re
 from searx.data import CURRENCIES
 from .online import OnlineProcessor
 
-
 parser_re = re.compile('.*?(\\d+(?:\\.\\d+)?) ([^.0-9]+) (?:in|to) ([^.0-9]+)', re.I)
 
+# pylint: disable=missing-function-docstring
 
 def normalize_name(name):
     name = name.lower().replace('-', ' ').rstrip('s')
     name = re.sub(' +', ' ', name)
     return unicodedata.normalize('NFKD', name).lower()
 
-
 def name_to_iso4217(name):
-    global CURRENCIES
+    global CURRENCIES  # pylint: disable=global-statement
     name = normalize_name(name)
     currency = CURRENCIES['names'].get(name, [name])
     if isinstance(currency, str):
         return currency
     return currency[0]
 
-
 def iso4217_to_name(iso4217, language):
-    global CURRENCIES
+    global CURRENCIES  # pylint: disable=global-statement
     return CURRENCIES['iso4217'].get(iso4217, {}).get(language, iso4217)
 
-
 class OnlineCurrencyProcessor(OnlineProcessor):
 
+    """Processor class used by ``online_currency`` engines."""
+
     engine_type = 'online_currency'
 
     def get_params(self, search_query, engine_category):

searx/search/processors/online_dictionary.py

@@ -1,15 +1,18 @@
 # SPDX-License-Identifier: AGPL-3.0-or-later
+# lint: pylint
+"""Processores for engine-type: ``online_dictionary``
+
+"""
 
 import re
 
 from searx.utils import is_valid_lang
 from .online import OnlineProcessor
 
-
 parser_re = re.compile('.*?([a-z]+)-([a-z]+) ([^ ]+)$', re.I)
 
-
 class OnlineDictionaryProcessor(OnlineProcessor):
+    """Processor class used by ``online_dictionnary`` engines."""
 
     engine_type = 'online_dictionnary'
 

searx/webapp.py

@@ -172,28 +172,34 @@ _category_names = (gettext('files'),
                    gettext('science'))
 
 #
-exception_classname_to_label = {
-    "searx.exceptions.SearxEngineCaptchaException": gettext("CAPTCHA"),
-    "searx.exceptions.SearxEngineTooManyRequestsException": gettext("too many requests"),
-    "searx.exceptions.SearxEngineAccessDeniedException": gettext("access denied"),
-    "searx.exceptions.SearxEngineAPIException": gettext("server API error"),
-    "httpx.TimeoutException": gettext("HTTP timeout"),
-    "httpx.ConnectTimeout": gettext("HTTP timeout"),
-    "httpx.ReadTimeout": gettext("HTTP timeout"),
-    "httpx.WriteTimeout": gettext("HTTP timeout"),
-    "httpx.HTTPStatusError": gettext("HTTP error"),
-    "httpx.ConnectError": gettext("HTTP connection error"),
-    "httpx.RemoteProtocolError": gettext("HTTP protocol error"),
-    "httpx.LocalProtocolError": gettext("HTTP protocol error"),
-    "httpx.ProtocolError": gettext("HTTP protocol error"),
-    "httpx.ReadError": gettext("network error"),
-    "httpx.WriteError": gettext("network error"),
-    "httpx.ProxyError": gettext("proxy error"),
-    "searx.exceptions.SearxEngineXPathException": gettext("parsing error"),
-    "KeyError": gettext("parsing error"),
-    "json.decoder.JSONDecodeError": gettext("parsing error"),
-    "lxml.etree.ParserError": gettext("parsing error"),
-    None: gettext("unexpected crash"),
+timeout_text = gettext('timeout')
+parsing_error_text = gettext('parsing error')
+http_protocol_error_text = gettext('HTTP protocol error')
+network_error_text = gettext('network error')
+exception_classname_to_text = {
+    None: gettext('unexpected crash'),
+    'timeout': timeout_text,
+    'asyncio.TimeoutError': timeout_text,
+    'httpx.TimeoutException': timeout_text,
+    'httpx.ConnectTimeout': timeout_text,
+    'httpx.ReadTimeout': timeout_text,
+    'httpx.WriteTimeout': timeout_text,
+    'httpx.HTTPStatusError': gettext('HTTP error'),
+    'httpx.ConnectError': gettext("HTTP connection error"),
+    'httpx.RemoteProtocolError': http_protocol_error_text,
+    'httpx.LocalProtocolError': http_protocol_error_text,
+    'httpx.ProtocolError': http_protocol_error_text,
+    'httpx.ReadError': network_error_text,
+    'httpx.WriteError': network_error_text,
+    'httpx.ProxyError': gettext("proxy error"),
+    'searx.exceptions.SearxEngineCaptchaException': gettext("CAPTCHA"),
+    'searx.exceptions.SearxEngineTooManyRequestsException': gettext("too many requests"),
+    'searx.exceptions.SearxEngineAccessDeniedException': gettext("access denied"),
+    'searx.exceptions.SearxEngineAPIException': gettext("server API error"),
+    'searx.exceptions.SearxEngineXPathException': parsing_error_text,
+    'KeyError': parsing_error_text,
+    'json.decoder.JSONDecodeError': parsing_error_text,
+    'lxml.etree.ParserError': parsing_error_text,
 }
 
 _flask_babel_get_translations = flask_babel.get_translations
@@ -786,15 +792,21 @@ def search():
 
 
 def __get_translated_errors(unresponsive_engines):
-    translated_errors = set()
-    for unresponsive_engine in unresponsive_engines:
-        error_msg = gettext(unresponsive_engine[1])
+    translated_errors = []
+    # make a copy unresponsive_engines to avoid "RuntimeError: Set changed size during iteration"
+    # it happens when an engine modifies the ResultContainer after the search_multiple_requests method
+    # has stopped waiting
+    for unresponsive_engine in list(unresponsive_engines):
+        error_user_text = exception_classname_to_text.get(unresponsive_engine[1])
+        if not error_user_text:
+            error_user_text = exception_classname_to_text[None]
+        error_msg = gettext(error_user_text)
         if unresponsive_engine[2]:
             error_msg = "{} {}".format(error_msg, unresponsive_engine[2])
         if unresponsive_engine[3]:
             error_msg = gettext('Suspended') + ': ' + error_msg
-        translated_errors.add((unresponsive_engine[0], error_msg))
-    return translated_errors
+        translated_errors.append((unresponsive_engine[0], error_msg))
+    return sorted(translated_errors, key=lambda e: e[0])
 
 
 @app.route('/about', methods=['GET'])
@@ -944,14 +956,14 @@ def preferences():
         # the first element has the highest percentage rate.
         reliabilities_errors = []
         for error in errors:
-            error_user_message = None
+            error_user_text = None
             if error.get('secondary') or 'exception_classname' not in error:
                 continue
-            error_user_message = exception_classname_to_label.get(error.get('exception_classname'))
+            error_user_text = exception_classname_to_text.get(error.get('exception_classname'))
             if not error:
-                error_user_message = exception_classname_to_label[None]
-            if error_user_message not in reliabilities_errors:
-                reliabilities_errors.append(error_user_message)
+                error_user_text = exception_classname_to_text[None]
+            if error_user_text not in reliabilities_errors:
+                reliabilities_errors.append(error_user_text)
         reliabilities[e.name]['errors'] = reliabilities_errors
 
     # supports