Merge pull request #2291 from dalf/settings2

[enh] user settings can relied on the default settings

docs/admin/settings.rst

@@ -206,3 +206,80 @@ Engine settings
 
    A few more options are possible, but they are pretty specific to some
    engines, and so won't be described here.
+
+
+.. _settings location:
+
+settings.yml location
+=====================
+
+First, searx will try to load settings.yml from these locations:
+
+1. the full path specified in the ``SEARX_SETTINGS_PATH`` environment variable.
+2. ``/etc/searx/settings.yml``
+
+If these files don't exist (or are empty or can't be read), searx uses the :origin:`searx/settings.yml` file.
+
+.. _ settings use_default_settings:
+
+use_default_settings
+====================
+
+.. note::
+
+   If searx is cloned from a git repository, most probably there is no need to have an user settings.
+
+The user defined settings.yml can relied on the default configuration :origin:`searx/settings.yml` using ``use_default_settings: True``.
+
+In the following example, the actual settings are the default settings defined in :origin:`searx/settings.yml` with the exception of the ``secret_key`` and the ``bind_address``:
+
+.. code-block:: yaml
+
+  use_default_settings: True
+  server:
+      secret_key: "uvys6bRhKHUdFF5CqbJonSDSRN8H0sCBziNSrDGNVdpz7IeZhveVart3yvghoKHA"
+      bind_address: "0.0.0.0"
+
+With ``use_default_settings: True``, each settings can be override in a similar way, the ``engines`` section is merged according to the engine ``name``.
+
+In this example, searx will load all the engine and the arch linux wiki engine has a :ref:`token<private engines>`:
+
+.. code-block:: yaml
+
+  use_default_settings: True
+  server:
+      secret_key: "uvys6bRhKHUdFF5CqbJonSDSRN8H0sCBziNSrDGNVdpz7IeZhveVart3yvghoKHA"
+  engines:
+    - name: arch linux wiki
+      tokens: ['$ecretValue']
+
+It is possible to remove some engines from the default settings. The following example is similar to the above one, but searx doesn't load the the google engine:
+
+.. code-block:: yaml
+
+  use_default_settings:
+      engines:
+         remove:
+           - google
+  server:
+      secret_key: "uvys6bRhKHUdFF5CqbJonSDSRN8H0sCBziNSrDGNVdpz7IeZhveVart3yvghoKHA"
+  engines:
+    - name: arch linux wiki
+      tokens: ['$ecretValue']
+
+As an alternative, it is possible to specify the engines to keep. In the following example, searx has only two engines:
+
+.. code-block:: yaml
+
+  use_default_settings:
+      engines:
+         keep_only:
+           - google
+           - duckduckgo
+  server:
+      secret_key: "uvys6bRhKHUdFF5CqbJonSDSRN8H0sCBziNSrDGNVdpz7IeZhveVart3yvghoKHA"
+  engines:
+    - name: google
+      tokens: ['$ecretValue']
+    - name: duckduckgo
+      tokens: ['$ecretValue']

docs/blog/private-engines.rst

@@ -7,6 +7,8 @@ enabled engines on their instances. It might be because they do not want to
 expose some private information through an offline engine. Or they
 would rather share engines only with their trusted friends or colleagues.
 
+.. _private engines:
+
 Private engines
 ===============
 

searx/__init__.py

@@ -16,39 +16,15 @@ along with searx. If not, see < http://www.gnu.org/licenses/ >.
 '''
 
 import logging
+import searx.settings_loader
 from os import environ
 from os.path import realpath, dirname, join, abspath, isfile
-from io import open
-from yaml import safe_load
 
 
 searx_dir = abspath(dirname(__file__))
 engine_dir = dirname(realpath(__file__))
 static_path = abspath(join(dirname(__file__), 'static'))
-
-
-def check_settings_yml(file_name):
-    if isfile(file_name):
-        return file_name
-    else:
-        return None
-
-
-# find location of settings.yml
-if 'SEARX_SETTINGS_PATH' in environ:
-    # if possible set path to settings using the
-    # enviroment variable SEARX_SETTINGS_PATH
-    settings_path = check_settings_yml(environ['SEARX_SETTINGS_PATH'])
-else:
-    # if not, get it from searx code base or last solution from /etc/searx
-    settings_path = check_settings_yml(join(searx_dir, 'settings.yml')) or check_settings_yml('/etc/searx/settings.yml')
-
-if not settings_path:
-    raise Exception('settings.yml not found')
-
-# load settings
-with open(settings_path, 'r', encoding='utf-8') as settings_yaml:
-    settings = safe_load(settings_yaml)
+settings, settings_load_message = searx.settings_loader.load_settings()
 
 if settings['ui']['static_path']:
     static_path = settings['ui']['static_path']
@@ -58,7 +34,6 @@ enable debug if
 the environnement variable SEARX_DEBUG is 1 or true
 (whatever the value in settings.yml)
 or general.debug=True in settings.yml
-
 disable debug if
 the environnement variable SEARX_DEBUG is 0 or false
 (whatever the value in settings.yml)
@@ -78,7 +53,7 @@ else:
     logging.basicConfig(level=logging.WARNING)
 
 logger = logging.getLogger('searx')
-logger.debug('read configuration from %s', settings_path)
+logger.info(settings_load_message)
 logger.info('Initialisation done')
 
 if 'SEARX_SECRET' in environ:

searx/exceptions.py

@@ -31,3 +31,11 @@ class SearxParameterException(SearxException):
         self.message = message
         self.parameter_name = name
         self.parameter_value = value
+
+
+class SearxSettingsException(SearxException):
+
+    def __init__(self, message, filename):
+        super().__init__(message)
+        self.message = message
+        self.filename = filename

searx/settings_loader.py

@@ -0,0 +1,129 @@
+# SPDX-License-Identifier: AGPL-3.0-or-later
+
+from os import environ
+from os.path import dirname, join, abspath, isfile
+from collections.abc import Mapping
+from itertools import filterfalse
+
+import yaml
+
+from searx.exceptions import SearxSettingsException
+
+
+searx_dir = abspath(dirname(__file__))
+
+
+def check_settings_yml(file_name):
+    if isfile(file_name):
+        return file_name
+    return None
+
+
+def load_yaml(file_name):
+    try:
+        with open(file_name, 'r', encoding='utf-8') as settings_yaml:
+            return yaml.safe_load(settings_yaml)
+    except IOError as e:
+        raise SearxSettingsException(e, file_name)
+    except yaml.YAMLError as e:
+        raise SearxSettingsException(e, file_name)
+
+
+def get_default_settings_path():
+    return check_settings_yml(join(searx_dir, 'settings.yml'))
+
+
+def get_user_settings_path():
+    # find location of settings.yml
+    if 'SEARX_SETTINGS_PATH' in environ:
+        # if possible set path to settings using the
+        # enviroment variable SEARX_SETTINGS_PATH
+        return check_settings_yml(environ['SEARX_SETTINGS_PATH'])
+
+    # if not, get it from searx code base or last solution from /etc/searx
+    return check_settings_yml('/etc/searx/settings.yml')
+
+
+def update_dict(default_dict, user_dict):
+    for k, v in user_dict.items():
+        if isinstance(v, Mapping):
+            default_dict[k] = update_dict(default_dict.get(k, {}), v)
+        else:
+            default_dict[k] = v
+    return default_dict
+
+
+def update_settings(default_settings, user_settings):
+    # merge everything except the engines
+    for k, v in user_settings.items():
+        if k not in ('use_default_settings', 'engines'):
+            update_dict(default_settings[k], v)
+
+    # parse the engines
+    remove_engines = None
+    keep_only_engines = None
+    use_default_settings = user_settings.get('use_default_settings')
+    if isinstance(use_default_settings, dict):
+        remove_engines = use_default_settings.get('engines', {}).get('remove')
+        keep_only_engines = use_default_settings.get('engines', {}).get('keep_only')
+
+    if 'engines' in user_settings or remove_engines is not None or keep_only_engines is not None:
+        engines = default_settings['engines']
+
+        # parse "use_default_settings.engines.remove"
+        if remove_engines is not None:
+            engines = list(filterfalse(lambda engine: (engine.get('name')) in remove_engines, engines))
+
+        # parse "use_default_settings.engines.keep_only"
+        if keep_only_engines is not None:
+            engines = list(filter(lambda engine: (engine.get('name')) in keep_only_engines, engines))
+
+        # parse "engines"
+        user_engines = user_settings.get('engines')
+        if user_engines:
+            engines_dict = dict((definition['name'], definition) for definition in engines)
+            for user_engine in user_engines:
+                default_engine = engines_dict.get(user_engine['name'])
+                if default_engine:
+                    update_dict(default_engine, user_engine)
+                else:
+                    engines.append(user_engine)
+
+        # store the result
+        default_settings['engines'] = engines
+
+    return default_settings
+
+
+def is_use_default_settings(user_settings):
+    use_default_settings = user_settings.get('use_default_settings')
+    if use_default_settings is True:
+        return True
+    if isinstance(use_default_settings, dict):
+        return True
+    if use_default_settings is False or use_default_settings is None:
+        return False
+    raise ValueError('Invalid value for use_default_settings')
+
+
+def load_settings(load_user_setttings=True):
+    default_settings_path = get_default_settings_path()
+    user_settings_path = get_user_settings_path()
+    if user_settings_path is None or not load_user_setttings:
+        # no user settings
+        return (load_yaml(default_settings_path),
+                'load the default settings from {}'.format(default_settings_path))
+
+    # user settings
+    user_settings = load_yaml(user_settings_path)
+    if is_use_default_settings(user_settings):
+        # the user settings are merged with the default configuration
+        default_settings = load_yaml(default_settings_path)
+        update_settings(default_settings, user_settings)
+        return (default_settings,
+                'merge the default settings ( {} ) and the user setttings ( {} )'
+                .format(default_settings_path, user_settings_path))
+
+    # the user settings, fully replace the default configuration
+    return (user_settings,
+            'load the user settings from {}'.format(user_settings_path))

tests/unit/settings/syntaxerror_settings.yml

@@ -0,0 +1,2 @@
+Test:
+  **********

tests/unit/settings/user_settings.yml

@@ -0,0 +1,111 @@
+general:
+    debug : False
+    instance_name : "searx"
+
+search:
+    safe_search : 0
+    autocomplete : ""
+    default_lang : ""
+    ban_time_on_fail : 5
+    max_ban_time_on_fail : 120
+
+server:
+    port : 9000
+    bind_address : "0.0.0.0"
+    secret_key : "user_settings_secret"
+    base_url : False
+    image_proxy : False
+    http_protocol_version : "1.0"
+    method: "POST"
+    default_http_headers:
+        X-Content-Type-Options : nosniff
+        X-XSS-Protection : 1; mode=block
+        X-Download-Options : noopen
+        X-Robots-Tag : noindex, nofollow
+        Referrer-Policy : no-referrer
+
+ui:
+    static_path : ""
+    templates_path : ""
+    default_theme : oscar
+    default_locale : ""
+    theme_args :
+        oscar_style : logicodev
+
+engines:
+  - name : wikidata
+    engine : wikidata
+    shortcut : wd
+    timeout : 3.0
+    weight : 2
+
+  - name : wikibooks
+    engine : mediawiki
+    shortcut : wb
+    categories : general
+    base_url : "https://{language}.wikibooks.org/"
+    number_of_results : 5
+    search_type : text
+
+  - name : wikinews
+    engine : mediawiki
+    shortcut : wn
+    categories : news
+    base_url : "https://{language}.wikinews.org/"
+    number_of_results : 5
+    search_type : text
+
+  - name : wikiquote
+    engine : mediawiki
+    shortcut : wq
+    categories : general
+    base_url : "https://{language}.wikiquote.org/"
+    number_of_results : 5
+    search_type : text
+
+locales:
+    en : English
+    ar : العَرَبِيَّة (Arabic)
+    bg : Български (Bulgarian)
+    bo : བོད་སྐད་ (Tibetian)
+    ca : Català (Catalan)
+    cs : Čeština (Czech)
+    cy : Cymraeg (Welsh)
+    da : Dansk (Danish)
+    de : Deutsch (German)
+    el_GR : Ελληνικά (Greek_Greece)
+    eo : Esperanto (Esperanto)
+    es : Español (Spanish)
+    et : Eesti (Estonian)
+    eu : Euskara (Basque)
+    fa_IR : (fārsī) فارسى (Persian)
+    fi : Suomi (Finnish)
+    fil : Wikang Filipino (Filipino)
+    fr : Français (French)
+    gl : Galego (Galician)
+    he : עברית (Hebrew)
+    hr : Hrvatski (Croatian)
+    hu : Magyar (Hungarian)
+    ia : Interlingua (Interlingua)
+    it : Italiano (Italian)
+    ja : 日本語 (Japanese)
+    lt : Lietuvių (Lithuanian)
+    nl : Nederlands (Dutch)
+    nl_BE : Vlaams (Dutch_Belgium)
+    oc : Lenga D'òc (Occitan)
+    pl : Polski (Polish)
+    pt : Português (Portuguese)
+    pt_BR : Português (Portuguese_Brazil)
+    ro : Română (Romanian)
+    ru : Русский (Russian)
+    sk : Slovenčina (Slovak)
+    sl : Slovenski (Slovene)
+    sr : српски (Serbian)
+    sv : Svenska (Swedish)
+    te : తెలుగు (telugu)
+    ta : தமிழ் (Tamil)
+    tr : Türkçe (Turkish)
+    uk : українська мова (Ukrainian)
+    vi : tiếng việt (Vietnamese)
+    zh : 中文 (Chinese)
+    zh_TW : 國語 (Taiwanese Mandarin)

tests/unit/settings/user_settings_keep_only.yml

@@ -0,0 +1,14 @@
+use_default_settings:
+    engines:
+        keep_only:
+            - wikibooks
+            - wikinews
+server:
+    secret_key: "user_secret_key"
+    bind_address: "0.0.0.0"
+    default_http_headers:
+        Custom-Header: Custom-Value
+engines:
+    - name: wikipedia
+    - name: newengine
+      engine: dummy

tests/unit/settings/user_settings_remove.yml

@@ -0,0 +1,10 @@
+use_default_settings:
+    engines:
+        remove:
+            - wikibooks
+            - wikinews
+server:
+    secret_key: "user_secret_key"
+    bind_address: "0.0.0.0"
+    default_http_headers:
+        Custom-Header: Custom-Value

tests/unit/settings/user_settings_remove2.yml

@@ -0,0 +1,15 @@
+use_default_settings:
+    engines:
+        remove:
+            - wikibooks
+            - wikinews
+server:
+    secret_key: "user_secret_key"
+    bind_address: "0.0.0.0"
+    default_http_headers:
+        Custom-Header: Custom-Value
+engines:
+    - name: wikipedia
+      tokens: ['secret_token']
+    - name: newengine
+      engine: dummy

tests/unit/settings/user_settings_simple.yml

@@ -0,0 +1,6 @@
+use_default_settings: True
+server:
+    secret_key: "user_secret_key"
+    bind_address: "0.0.0.0"
+    default_http_headers:
+        Custom-Header: Custom-Value

tests/unit/test_settings_loader.py

@@ -0,0 +1,122 @@
+# SPDX-License-Identifier: AGPL-3.0-or-later
+
+from os.path import dirname, join, abspath
+from unittest.mock import patch
+
+from searx.testing import SearxTestCase
+from searx.exceptions import SearxSettingsException
+from searx import settings_loader
+
+
+test_dir = abspath(dirname(__file__))
+
+
+class TestLoad(SearxTestCase):
+
+    def test_load_zero(self):
+        with self.assertRaises(SearxSettingsException):
+            settings_loader.load_yaml('/dev/zero')
+
+        with self.assertRaises(SearxSettingsException):
+            settings_loader.load_yaml(join(test_dir, '/settings/syntaxerror_settings.yml'))
+
+        with self.assertRaises(SearxSettingsException):
+            settings_loader.load_yaml(join(test_dir, '/settings/empty_settings.yml'))
+
+    def test_check_settings_yml(self):
+        self.assertIsNone(settings_loader.check_settings_yml('/dev/zero'))
+
+        bad_settings_path = join(test_dir, 'settings/syntaxerror_settings.yml')
+        self.assertEqual(settings_loader.check_settings_yml(bad_settings_path), bad_settings_path)
+
+
+class TestDefaultSettings(SearxTestCase):
+
+    def test_load(self):
+        settings, msg = settings_loader.load_settings(load_user_setttings=False)
+        self.assertTrue(msg.startswith('load the default settings from'))
+        self.assertFalse(settings['general']['debug'])
+        self.assertTrue(isinstance(settings['general']['instance_name'], str))
+        self.assertEqual(settings['server']['secret_key'], "ultrasecretkey")
+        self.assertTrue(isinstance(settings['server']['port'], int))
+        self.assertTrue(isinstance(settings['server']['bind_address'], str))
+        self.assertTrue(isinstance(settings['engines'], list))
+        self.assertTrue(isinstance(settings['locales'], dict))
+        self.assertTrue(isinstance(settings['doi_resolvers'], dict))
+        self.assertTrue(isinstance(settings['default_doi_resolver'], str))
+
+
+class TestUserSettings(SearxTestCase):
+
+    def test_is_use_default_settings(self):
+        self.assertFalse(settings_loader.is_use_default_settings({}))
+        self.assertTrue(settings_loader.is_use_default_settings({'use_default_settings': True}))
+        self.assertTrue(settings_loader.is_use_default_settings({'use_default_settings': {}}))
+        with self.assertRaises(ValueError):
+            self.assertFalse(settings_loader.is_use_default_settings({'use_default_settings': 1}))
+        with self.assertRaises(ValueError):
+            self.assertFalse(settings_loader.is_use_default_settings({'use_default_settings': 0}))
+
+    def test_user_settings_not_found(self):
+        with patch.dict(settings_loader.environ,
+                        {'SEARX_SETTINGS_PATH': '/dev/null'}):
+            settings, msg = settings_loader.load_settings()
+            self.assertTrue(msg.startswith('load the default settings from'))
+            self.assertEqual(settings['server']['secret_key'], "ultrasecretkey")
+
+    def test_user_settings(self):
+        with patch.dict(settings_loader.environ,
+                        {'SEARX_SETTINGS_PATH': join(test_dir, 'settings/user_settings_simple.yml')}):
+            settings, msg = settings_loader.load_settings()
+            self.assertTrue(msg.startswith('merge the default settings'))
+            self.assertEqual(settings['server']['secret_key'], "user_secret_key")
+            self.assertEqual(settings['server']['default_http_headers']['Custom-Header'], "Custom-Value")
+
+    def test_user_settings_remove(self):
+        with patch.dict(settings_loader.environ,
+                        {'SEARX_SETTINGS_PATH': join(test_dir, 'settings/user_settings_remove.yml')}):
+            settings, msg = settings_loader.load_settings()
+            self.assertTrue(msg.startswith('merge the default settings'))
+            self.assertEqual(settings['server']['secret_key'], "user_secret_key")
+            self.assertEqual(settings['server']['default_http_headers']['Custom-Header'], "Custom-Value")
+            engine_names = [engine['name'] for engine in settings['engines']]
+            self.assertNotIn('wikinews', engine_names)
+            self.assertNotIn('wikibooks', engine_names)
+            self.assertIn('wikipedia', engine_names)
+
+    def test_user_settings_remove2(self):
+        with patch.dict(settings_loader.environ,
+                        {'SEARX_SETTINGS_PATH': join(test_dir, 'settings/user_settings_remove2.yml')}):
+            settings, msg = settings_loader.load_settings()
+            self.assertTrue(msg.startswith('merge the default settings'))
+            self.assertEqual(settings['server']['secret_key'], "user_secret_key")
+            self.assertEqual(settings['server']['default_http_headers']['Custom-Header'], "Custom-Value")
+            engine_names = [engine['name'] for engine in settings['engines']]
+            self.assertNotIn('wikinews', engine_names)
+            self.assertNotIn('wikibooks', engine_names)
+            self.assertIn('wikipedia', engine_names)
+            wikipedia = list(filter(lambda engine: (engine.get('name')) == 'wikipedia', settings['engines']))
+            self.assertEqual(wikipedia[0]['engine'], 'wikipedia')
+            self.assertEqual(wikipedia[0]['tokens'], ['secret_token'])
+            newengine = list(filter(lambda engine: (engine.get('name')) == 'newengine', settings['engines']))
+            self.assertEqual(newengine[0]['engine'], 'dummy')
+
+    def test_user_settings_keep_only(self):
+        with patch.dict(settings_loader.environ,
+                        {'SEARX_SETTINGS_PATH': join(test_dir, 'settings/user_settings_keep_only.yml')}):
+            settings, msg = settings_loader.load_settings()
+            self.assertTrue(msg.startswith('merge the default settings'))
+            engine_names = [engine['name'] for engine in settings['engines']]
+            self.assertEqual(engine_names, ['wikibooks', 'wikinews', 'wikipedia', 'newengine'])
+            # wikipedia has been removed, then added again with the "engine" section of user_settings_keep_only.yml
+            self.assertEqual(len(settings['engines'][2]), 1)
+
+    def test_custom_settings(self):
+        with patch.dict(settings_loader.environ,
+                        {'SEARX_SETTINGS_PATH': join(test_dir, 'settings/user_settings.yml')}):
+            settings, msg = settings_loader.load_settings()
+            self.assertTrue(msg.startswith('load the user settings from'))
+            self.assertEqual(settings['server']['port'], 9000)
+            self.assertEqual(settings['server']['secret_key'], "user_settings_secret")
+            engine_names = [engine['name'] for engine in settings['engines']]
+            self.assertEqual(engine_names, ['wikidata', 'wikibooks', 'wikinews', 'wikiquote'])