[enh] settings.yml: add use_default_settings option

This change is backward compatible with the existing configurations.

If a settings.yml loaded from an user defined location (SEARX_SETTINGS_PATH or /etc/searx/settings.yml),
then this settings can relied on the default settings.yml with this option:
user_default_settings:True

Makefile

@@ -266,4 +266,19 @@ test.clean:
 travis.codecov:
 	$(Q)$(PY_ENV_BIN)/python -m pip install codecov
 
-.PHONY: $(PHONY)
+
+# user-settings
+# -------------
+
+PHONY += user-settings.create user-settings.update
+
+user-settings.update:  pyenvinstall
+	$(Q)$(PY_ENV_ACT); pip install ruamel.yaml
+	$(Q)$(PY_ENV_ACT); python utils/update_user_settings.py ${SEARX_SETTINGS_PATH}
+
+user-settings.update.engines:  pyenvinstall
+	$(Q)$(PY_ENV_ACT); pip install ruamel.yaml
+	$(Q)$(PY_ENV_ACT); python utils/update_user_settings.py --add-engines ${SEARX_SETTINGS_PATH}
+
+
+.PHONY: $(PHONY)

docs/admin/settings.rst

@@ -206,3 +206,97 @@ 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"
+  server:
+      bind_address: "0.0.0.0"
+
+With ``use_default_settings: True``, each settings can be override in a similar way with one exception, the ``engines`` section:
+
+* If the ``engines`` section is not defined in the user settings, searx uses the engines from the default setttings (the above example).
+* If the ``engines`` section is defined then:
+
+   * searx loads only the engines declare in the user setttings.
+   * searx merges the configuration according to the engine name.
+
+In the following example, only three engines are available. Each engine configuration is merged with the default configuration.
+
+.. code-block:: yaml
+
+  use_default_settings: true
+  server:
+      secret_key: "uvys6bRhKHUdFF5CqbJonSDSRN8H0sCBziNSrDGNVdpz7IeZhveVart3yvghoKHA"
+  engines:
+    - name: wikipedia
+    - name: wikidata
+    - name: ddg definitions
+
+Another example where four engines are available. 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']
+    - name: wikipedia
+    - name: wikidata
+    - name: ddg definitions
+
+automatic update
+----------------
+
+The following comand creates or updates a minimal user settings (a secret key is defined if it is not already the case):
+
+.. code-block:: sh
+
+  make SEARX_SETTINGS_PATH=/etc/searx/settings.yml user-settings.update
+
+Set ``SEARX_SETTINGS_PATH`` to your user settings path.
+
+As soon the user settings contains an ``engines`` section, it becomes difficult to keep the engine list updated.
+The following command creates or updates the user settings including the ``engines`` section:
+
+.. code-block:: sh
+
+  make SEARX_SETTINGS_PATH=/etc/searx/settings.yml user-settings.update.engines
+
+After that ``/etc/searx/settings.yml``
+
+* has a ``secret key``
+* has a ``engine`` section if it is not already the case, moreover the command:
+
+  * has deleted engines that do not exist in the default settings.
+  * has added engines that exist in the default settings but are not declare in the user settings.

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
 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.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.py

@@ -0,0 +1,91 @@
+import collections.abc
+
+import yaml
+from searx.exceptions import SearxSettingsException
+from os import environ
+from os.path import dirname, join, abspath, isfile
+
+
+searx_dir = abspath(dirname(__file__))
+
+
+def check_settings_yml(file_name):
+    if isfile(file_name):
+        return file_name
+    else:
+        return None
+
+
+def load_yaml(file_name):
+    try:
+        with open(file_name, 'r', encoding='utf-8') as settings_yaml:
+            settings = yaml.safe_load(settings_yaml)
+            if not isinstance(settings, dict) or len(settings) == 0:
+                raise SearxSettingsException('Empty file', file_name)
+            return settings
+    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'])
+    else:
+        # 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(d, u):
+    for k, v in u.items():
+        if isinstance(v, collections.abc.Mapping):
+            d[k] = update_dict(d.get(k, {}), v)
+        else:
+            d[k] = v
+    return d
+
+
+def update_settings(default_settings, user_settings):
+    for k, v in user_settings.items():
+        if k == 'use_default_settings':
+            continue
+        elif k == 'engines':
+            default_engines = default_settings[k]
+            default_engines_dict = dict((definition['name'], definition) for definition in default_engines)
+            default_settings[k] = [update_dict(default_engines_dict[definition['name']], definition)
+                                   for definition in v]
+        else:
+            update_dict(default_settings[k], v)
+
+    return 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 user_settings.get('use_default_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))

utils/update_user_settings.py

@@ -0,0 +1,98 @@
+#!/usr/bin/env python
+
+# set path
+from sys import path
+from os.path import realpath, dirname, join
+path.append(realpath(dirname(realpath(__file__)) + '/../'))
+
+import argparse
+import sys
+import string
+import ruamel.yaml
+import secrets
+import collections
+from ruamel.yaml.scalarstring import SingleQuotedScalarString, DoubleQuotedScalarString
+from searx.settings import load_settings, check_settings_yml, get_default_settings_path
+from searx.exceptions import SearxSettingsException
+
+
+RANDOM_STRING_LETTERS = string.ascii_lowercase + string.digits + string.ascii_uppercase
+
+
+def get_random_string():
+    r = [secrets.choice(RANDOM_STRING_LETTERS) for _ in range(64)]
+    return ''.join(r)
+
+
+def main(prog_arg):
+    yaml = ruamel.yaml.YAML()
+    yaml.preserve_quotes = True
+    yaml.indent(mapping=4, sequence=1, offset=2)
+    user_settings_path = prog_args.get('user-settings-yaml')
+
+    try:
+        default_settings, _ = load_settings(False)
+        if check_settings_yml(user_settings_path):
+            with open(user_settings_path, 'r', encoding='utf-8') as f:
+                user_settings = yaml.load(f.read())
+            new_user_settings = False
+        else:
+            user_settings = yaml.load('use_default_settings: True')
+            new_user_settings = True
+    except SearxSettingsException as e:
+        sys.stderr.write(str(e))
+        return
+
+    if not new_user_settings and not user_settings.get('use_default_settings'):
+        sys.stderr.write('settings.yml already exists and use_default_settings is not True')
+        return
+
+    user_settings['use_default_settings'] = True
+    use_default_settings_comment = "settings based on " + get_default_settings_path()
+    user_settings.yaml_add_eol_comment(use_default_settings_comment, 'use_default_settings')
+
+    if user_settings.get('server', {}).get('secret_key') in [None, 'ultrasecretkey']:
+        user_settings.setdefault('server', {})['secret_key'] = DoubleQuotedScalarString(get_random_string())
+
+    user_engines = user_settings.get('engines')
+    if user_engines:
+        has_user_engines = True
+        user_engines_dict = dict((definition['name'], definition) for definition in user_engines)
+    else:
+        has_user_engines = False
+        user_engines_dict = {}
+        user_engines = []
+
+    # remove old engines
+    if prog_arg.get('add-engines') or has_user_engines:
+        default_engines_dict = dict((definition['name'], definition) for definition in default_settings['engines'])
+        for i, engine in enumerate(user_engines):
+            if engine['name'] not in default_engines_dict:
+                del user_engines[i]
+
+    # add new engines
+    if prog_arg.get('add-engines'):
+        for engine in default_settings.get('engines', {}):
+            if engine['name'] not in user_engines_dict:
+                user_engines.append({'name': engine['name']})
+        user_settings['engines'] = user_engines
+
+    # output
+    if prog_arg.get('dry-run'):
+        yaml.dump(user_settings, sys.stdout)
+    else:
+        with open(user_settings_path, 'w', encoding='utf-8') as f:
+            yaml.dump(user_settings, f)
+
+
+def parse_args():
+    parser = argparse.ArgumentParser(description='Update user settings.yml')
+    parser.add_argument('--add-engines', dest='add-engines', default=False, action='store_true', help='Add new engines')
+    parser.add_argument('--dry-run', dest='dry-run', default=False, action='store_true', help='Dry run')
+    parser.add_argument('user-settings-yaml', type=str)
+    return vars(parser.parse_args())
+
+
+if __name__ == '__main__':
+    prog_args = parse_args()
+    main(prog_args)