searxngRebrandZaclys/docs/admin/engines/sql-engines.rst
Markus Heiser ecb7e73e03 [brand] docs - normalize project name to SearXNG
Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>
2021-09-13 14:36:24 +02:00

5.0 KiB

SQL Engines

further read

With the SQL engines you can bind SQL databases into SearXNG. The following Relational Database Management System (RDBMS) are supported:

  • engine sqlite
  • engine postgresql
  • engine mysql_server

All of the engines above are just commented out in the settings.yml <searx/settings.yml>, as you have to set the required attributes for the engines, e.g. database: ...

- name: ...
  engine: {sqlite|postgresql|mysql_server}
  database: ...
  result_template: {template_name}
  query_str: ...

By default, the engines use the key-value template for displaying results / see oscar <searx/templates/oscar/result_templates/key-value.html> & simple <searx/templates/simple/result_templates/key-value.html> themes. If you are not satisfied with the original result layout, you can use your own template, set result_template attribute to {template_name} and place the templates at:

searx/templates/{theme_name}/result_templates/{template_name}

If you do not wish to expose these engines on a public instance, you can still add them and limit the access by setting tokens as described in section private engines.

Configure the engines

The configuration of the new database engines are similar. You must put a valid SQL-SELECT query in query_str. At the moment you can only bind at most one parameter in your query. By setting the attribute limit you can define how many results you want from the SQL server. Basically, it is the same as the LIMIT keyword in SQL.

Please, do not include LIMIT or OFFSET in your SQL query as the engines rely on these keywords during paging. If you want to configure the number of returned results use the option limit.

SQLite

info

  • sqlite.py <searx/engines/sqlite.py>

SQLite is a small, fast and reliable SQL database engine. It does not require any extra dependency. To demonstrate the power of database engines, here is a more complex example which reads from a MediathekView (DE) movie database. For this example of the SQlite engine download the database:

and unpack into searx/data/filmliste-v2.db. To search the database use e.g Query to test: !mediathekview concert

- name: mediathekview
  engine: sqlite
  disabled: False
  categories: general
  result_template: default.html
  database: searx/data/filmliste-v2.db
  query_str:  >-
    SELECT title || ' (' || time(duration, 'unixepoch') || ')' AS title,
           COALESCE( NULLIF(url_video_hd,''), NULLIF(url_video_sd,''), url_video) AS url,
           description AS content
      FROM film
     WHERE title LIKE :wildcard OR description LIKE :wildcard
     ORDER BY duration DESC

Extra Dependencies

For using engine postgresql or engine mysql_server you need to install additional packages in Python's Virtual Environment of your SearXNG instance. To switch into the environment (searx-src) you can use searx.sh:

$ sudo utils/searx.sh shell
(searx-pyenv)$ pip install ...

PostgreSQL

info

  • postgresql.py <searx/engines/postgresql.py>
  • pip install psycopg2

PostgreSQL is a powerful and robust open source database. Before configuring the PostgreSQL engine, you must install the dependency psychopg2. You can find an example configuration below:

- name: my_database
  engine: postgresql
  database: my_database
  username: searx
  password: password
  query_str: 'SELECT * from my_table WHERE my_column = %(query)s'

MySQL

info

MySQL is said to be the most popular open source database. Before enabling MySQL engine, you must install the package mysql-connector-python.

The authentication plugin is configurable by setting auth_plugin in the attributes. By default it is set to caching_sha2_password. This is an example configuration for quering a MySQL server:

- name: my_database
  engine: mysql_server
  database: my_database
  username: searx
  password: password
  limit: 5
  query_str: 'SELECT * from my_table WHERE my_column=%(query)s'

Acknowledgment

This development was sponsored by Search and Discovery Fund of NLnet Foundation.