asql

Query a database in natural language. Uses docker images from https://github.com/paulfitz/mlsql - make sure Docker RAM limits are not too restrictive (should be above 3GB).

Installation

• pip install asql to work with Sqlite or CSV files.
• pip install asql[postgres] to work with PostgreSQL databases.
• pip install asql[mysql] to work with MySQL databases.
• For other databases, see SQLAlchemy supported dialects.

Use

Tell asql which model to use:

# pick one of these
asql --docker sqlova
asql --docker valuenet

I suggest you start with just one of these, although you can start both. The sqlova model works on single tables (e.g. a csv file) and can handle queries that take parameters. The valuenet model works on many tables, but is less strong at queries that take parameters.

Tell asql which data to use:

$ pick one of these
asql --db your_data.csv
asql --db your_data.sqlite
asql --db postgres://user:password@host/database
asql --db mysql://user:password@host/database

Don't use on an enormous database just yet.

Now, ask whatever questions you like in plain English, and see what happens:

asql how many players are there?
asql which is the longest bridge?