Shillelagh

Shillelagh (ʃɪˈleɪlɪ) is a Python library and CLI that allows you to query many resources (APIs, files, in memory objects) using SQL. It's both user and developer friendly, making it trivial to access resources and easy to add support for new ones.

Learn more on the documentation.

The library is an implementation of the Python DB API 2.0 based on SQLite (using the APSW library):

from shillelagh.backends.apsw.db import connect

connection = connect(":memory:")
cursor = connection.cursor()

query = "SELECT * FROM a_table"
for row in cursor.execute(query):
    print(row)

There is also a SQLAlchemy dialect:

from sqlalchemy.engine import create_engine

engine = create_engine("shillelagh://")
connection = engine.connect()

query = "SELECT * FROM a_table"
for row in connection.execute(query):
    print(row)

And a command-line utility:

$ shillelagh
sql> SELECT * FROM a_table

There is also an experimental backend that uses Postgres with the Multicorn2 extension. First, install the additional dependencies:

$ pip install 'shillelagh[multicorn]'
$ pip install 'multicorn @ git+https://github.com/pgsql-io/multicorn2.git@v2.5'

Then run:

from shillelagh.backends.multicorn.db import connect

connection = connect(
    user="username",
    password="password",
    host="localhost",
    port=5432,
    database="examples",
)

Or:

from sqlalchemy import create_engine
engine = create_engine("shillelagh+multicorn2://username:password@localhost:5432/examples")

Why SQL?

Sharks have been around for a long time. They're older than trees and the rings of Saturn, actually! The reason they haven't changed that much in hundreds of millions of years is because they're really good at what they do.

SQL has been around for some 50 years for the same reason: it's really good at what it does.

Why "Shillelagh"?

Picture a leprechaun hitting APIs with a big stick so that they accept SQL.

How is it different?

Shillelagh allows you to easily query non-SQL resources. For example, if you have a Google Spreadsheet you can query it directly as if it were a table in a database:

SELECT country, SUM(cnt)
FROM "https://docs.google.com/spreadsheets/d/1_rN3lm0R_bU3NemO0s9pbFkY5LQPcuy1pscv8ZXPtg8/edit#gid=0"
WHERE cnt > 0
GROUP BY country

You can even run INSERT/DELETE/UPDATE queries against the spreadsheet:

UPDATE "https://docs.google.com/spreadsheets/d/1_rN3lm0R_bU3NemO0s9pbFkY5LQPcuy1pscv8ZXPtg8/edit#gid=0"
SET cnt = cnt + 1
WHERE country != 'BR'

Queries like this are supported by adapters. Currently Shillelagh has the following adapters:

There are also 3rd-party adapters:

• Airtable
• GraphQL

A query can combine data from multiple adapters:

INSERT INTO "/tmp/file.csv"
SELECT time, chance_of_rain
FROM "https://api.weatherapi.com/v1/history.json?q=London"
WHERE time IN (
  SELECT datetime
  FROM "https://docs.google.com/spreadsheets/d/1_rN3lm0R_bU3NemO0s9pbFkY5LQPcuy1pscv8ZXPtg8/edit#gid=1648320094"
)

The query above reads timestamps from a Google sheet, uses them to filter weather data from WeatherAPI, and writes the chance of rain into a (pre-existing) CSV file.

New adapters are relatively easy to implement. There's a step-by-step tutorial that explains how to create a new adapter to an API or filetype.

Installation

Install Shillelagh with pip:

$ pip install 'shillelagh'

You also need to install optional dependencies, depending on the adapter you want to use:

$ pip install 'shillelagh[console]'        # to use the CLI
$ pip install 'shillelagh[genericjsonapi]' # for Generic JSON
$ pip install 'shillelagh[genericxmlapi]'  # for Generic XML
$ pip install 'shillelagh[githubapi]'      # for GitHub
$ pip install 'shillelagh[gsheetsapi]'     # for GSheets
$ pip install 'shillelagh[htmltableapi]'   # for HTML tables
$ pip install 'shillelagh[pandasmemory]'   # for Pandas in memory
$ pip install 'shillelagh[s3selectapi]'    # for S3 files
$ pip install 'shillelagh[systemapi]'      # for CPU information

Alternatively, you can install everything with:

$ pip install 'shillelagh[all]'