tk0miya / testing.postgresql

Apache License 2.0
292 stars 46 forks source link

About

testing.postgresql automatically setups a postgresql instance in a temporary directory, and destroys it after testing.

.. image:: https://travis-ci.org/tk0miya/testing.postgresql.svg?branch=master :target: https://travis-ci.org/tk0miya/testing.postgresql

.. image:: https://coveralls.io/repos/tk0miya/testing.postgresql/badge.png?branch=master :target: https://coveralls.io/r/tk0miya/testing.postgresql?branch=master

.. image:: https://codeclimate.com/github/tk0miya/testing.postgresql/badges/gpa.svg :target: https://codeclimate.com/github/tk0miya/testing.postgresql

Documentation https://github.com/tk0miya/testing.postgresql Issues https://github.com/tk0miya/testing.postgresql/issues Download https://pypi.python.org/pypi/testing.postgresql

Install

Use pip::

$ pip install testing.postgresql

And testing.postgresql requires PostgreSQL server in your PATH.

Usage

Create PostgreSQL instance using testing.postgresql.Postgresql::

import testing.postgresql from sqlalchemy import create_engine

Lanuch new PostgreSQL server

with testing.postgresql.Postgresql() as postgresql:

connect to PostgreSQL

  engine = create_engine(postgresql.url())

  # if you use postgresql or other drivers:
  #   import psycopg2
  #   db = psycopg2.connect(**postgresql.dsn())

  #
  # do any tests using PostgreSQL...
  #

PostgreSQL server is terminated here

testing.postgresql.Postgresql executes initdb and postgres on instantiation. On deleting Postgresql object, it terminates PostgreSQL instance and removes temporary directory.

If you want a database including tables and any fixtures for your apps, use copy_data_from keyword::

uses a copy of specified data directory of PostgreSQL.

postgresql = testing.postgresql.Postgresql(copy_data_from='/path/to/your/database')

For example, you can setup new PostgreSQL server for each testcases on setUp() method::

import unittest import testing.postgresql

class MyTestCase(unittest.TestCase): def setUp(self): self.postgresql = testing.postgresql.Postgresql()

  def tearDown(self):
      self.postgresql.stop()

To make your tests faster

testing.postgresql.Postgresql invokes initdb command on every instantiation. That is very simple. But, in many cases, it is very waste that generating brandnew database for each testcase.

To optimize the behavior, use testing.postgresql.PostgresqlFactory. The factory class is able to cache the generated database beyond the testcases, and it reduces the number of invocation of initdb command::

import unittest import testing.postgresql

Generate Postgresql class which shares the generated database

Postgresql = testing.postgresql.PostgresqlFactory(cache_initialized_db=True)

def tearDownModule(self):

clear cached database at end of tests

  Postgresql.clear_cache()

class MyTestCase(unittest.TestCase): def setUp(self):

Use the generated Postgresql class instead of testing.postgresql.Postgresql

      self.postgresql = Postgresql()

  def tearDown(self):
      self.postgresql.stop()

If you want to insert fixtures to the cached database, use initdb_handler option::

create initial data on create as fixtures into the database

def handler(postgresql): conn = psycopg2.connect(**postgresql.dsn()) cursor = conn.cursor() cursor.execute("CREATE TABLE hello(id int, value varchar(256))") cursor.execute("INSERT INTO hello values(1, 'hello'), (2, 'ciao')") cursor.close() conn.commit() conn.close()

Use handler() on initialize database

Postgresql = testing.postgresql.PostgresqlFactory(cache_initialized_db=True, on_initialized=handler)

Requirements

License

Apache License 2.0

History

1.3.0 (2016-02-03)

1.2.1 (2015-08-22)

1.2.0 (2015-05-17)

1.1.2 (2015-04-06)

1.1.1 (2015-01-18)

1.1.0 (2014-12-20)

1.0.6 (2014-07-19)

1.0.5 (2014-07-19)

1.0.4 (2014-06-19)

1.0.3 (2014-06-11)

1.0.2 (2013-12-06)

1.0.1 (2013-12-05)

1.0.0 (2013-12-04)

0.1.0 (2013-11-26)