Jose Idar d8a0dcb5e6 Modernize setup.py and add wheel support
* Incremented version to 0.2.2 for pypi release.
  * Removed all post-install hooks.  All initialization and configuration
    should be done through the cafe-config cli tool.
  * Added setup.cfg with support for universal bdist wheel.
  * Updated MANIFEST.in file.
  * Updated README.rst to reflect new install procedures.
  * Modified cafe-config so that initialization command is now just
      "cafe-config init".
  * Since the ".opencafe" directory is no longer initialized at install
    while access to the source code is guaranteed, the plugins are now
    distributed as package_data, and installed as such to site-packages
    under the new "plugins" directory within the "cafe" namespace.
  * The plugins directory is moved to the cafe package directory as
     package_data.
  * The plugin cache is no longer created at initialization, and all
    code relating to it in the cli.py and managers.py file has been
    removed.
  * Removed pip-requires file in favor of including the only requirement,
    'six', in setup.py.  The plan is to refactor so as to remove the
    dependency on six eventually.
  * Renamed test-requirements.txt to test-requires.
  * Added Authors.rst

Change-Id: I28a605f926ae5f2d972a6a36171d0e4eb92cac09
2015-10-11 22:15:08 -05:00
2013-04-01 02:02:01 -05:00
2014-03-05 18:35:55 -06:00
2014-05-15 21:53:20 -05:00

Open CAFE Core

( (

) )

      |___
      
 :-) |
      

=== CAFE Core ===

OpenCAFE, the Open Common Automation Framework Engine, is designed to be used as the base for building an automated testing framework for API and other (non-UI) testing. It is designed to support all kinds of testing methodologies, such as unit, functional and integration testing, using a model-based approach. Although the engine is not designed with performance or load testing in mind, as it prioritizes repeatability and (verbose) logging over performance, it can be used to that end.

Installation

>Source code is available at https://github.com/stackforge/opencafe

Supported Operating Systems

Open CAFE Core has been developed primarily on and for Linux, but supports installation and execution on BSD and other *nix's, as well as OS X and modern Windows. It can be installed from pypi via pip or from source.

>It is recommended that you install OpenCAFE in a python virtualenv.

via pip

pip install opencafe

from source

$ git clone https://github.com/stackforge/opencafe.git

$ cd opencafe $ python setup.py install

Post-install Configuration

Post-install, the cafe-config cli tool will become available. It is used for installing plugins and initializing the engine's default .opencafe directory.

Initialization

OpenCAFE uses a set of default locations for logging, storing test configurations, test data, statistics, and the like; all of which are set in, and read from, the engine.config file (in order to make it easy for the end user to override the default behavior). The engine.config file, and the directories it references, can be created on demand by running:

cafe-config init

> This will create a directory named .opencafe in the user's home directory, or in the case of a python virtualenv, in the virtualenv root folder.

Installing Plugins

The cafe-config plugins command is used to list and install plugins.

Example:

$ cafe-config plugins list

* Available Plugins

... http ... mongo ... winrm ... elasticsearch ... soap ... skip_on_issue ... rsyslog ... ssh


$ cafe-config plugins install http ================================= * Installing Plugins ... http =================================

> Note: There is currently no way to reliably have an implementation of

OpenCAFE require specific plugins at install time, due to issues with pip=>7.0.0. We are working on an overhaul of the plugin system that should remedy the situation.

Package Structure Overview

cafe.common.reporting

Provides tools for logging and reporting. This namespace should be used by plugins to add logging and reporting features.

cafe.configurator

Used by the cafe-config cli tool for setting up new installations of opencafe.

cafe.drivers

Houses various test runner wrappers and supporting tools. This namespace should be used by plugins to add new test runner support.

cafe.engine

Includes the base classes that OpenCAFE implementations will use to create behaviors, configs, clients and models. This namespace should be used by plugins to add new clients.

cafe.resources

Deprecated. Historically contained all modules that reference external resources to OpenCAFE. Currently acts only as a namespace for backward compatability with some plugins.

Terminology

Following are some notes on Open CAFE lingo and concepts.

  • Implementation

    Although the engine can serve as a basic framework for testing, it's meant to be used as the base for the implementation of a product-specific testing framework.

  • Product

    Anything that's being tested by an implementation of Open CAFE Core. If you would like to see a refernce implementation, there is an Open Source implementation based on OpenStack.

  • Client / Client Method

    A client is an "at-least-one"-to-"at-most-one" mapping of a product's functionality to a collection of client methods. Using a REST API as an example, a client that represents that API in CAFE will contain at least one (but possibly more) method(s) for every function exposed by that API. Should a call in the API prove to be too difficult or cumbersome to define via a single client method, then multiple client methods can be defined such that as a whole they represent the complete set of that API call's functionality. A client method should never be a superset of more than one call's functionality.

  • Behavior

    A behavior is a many-to-many mapping of client methods to business logic, functioning as compound methods. An example behavior might be to POST content, perform a GET to verify the POST, and then return the verified data

  • Model

    A model can be many things, but generally is a class that describes a specific data object. An example may be a collection of logic for converting an XML or JSON response into a data object, so that a single consumer can be written to consume the model.

  • Provider

    This is meant to be a convenience facade that performs configuration of clients and behaviors to provide configuration-based default combinations of different clients and behaviors.

Description
Common Automation Framework Engine. Core driver for CloudCAFE
Readme 3.2 MiB
Languages
Python 99.1%
reStructuredText 0.9%