#Batch Processing RESTFul Client

A Dev/DevOps tool to automate calling a set of RESTFul JSON endpoints, with the ability to grab data from the JSON payload and use it in subsequent calls in the recipe.

If you're a Dev/DevOps engineer you may have been faced with a situation where you find yourself writing a shell script to automate getting JSON data from one or more RESTful endpoints using curl or wget or httpie, and then parsing the output using sed, grep or jq. This tool is designed to provide a generic, simplied, yet powerful means of writing such scripts in the form of a simple recipe specification, rather than a shell script.

It relies on the excellent Pygments and Requests libraries.

This is a Python application which is written in Python 3.

It has only currently been tested on Linux (Ubuntu 14.04) and MacOS X (El Capitan)

Make sure you've installed Python 3 and Pip 3

$ apt-get install python3
$ apt-get install python3-pip

On the mac you need homebrew first. Then just install python3. The homebrew package ships with pip3 so nothing additional should be required.

$ brew install python3

Install like this on either Linux or Mac OS X:

$ pip3 install bprc

Or a tagged version directly from GitHub

$ # replace x.y.z with GitHub version tag -- no tar.gz extension needed
$ pip3 install https://github.com/bradwood/BPRC/tarball/x.y.x 

The recipe is specified in a single YAML file which describes:

• A list of variables that are initialised at the top of the recipe for use in the steps below.
• A list of ordered steps that comprise the recipe, each containing some or all of the below:
  • the URL that need to be visited
  • the HTTP method to use
  • the headers, querystring and body data to include in the step
  • specific options to pass into the this step -- currently only the following options are supported:
    • request.retries - set to the number of retries to attempt on the step in question. Only works for non-mutating calls (e.g., GETs), defaults to 3.
    • request.body_format - can be set to json (default) or form. Will assemble the request body as either form-encoded or json and set the Content-type: header to application/x-www-form-urlencoded or application/json respectively.

Additionally, the YAML recipe file supports the ability to grab data from any part of any of the HTTP requests or responses in earlier steps in the recipe and insert them into later steps using a PHP-like construct. For example, say I have a 10-step recipe specified and in step 7 I need to POST some data that I received in step 3's reponse. I can include a construct like this in any part of the YAML file:

<%=steps[3].response.body["id"]%>

Assuming that step 3 did indeed contain a parameter called id in it's JSON response payload, this data will then be substituted in the specified part of step 10's request.

The insertion of variables anywhare in the recipe is done as shown:

<%!varname%>

Files can also be included into a recipe using a construct like this:

<%f/path/to/file.txt%>

This functionality is best illustrated with a complete recipe file as shown below.

--- #sample recipe
variables:
	name: brad
	favourite_colour: red
	age: 345
	drinks_beer: true
	age_and_colour: <%!age%>:<%!favourite_colour%>
	lorumfile: <%f./examples/lorum.txt%>

recipe:
	-  # step0
		name: Call Mockbin # see http://mockbin.org/docs#http-request
		httpmethod: POST
		# using the HTTP Basic auth process to see if the Authorization: header
		# is visible in the output file.
		URL: http://Aladdin:OpenSesame@mockbin.org/request/path/to/<%!name%>
		request:
			body:
				name: My name is <%!name%>
				age: I am <%!age%> years old.
				beer_drinker: <%!drinks_beer%>
				lorum_impsum: <%!lorumfile%>
			querystring:
				colors: blue, green,  <%!favourite_colour%>
			headers:
				X-info: <%!age_and_colour%>
	- 
		name: Call Mockbin with data from the previous call.
		httpmethod: GET
		options:
			request.retries: 10  #set retries to 10, overriding the default of 3.
		URL: http://mockbin.org/request/path/to/<%!name%>
		request:
			headers:
				date_header_from_previous_call: <%=steps[0].response.headers["Date"]%>
			body:
				http_response_code_from_previous_call: <%=steps[0].response.code%>

bprc provides the following features:

• robust logging support
• saving output files as raw HTTP (response only, or both request and response) or JSON
• SSL support (including the ability to ingore invalid server certificates)
• verbose and/or debug output
• HTTP request bodies formatted either as JSON or form-encoded
• Pretty-printed output

• improving error handling
• better test coverage
• Implementing --dry-run
• passing an entire payload, rather than just a single parameter, using a file include
• setting a recipe variable via cli and/or environment variable
• cleaning up the logging file format

Contributions are welcome! Please fork, make your changes, add tests to cover your work and then raise a pull request.