zc.customdoctests 0.1.0

Custom doctest parsers

zc.customdoctests provides a little bit of help with creating custom doctest parsers that work pretty muct like regular doctests, but that use an alternate means of evaluating examples. To use it, you call zc.customdoctests.DocTestParser and pass any of the following options:

ps1

The first-line prompt, which defaultd to '>>>'.

This must be a regular expression that matches exact;y 3 characters.

(Note that you can’t override the second-line prompt.)

comment_prefix

The comment prefix regular expression, which defaults to ‘#’.

transform

A function used to transform example source, which defaults to a no-operation function.

The js module provides support for using JavaScript in doctests using python-spidermonkey. It provides some examples of defining custom doctest parsers.

Javascript and Python-Spidermonkey support

This file shows some examples of using spidermonkey APIs in doctests.

To wire this up, you’d use something like:

import doctest, zc.customdoctests.js

test_suite = doctest.DocTestSuite(
    parser = zc.customdoctests.js.parser,
    setUp = zc.customdoctests.js.spidermonkeySetUp,
    )

Or, with manuel:

test_suite = manuel.testing.TestSuite(
        manuel.doctest.Manuel(parser=zc.customdoctests.js.parser) +
        manuel.doctest.Manuel(parser=zc.customdoctests.js.eq_parser) +
        manuel.doctest.Manuel() +
        manuel.capture.Manuel(),
        'spidermonkey.txt',
        setUp=zc.customdoctests.js.spidermonkeySetUp)
        )

Note that zc.customdoctests doesn’t require spidermonkey, so you need to install spidermonkey seperately if you want to use it.

An advantage of using manuel is that you can use multiple parsers in the same document. In the example, above, 2 javascript example syntaxes (described below) as well as the standard doctest syntax are supported. This document is run with manuel to allow all 3 syntaxes.

For the rest of this document, we’ll show examples of JavaScript doctests as well as helper APIs used to support JavaScript and to integrate JavaScript and Python.

Javascript doctests use a “js>” prompt (as used in rhino and the spidermonkey interpreter):

js> 2 +
... 'Hi world' // doctest: +ELLIPSIS
u'2Hi...

Assignments return values. This can generate annoying output in doctests:

js> ob = {a: 1, b: 2}
[object Object]

If you’re using manuel, you can avoid this by using js!:

js! x = 3

which suppresses expression values.

load and print functions (similar to those found in rhino) are provided. For example, given a javascript file, double.js:

function double (x) {
    return x*2;
}

We can load the file:

js> load('double.js')
js> double(10)
20

We can print values:

js> print('Hi')
Hi

A python object provides access to the open function and the os module:

js> python.os.path.exists('double.js')
True

js! f = python.open('double.js')
js> print(f.read())
function double (x) {
    return x*2;
}
<BLANKLINE>

js> f.close()

If you’re using manuel, you can intermix Python and and JavaScript examples and there are a number of APIs to facilitate using Python and JavaScript together.

There’s an add_js_global function to copy data from Python:

>>> add_js_global('y', 1)

js> y
1

There’s also a js object that provides attribute access to js globals:

>>> js.x
3

>>> js.z = 4

js> z
4

You can also call this to run JS code without returning the resulting value:

>>> js('a = x + y')

js> a
4