Parse BDD-style tests (Mocha, Jasmine) to generate documentation
Acquit's parse()
function takes in mocha tests as a string, and outputs
a list of "blocks", which are either describe
or it
calls. A describe
call contains a list of "blocks", whereas an it
call contains the actual
code
in order to provide an effective, well-tested example.
var contents = `
/**
* A Model is a convenience wrapper around objects stored in a
* collection
*/
describe('Model', function() {
/**
* Model **should** be able to save
**/
it('can save', function() {
assert.ok(1);
});
it('can save with a parameter', function() {
});
});
`;
var ret = acquit.parse(contents);
// One top-level block: describe('Model')
assert.equal(1, ret.length);
assert.equal('describe', ret[0].type);
assert.equal(1, ret[0].comments.length);
assert.ok(ret[0].comments[0].indexOf('Model') != -1);
// Top-level block contains the `it('can save')` block, which contains
// the code
assert.equal(2, ret[0].blocks.length);
assert.equal('it', ret[0].blocks[0].type);
assert.equal(1, ret[0].blocks[0].comments.length);
assert.ok(ret[0].blocks[0].code.indexOf('assert.ok(1)') !== -1);
assert.equal('can save', ret[0].blocks[0].contents);
assert.equal('it', ret[0].blocks[1].type);
assert.equal('can save with a parameter', ret[0].blocks[1].contents);
assert.equal(0, ret[0].blocks[1].comments.length);
Acquit can also take a callback as second parameter. This callback gets executed on every block and can transform the block however you want.
var contents = `
describe('ES6', function() {
// ES6 has a yield keyword
it(\'should be able to yield\', function() {
// some code
});
});
`;
var cb = function(block) {
block.code = 'return value from callback';
};
var ret = acquit.parse(contents, cb);
assert.equal('return value from callback', ret[0].blocks[0].code);
Want to chain multiple callbacks together and/or develop re-usable
plugins? acquit.transform()
allows you to add transformations that
are executed each time you call .parse()
.
Transform functions are executed in order before the callback
function passed to .parse()
.
var contents = `
describe('ES6', function() {
// ES6 has a yield keyword
it('should be able to yield', function() {
// some code
});
});
`;
var cb = function(block) {
block.code = 'my transformed code';
};
acquit.transform(cb);
var ret = acquit.parse(contents);
assert.equal('my transformed code', ret[0].blocks[0].code);
acquit.removeAllTransforms();
Acquit can also parse ES6 code
var contents = `
describe('ES6', function() {
// ES6 has a yield keyword
it('should be able to yield', function() {
co(function*() {
yield 1;
})();
});
});
`;
var ret = acquit.parse(contents);
assert.equal(1, ret.length);
assert.equal('describe', ret[0].type);
assert.equal(0, ret[0].comments.length);
assert.equal(1, ret[0].blocks.length);
assert.equal('it', ret[0].blocks[0].type);
assert.equal(1, ret[0].blocks[0].comments.length);
assert.ok(ret[0].blocks[0].code);
Acquit can parse Mocha alias:
context
=describe
specify
=it
var contents = `
context('Mocha aliases', function() {
specify('should be parsed', function() {
assert.equal(1, 1);
});
});
`;
var ret = acquit.parse(contents);
assert.equal(1, ret.length);
assert.equal('describe', ret[0].type);
assert.equal(0, ret[0].comments.length);
assert.equal(1, ret[0].blocks.length);
assert.equal('it', ret[0].blocks[0].type);
assert.equal(0, ret[0].blocks[0].comments.length);
assert.ok(ret[0].blocks[0].code);
trimEachLine()
is a helper function for trimming whitespace and asterisks
from JSdoc-style comments
var str = ` * This comment looks like a
* parsed JSdoc-style comment`;
assert.equal(acquit.trimEachLine(str), 'This comment looks like a\n' +
'parsed JSdoc-style comment');
You don't have to use JSdoc-style comments: trimEachLine()
also trims
leading and trailing whitespace.
var str = `This comment looks like a
* parsed JSdoc-style comment`;
assert.equal(acquit.trimEachLine(str), 'This comment looks like a\n' +
'parsed JSdoc-style comment');
You can use the .output()
function to attach output processors,
which transform the output from acquit.parse()
before you get it.
var contents = `
describe("My feature", function() {
it("works", function() {
// some code
});
});
`;
acquit.output(function(res) {
return `
# ${res[0].contents}
## ${res[0].blocks[0].contents}
`;
});
var res = acquit.parse(contents);
assert.equal(res.trim(), `
# My feature
## works
`.trim());
acquit.removeAllTransforms();
You can also use acquit as a constructor, in case you need multiple sets of transforms.
acquit.transform(function(block) {});
assert.equal(acquit.getTransforms().length, 1);
var parser = acquit();
assert.equal(parser.getTransforms().length, 1);
parser.transform(function(block) {});
assert.equal(parser.getTransforms().length, 2);
parser.removeAllTransforms();
assert.equal(parser.getTransforms().length, 0);
assert.equal(acquit.getTransforms().length, 1);
assert.equal(parser.parse('describe("test", function() {});').length,
1);
parser.output(function(res) {
return 'myFakeOutput';
});
assert.equal(parser.parse('describe("test", function() {});'),
'myFakeOutput');
parser.removeAllOutputProcessors();
assert.equal(parser.parse('describe("test", function() {});').length,
1);