traverse-folders
v1.1.3
Published
Traverse nested folders and process each of the discovered files.
Downloads
1,027
Readme
Traverse nested folders and process each of the discovered files.
Prerequisites
This library assumes:
- You are using NodeJS 8+
Install
Add traverse-folders
as a dependency
:
npm i traverse-folders
Examples
A common use of traverse-folders
is to automatically load a nested hierarchy of functions into an index.js
file.
Consolidating API route controllers
For example, let's say you are writing an API server with the following folder hierarchy.
src/
api/
index.js
ping.js
version.js
/things
/createThing.js
/deleteThings.js
/getThing.js
/listThings.js
/updateThing.js
In src/api/index.js
you could put the following:
const path = require('path')
const traverse = require('traverse-folders')
const pathSeparator = new RegExp(path.sep, 'g')
const apis = {}
const base = __dirname
const processor = file => {
const name = file.slice(base.length + 1, -3).replace(pathSeparator, '_')
apis[name] = require(file)
}
traverse(base, processor)
module.exports = apis
Then when index.js
is first required it will load all the underlying code and expose
{
ping,
version,
things_createThing,
things_deleteThing,
things_getThing,
things_listThings,
things_updateThing,
}
with each api correctly linked to the underlying function.
Making a mockAPI that stays in sync with your real API
Now let's suppose, in your tests, you want to create a mock API that has the same function names, but instead of actually loading the functions, it associates each name with a stub
In test/utils/mockAPI.js
you could write
const path = require('path')
const { stub } = require('sinon')
const traverse = require('traverse-folders')
const pathSeparator = new RegExp(path.sep, 'g')
const mockApi = {}
const apiPath = 'src/api'
const processor = file => {
const name = file.slice(apiPath.length + 1, -3).replace(pathSeparator, '_')
names[name] = stub()
}
traverse(apiPath, processor)
module.exports = mockApi
Now your mockAPI can be used in unit tests in place of the real API, without referencing the real API at all. This can be important if your API controllers refer to Sequelize
models that might trigger an unwanted database connection. (Unit tests must not depend on external services.)
By customising the processor
function you can use traverse-folders
to auto-load Sequelize
models, ExpressJS
middleware, and all manner of other things.
Options
By default traverse
will ignore any index.js
files and only process files ending in '.js'
. To override this behaviour you can pass an options
object as the final parameter.
The defaults are:
{
ignore: 'index.js',
suffix: '.js'
}
So traverse(apiPath, processor, { ignore: 'index.jxs', suffix: '.jsx' })
will ensure that only files ending in .jsx
get loaded, but will ignore 'index.jsx'
.
Other ignore
options
ignore
can be a regular expression, sotraverse(apiPath, processor, { ignore: /-/ })
will ensure that only files ending in.js
get loaded, and will ignore any files with a dash in their name.ignore
can also be a function, sotraverse(apiPath, processor, { ignore: file => file === 'index.js' })
will ensure that only files ending in.js
get loaded, but will ignore'index.js'
.- if
ignore
is not a string, regular expression, or function, it's ignored.
Development
Branches
| Branch | Status | Coverage | Audit | Notes |
| ------ | ------ | -------- | ----- | ----- |
| develop
| | | | Work in progress |
| main
| | | | Latest stable release |
Prerequisites
- NodeJS, 15.1.0+ (I use
nvm
to manage Node versions —brew install nvm
.) You must use npm version 7.0.8 or better.
Test it
npm test
— runs the unit tests.npm run test:unit:cov
— runs the unit tests with coverage reporting.
Lint it
npm run lint
Contributing
Please see the contributing notes.
Thanks
- Thanks to
reallinfo
for the logo.