melvin

file mutation stream

what

a duplex stream interface to mutating directories of files

why

this module is specifically optimized for use-cases where the fully buffered file contents are necessary and/or you will be doing lots of file renaming. If these things are not the case, you might try mutiny instead.

example

var path = require('path')

var through = require('through2')
var melvin = require('melvin')

var fileStream = melvin(process.cwd())

fileStream.pipe(through.obj(toUpperCase)).pipe(fileStream)

function toUpperCase (file, x, next) {
  var basename = path.basename(file.filename)

  file.filename = path.resolve(
    process.cwd(),
    '..',
    'transformed',
    basename.toUpperCase()
  )

  file.data = file.data.toString().toUpperCase()

  this.push(file)

  next()
}

will take any file in the CWD and put it in ../transformed/ with its filename and contents all upper-cased.

API

melvin([sourceDir, ] [dirFilter, ] [fileFilter]) -> Stream

• sourceDir is the directory to read files from. If not provided, no directories will be read and the return value will be simply a Writable stream. Otherwise, a Duplex stream will be returned.
• dirFilter is a function that will be called on every encountered directory to determine whether or not to descend into it. If not provided, every directory will be recursed into.
• fileFilter is a function that will be called with every file encountered to determine whether or not to read the file and emit its contents. If not provided, every file will be read.

Notes

• The filter functions must operate synchronously.

File Objects

melvin emits File Objects on read, and accepts them on write. They look like this:

{
  filename: String,
  data: Buffer | String
}

• filename is the full path to the file
• data is a Buffer or String of the contents of said file. (Will always be emitted as a Buffer)

events

melvin emits written events with filename whenever it writes a file to disk.

license

MIT