Fife

Fife is a multiple IO pipelining tool, originally designed for file uploading through HTTP, but can be used in any other cases.

This gem is in early development stage, and the interface could be changed enormously in future releases.

Add this line to your application's Gemfile:

gem 'fife'

And then execute:

$ bundle

Or install it yourself as:

$ gem install fife

Create a Fife instance by calling Fife(io_ary), then chain as many pipe calls as you want.

io_ary = [
  File.new('/path/to/my/file'),
  StringIO.new('Hello world'),
  Tempfile.new('temp')
]

Fife(io_ary).pipe(-> io {io})
            .pipe(-> io {[io, io]})
            .pipe(:close)

This method returns a Fife::Pipe instance for chaining.

This method expect 0 or more IO-like objects as arguments. Alternatively it can take an array of IO-like objects

This method returns a new Fife::Pipe instance.

It has 2 forms of usage:

1. pass an operation as the argument Fife(io_ary).pipe(->io{io})
2. pass the underscored name of an operation class, and 0 or more arguments as the constructor arguments of that class. Fife(io_ary).pipe(:store, my_storage) see Define custom operations for detail.

An operation is an object that responds to call(io), and returns 0 or more IO-like objects. A lambda is often used as an operation, but an operation can be of any type.

Currently, Fife ships with 4 operations: :noop, :name, :store and :close

• noop Performs no operation on the IO, and returns the IO directly.
• close Closes the IO if not already closed.
• store Stores the content of the IO.
• name Gives the IO a name according to the per-io naming strategy you specified. You can retrieve it's name by calling #name on the IO instance.

name = Fife::Operations::Name.new(->io { 'some_name' })
Fife(io_ary).pipe(name)

# Or
Fife(io_ary).pipe(:name, ->io { 'some_name' })

If you feel lambdas are not enough for your job, you can easily define your own operations.

class MyOperation
  def initialize(arg1, arg2)
    # Initialize the Operation
  end

  def call(io)
    # Handle the io and return some IO instances
  end
end

# Register it
Fife::Operations.register(:my_op, MyOperation)

Then you can use it like

Fife(io_ary).pipe(:my_op, 1, 2)

To leverage the :store operation, you need a storage.

A storage is just an object that responds to store(io).

Currently, Fife ships with 2 storage classes:

• Fife::Storage::Null Simply does nothing
• Fife::Storage::Sftp Streams the IO to a remote file via SFTP. This storage should be used with :name operation because the remote filename depends on the name of the IO. (I know it's nasty, but it's still under developing)

storage = Fife::Storage::Sftp.new do
  host 'localhost'
  user 'me'
  remote_dir '/path/to/my/remote/dir'
  naming -> io {"lorem_ipsum_#{rand}"}
  ssh_options password: 's3cret'
end
Fife(io_ary).pipe(:store, storage)

By default, Fife ignores all failures and keeps on executing further operations on succeeded operations.
This behavior can be changed by passing abort_on_fail: true to Kernel#Fife:

Fife(io_ary, abort_on_fail: true)

When abort_on_fail is turned on, if some operation raises error, no further operation will be executed, and an Fife::Pipe::Aborted will be raised. You can get the real causes by accessing Fife::Pipe::Aborted#causes

After checking out the repo, run bin/setup to install dependencies. Then, run rake false to run the tests. You can also run bin/console for an interactive prompt that will allow you to experiment.

To install this gem onto your local machine, run bundle exec rake install. To release a new version, update the version number in version.rb, and then run bundle exec rake release, which will create a git tag for the version, push git commits and tags, and push the .gem file to rubygems.org.

Bug reports and pull requests are welcome on GitHub at https://github.com/aetherus/fife.