0.0.5 • Published 7 years ago

stagepipe v0.0.5

Weekly downloads
1
License
MIT
Repository
-
Last release
7 years ago

Stagepipe

Build Status

Stagepipe is an asynchronous flow control library for JavaScript applications.

Application code is organized into "stages" and "pipes", represented by a multidimensional array. You can think of stages as the rows of the array and pipes as the columns. The entire data structure is caled a "pipeline".

//    Pipe 1         Pipe 2
//      |              |
//      V              V
[
  [ getFirstName, getEmail      ], // <-- Stage 1
  [ upperCase,    validateEmail ], // <-- Stage 2
  [ greetUser,    sendPromo     ]  // <-- Stage 3
]

Each stage is executed in sequence until all stages are completed or an unhandled exception occurs. Subsequent stages are not started until all pipes in the current stage are completed. For example, suppose that getEmail in the above example returned a promise that resolves after making an HTTP call to a remote server. Stage 2 will not begin until that HTTP call comes back and the data is ready.

The previous example could be called from code like so:

const { stagepipe } = require('stagepipe')

const pipes = // Contents from the previous example
const pipeline = stagepipe(pipes)

// Invoke the pipeline with a user ID as input
pipeline(userId)
  .then(() => console.log('Done!'))
  .catch(err => console.error(err))

Pipelines always have exactly one input and produce exactly one output. Because of this, the first example is not valid and needs to be rewritten like so:

[
  [ split()                     ],
  [ getFirstName, getEmail      ],
  [ upperCase,    validateEmail ],
  [ greetUser,    sendPromo     ]
]

This pipeline takes its input and splits it into two pipes which produce the same result. In the next stage, both getFirstName and validateEmail will receive as input the data passed to the pipeline.

You can split streams into as many concurrent pipes as you'd like:

[
  [ split(4)                                               ],
  [ getFirstName, getEmail, getFavoriteColor, getBirthDate ]
]

Note that it is valid to terminate a pipeline with multiple pipes even though only one output will be produced. In the previous example, the output of the pipeline would be the output of the getFirstName function.

Stages

Each stage is an array of pipes to execute. If a stage consists of only one pipe you can omit the array notation for convenience:

[
  split(), // <-- There is only one pipe in this stage, don't need [ ]
  [ getFirstName, getEmail ] // <-- Multiple pipes, must wrap in [ ]!
]

Pipes

Pipe elements are just functions. If a function returns a promise, the current stage of the pipe will not be completed until the promise resolves. For any other return value (including undefined), the result will be synchronously passed to the next stage. In other words, if you're doing something async, you need to return a Promise.

Data is fed through each pipe top to bottom, left to right. By default, each function is assumed to have an arity of 1. If you need to call a function with more than one argument (thus merging two pipes from the previous stage), use the arity function.

[
  [ split()    ],
//   |\
//   |  \
//   |    \
  [ foo,   bar ],
//   |      |
  [ baz,   qux ],
//   |     /
//   |   /
//   |/
  arity(2, myFunc)
]

API

  • stagepipe(pipeline) Returns a function that will execute the given pipeline. The function takes as input the starting state of the pipeline and returns a promise that will resolve to the output value of the pipeline.
  • split([count=2]) Splits a pipe into count pipes.
  • pass() A noop function that passes its input out to the next stage
  • arity(count, fn) Joins two or more pipes into one by calling fn with outputs of the previous stage. If there are not enough outputs in the previous stage, undefined will be passed as the remaining argument values.
  • print([count=1]) A debugging utility. Prints the state of count pipes and passes the data along to the next stage.

License

MIT

0.0.5

7 years ago

0.0.4

7 years ago

0.0.3

7 years ago

0.0.2

7 years ago

0.0.1

7 years ago