streamtools

Streamtools is a creative tool for working with streams of data. It provides a vocabulary of data processing operations, called blocks, that can be connected to create online data processing systems without the need for programming or complicated infrastructure.

Streamtools is built upon a few core principles:

• Working with data should be a responsive, exploratory practice. Streamtools allows you to immediately ask questions of the data as it flows through the system (see: Creative Abduction).
• In the real world, the character of your data is constantly changing. We designed Streamtools not only to reflect how your data is changing but to let you work with that change (see: Non-Stationarity).
• Working with data should not require complex engineering. Streamtools provides a visual interface and an expressive toolset for working with streams of data.

Streamtools is an open source project written in Go and is intended to be used with streams of JSON.

getting started

1. Find a computer to play with. It needs to be Linux or OSX.
2. Download the latest release. You need st-linux if you're on linux or st-darwin if you're on osx.
3. In a terminal, change directory to wherever you downloaded the file.
4. Run chmod +x st-linux if you're on linux or chmod +x st-darwin on osx. This makes the file you downloaded executeable.
5. Now launch streamtools by typing ./st-linux if you're on linux or ./st-darwin if you're on osx. Your terminal should say starting stream tools on port 7070.
6. To find the UI visit http://localhost:7070 in a browser. If you're not running streamtools locally you need some way of accessing port 7070 on your remote box.
7. Go through our Hello World pattern!
8. Look through the rest of our patterns for inspiration and guidance.

Good luck!

health warning

Note that streamtools is very new! This means we're developing it very rapidly, and some things aren't going to work. If you find a bug please do let us know! And, if you think of something you'd like to see, please do request it! Both of these things can be done on our issues page.

contributing

As always: pull requests are welcome! Our focus at the moment (Spring '14) is to get a fully functioning system together that we can demonstrate. Therefore new blocks are likely to be merged in with more energy than large re-writes of the back-end. Having said that, there is plenty that can and should be done behind the scenes, and we always have an eye to the next major re-factor.

If you'd like to make a new block the best place to start is to look at the skeleton blocks. We have a skeleton state block which demonstrates how to lay out a block that maintains a state. We also have a skeleton transfer block which demonstrates how to lay out a block that emits zero or one messages upon reciept of an inbound message.