NOTE: Brewery was replaced by Bubbles
Flow based data processing with support for variety of data sources, data targets.
Data are treated as file-like objects where structured data is being passed instead of bytes.
Read more about stores.
Processing network is described as Stream composed of connected processing nodes. Besides traditional network construction there is an option to use higher order messaging to construct a stream:
audit = Stream().fork() audit.sample(1000) audit.audit() audit.formatted_printer() audit.run()
Brewery uses metadata extensively – they describe the structure of processed data flow.
brewery is a command line tool for running data processing streams, for example:
cat data.csv | brewery pipe audit pretty_printer