jessitron / aqueductron Goto Github PK

The current endpiece split is sufficient for all cases, but it would be more efficient if we could selectively send the data down exactly one pipe.

Implement a partition endpiece that accepts an ordered hash of [Predicate, Pipe] and a default Pipe. Send the data down the first Pipe whose Predicate is true, or the default if none are.

I moved the book-in-stock example files around. It'll take some tweaking for stock_stats.rb to run correctly again.

Make stock_stats.rb run, and then put a readme or something in its example directory describing the command line syntax for running it.

the map functionality should be available as "through", "map", or "transform."
Filtering should be available as "filter" or "keeping"

Another endpiece: partition_dynamic(?) it takes two functions: data => category and category => Duct. The ducts are created as necessary as data comes in.

Use case: count characters in a string, without knowing what characters will appear in it.

If you can think of a more descriptive name, fabulous.

Rename Pipeline to Aqueductron and Pipe to Duct. Unless you can think of even better naming.

Result needs to provide its contents dynamically.
This is especially important when ducts can be created dynamically (#5)
This requires more thought. It needs to supply the path as well as the result.

Perhaps it implements each, supplying a tuple?