pipeline

STATUS: Pre-alpha, in design and prototyping phase.

Description will come later. See tests for examples.

TODO

Move examples from test to examples
Replace malli with spec
Add action and transformation construction functions
Restructure internal context
Make better example with sql and http
Fake sql and http in test
allow both simple and qualified keywords as input paths
allow both simple and qualified keywords as output path
Add missing accessors
add pipeline binding and get-binding accessor
add step binding
param "spec" (required keys and preds)
print/return pipeline (static structure)
print/return pipeline run (results of steps)
print-call, not only print-failed-call (for putting in logs)
validate pipeline w/ explain
repl support for platform objects, e.g. data sources
option can contain "spec" validation function (malli, spec, schema, ...)
store only changed "keys" in trace
validate dependencies (all should resolve)
fork/join, paralellisation when executing
validate that step names are unique
decide if it's ok or not to overwrite keys
validate that if a key is overwritten then it's in the same path if there are forks. No race conditions allowed.