Difference between revisions of "TaskFlow/Persistence/Objects, not Collections"

Revision as of 08:11, 31 March 2014

What

blueprint: https://blueprints.launchpad.net/taskflow/+spec/lb-not-collections

It is proposed to stop treating logbook object as collection of smaller objects: LogBook and FlowDetails should be made simple data-transfer objects, not containers for FlowDetails and AtomDetails respectively. Corresponding API should be moved to backends.

Why

• general sanity
• optimization for certain tasks (monitoring, UI)
  • while perserving good performance for engines
• better cache semantics
  • backend should know that better

TBD: more descriptive text here.

Implementation Notes

• remove __iter__ and find from LogBook and FlowDetail
• add methods to Connection instead:
  • instead of FlowDetail methods:
    • get_atom_by_uuid(uuid) -> atom_detail
    • get_atoms_for_flow(flow_uuid) -> [atom_detail]
    • get_atoms_by_name(flow_uuid, name) -> [atom_detail] ^[1]
    • save_atom(atom_detail, ...) -> atom_detail ^[2][3]
  • instead of LogBook methods:
    • get_flow_by_uuid(uuid) -> flow_detail
    • get_flows_for_book(logbook_uuid) -> [flow_detail]
    • save_flow(flow_uuid)

When: 0.3 series

How:

• add methods to impl_memory.Connection
• modify storage.Storage, make engine and storage tests run
• modify other backends one by one
• modify backends' base class

Notes

1. ↑ backend does not guarantee that names will be unique in the flow, API must reflect that
2. ↑ returns updated atom detail
3. ↑ may need O_CREATE-like option (raise exception when atom already exists)