Reimplement legacy states in data seeding with shep
Problem
Presently main thread state of account import lifetime is handled by explicitly moving and linking records between directories.
With the introduction of the shep
state backend, the full lifetime can be synchronized using this instead.
This will simultaneously remove the need for the cic_seeding.legacy
module.
Solution
Re-implement using shep
, e.g. with the following registered states:
original location | state replacement | alias of | description |
---|---|---|---|
src | NEW | - | pre-import state of user (e.g. output of create_import_users.py) |
new | ADDRESS | - | new address for user has been created and is known |
- | REGISTRY | - | new address for user has been registered on network |
meta, symlink to new | META | - | Metadata has been processed |
custom, symlink to new | - | - | (Merge with META) |
- | BALANCE | - | Initial balance tx has been sent |
- | REGISTERED | ADDRESS v REGISTRY |
Account registration is complete |
In the above, meta and balance executors could poll for match on the REGISTERED
state before continuing with META
and BALANCE
. This requires a feature addition for shep
enabling listings for partial state matches.
Edited by Louis Holbrook