* Improve alignment of live logbook and historical logbook models
- Make EventAsRow as NamedType which is better aligned with
sqlalchemy Row
- Use getitem to fetch results for both Row and EventAsRow
since its an order of magnitude faster fetching sqlalchemy
Row object values.
* final
* fix
* unused
* fix more tests
* cleanup
* reduce
* tweak
Avoid addtional timestamp conversion to set state
Since we already have the timestamp, we can pass it on to the State
object and avoid the additional timestamp conversion which can be as
much as 30% of the state write runtime.
Since datetime objects are limited to microsecond precision, we need
to adjust some tests to account for the additional precision that we
will now be able to get in the database
Co-authored-by: Sid <27780930+autinerd@users.noreply.github.com>
Co-authored-by: Marc Mueller <30130371+cdce8p@users.noreply.github.com>
Co-authored-by: J. Nick Koston <nick@koston.org>
* Ensure logbook still responds if describe event throws
If describe fails, the logbook stream should not collapse
* Ensure logbook still responds if describe event throws
If describe fails, the logbook stream should not collapse
* Remove legacy context lookup implemention from logbook
This object can now be replaced with a simple dict
* Remove legacy context lookup implemention from logbook
This object can now be replaced with a simple dict
* scope
* fix order issue
* Force live logbook to send an empty message to indicate no results
Since the sync task can take a while if the recorder is
busy, the logbook will appear to hang if we do not send
the first partial message even if its empty.
This work is in preparation for a higher database
commit interval where this issue is most obvious.
The historical only path did not have this issue because
it never had to wait for the db sync.
* update tests
The CI sets the timezone to US/Pacific and the logbook
uses start_of_local_day when called without a time.
We now call the logbook api with a specific time to
avoid them being out of sync since the test would
fail at CET 8:55am on Mon Nov 7th 2022 (and probably
other dates)
epenet
J. Nick Koston
Joost Lekkerkerker
Erik Montnemery
Marc Mueller
Aarni Koskela
Ville Skyttä
Franck Nijhof
Jan Bouwhuis