Repository navigation
SimStore performance improvements and fixes - #1157
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #1157 +/- ##
==========================================
- Coverage 81.29% 81.28% -0.01%
==========================================
Files 142 142
Lines 15643 15643
==========================================
- Hits 12717 12716 -1
- Misses 2926 2927 +1 ☔ View full report in Codecov by Sentry. |
|
This is ready for review and comment. Without review, I will merge this after at least 18 hours, no sooner than Sat 31 Aug 11:00 GMT (13:00 my local). |
… into uuid-indices
sroet
left a comment
sroet
left a comment
There was a problem hiding this comment.
One import can be dropped, feel free to ignore. LGTM otherwise
| ) | ||
| from .proxy import GenericLazyLoader | ||
| from .uuids import ( | ||
| has_uuid, get_uuid, set_uuid, encode_uuid, decode_uuid, encoded_uuid_re, |
There was a problem hiding this comment.
shouldn't has_uuid be dropped from the imports? (only reference is in commented out code)
Bit surprised linting did not catch this
There was a problem hiding this comment.
Turns out this was trickier than it appeared. In an early version of the file, has_uuid was defined in serialization_helpers.py. Then I split it off into uuids.py. But I allowed the transitive import.
Fixed that in 8bf1f31.
|
|
||
| universal_sql_meta = { | ||
| 'uuid': {'uuid': {'primary_key': True}}, | ||
| 'uuid': {'uuid': {'primary_key': True, 'index': True}}, |
This includes several performance improvements and minor fixes for SimStore. Quick manifest:
is_storage_stringandhas_storage_uuid, which use the fastClassIsSomethingmodel to cache classes with given attributes. The normalis_stringandhas_uuidmethods were taking a larger fraction of time than they should.This also helps clear up some of the external testing in ops-storage-notebooks which has been off for a while.