Difference between revisions of "Staff Priorities - Masha"
From SCECpedia
Jump to navigationJump to searchLine 3: | Line 3: | ||
=== Yesterday === | === Yesterday === | ||
* Trac ticket #370: Replace existing __del__ methods of CSEP classes with atexit.register() handlers: | * Trac ticket #370: Replace existing __del__ methods of CSEP classes with atexit.register() handlers: | ||
− | ** | + | ** Committed code changes to SVN: use weakref.ref for circular dependencies and keep __del__() method for CSEP classes |
+ | * miniCSEP notebook to invoke T and W evaluation tests for CA one-day forecasts | ||
+ | |||
=== Today === | === Today === | ||
− | * Trac ticket #370: Replace existing __del__ methods of CSEP classes with atexit.register() handlers | + | * 9am meeting with Max on Bayesian ensemble modeling |
+ | * Nightly build failed on csep-cert due to committed code changes, investigating | ||
+ | ** Trac ticket #370: Replace existing __del__ methods of CSEP classes with atexit.register() handlers | ||
* Trac ticket #371: Apply 2-step optimization for locating forecast bin by RELM evaluaiton tests | * Trac ticket #371: Apply 2-step optimization for locating forecast bin by RELM evaluaiton tests | ||
'''Blocked''': No | '''Blocked''': No | ||
− | '''Followups''': | + | '''Followups''': |
== TODO == | == TODO == |
Revision as of 19:38, 2 February 2016
Current Activities
Yesterday
- Trac ticket #370: Replace existing __del__ methods of CSEP classes with atexit.register() handlers:
- Committed code changes to SVN: use weakref.ref for circular dependencies and keep __del__() method for CSEP classes
- miniCSEP notebook to invoke T and W evaluation tests for CA one-day forecasts
Today
- 9am meeting with Max on Bayesian ensemble modeling
- Nightly build failed on csep-cert due to committed code changes, investigating
- Trac ticket #370: Replace existing __del__ methods of CSEP classes with atexit.register() handlers
- Trac ticket #371: Apply 2-step optimization for locating forecast bin by RELM evaluaiton tests
Blocked: No
Followups:
TODO
- Trac ticket #353: Optimize RELM N-test by avoiding event to bin look up when masking bit is set for the whole forecast