Difference between revisions of "Staff Priorities - Masha"

From SCECpedia
Jump to navigationJump to search
Line 2: Line 2:
  
 
=== Yesterday ===
 
=== Yesterday ===
* CSEP V15.10.0 release: could not install on csep-op due to the nightly processing still running
+
* CSEP V15.10.0 release:  
** Tested NZ image successfully
+
** Installed on csep-op, ran all unit and acceptance tests successfully
* Updated/checked David Jackson's email on CSEP lists
+
** John brought up Japan image for testing: scipy files are corrupted
 
* Read papers for the CSEP status meeting on Wednesday
 
* Read papers for the CSEP status meeting on Wednesday
* 'pandas' data analysis Python package for optimized file IO functionality: problems reading forecast from gzip archive because of storing original filename as part of the archive
+
* 'pandas' data analysis Python package for optimized file IO functionality:
 +
** Posted question to user group if can load data directly from archive with Pandas when filename is stored within archive
 
* Re-run test case using best available catalog for Canterbury experiment with zero days of delay for providing slip models to the forecasts:
 
* Re-run test case using best available catalog for Canterbury experiment with zero days of delay for providing slip models to the forecasts:
 
** Generating one-month and one-day forecasts
 
** Generating one-month and one-day forecasts
Line 12: Line 13:
 
=== Today ===
 
=== Today ===
 
* CSEP V15.10.0:
 
* CSEP V15.10.0:
** Install and test on csep-op
+
** csep-op: error from SVN about missing pristine file, seems to be bug in SVN, checked out fresh working copy, reprocessing failed Dispatcher
** Test EU, Japan and miniCSEP VM images
+
** Test Japan and miniCSEP VM images
 
* Read papers for the CSEP status meeting on Wednesday
 
* Read papers for the CSEP status meeting on Wednesday
* 'pandas' data analysis Python package for reading forecast from gzip archive file:
+
* Trac ticket #366: Adopt Pandas Python package for reading forecasts files in
** tested with iPython that adding data extraction from archive adds only 100ms to 500ms to read data with Pandas vs. 3.5s to load data with numpy.loadtxt
 
** post question to user group if can load data directly from archive with Pandas when filename is stored within archive
 
 
* Re-run test case using best available catalog for Canterbury experiment with zero days of delay for providing slip models to the forecasts:
 
* Re-run test case using best available catalog for Canterbury experiment with zero days of delay for providing slip models to the forecasts:
 
** Generating one-month and one-day forecasts
 
** Generating one-month and one-day forecasts

Revision as of 17:58, 14 October 2015

Current Activities

Yesterday

  • CSEP V15.10.0 release:
    • Installed on csep-op, ran all unit and acceptance tests successfully
    • John brought up Japan image for testing: scipy files are corrupted
  • Read papers for the CSEP status meeting on Wednesday
  • 'pandas' data analysis Python package for optimized file IO functionality:
    • Posted question to user group if can load data directly from archive with Pandas when filename is stored within archive
  • Re-run test case using best available catalog for Canterbury experiment with zero days of delay for providing slip models to the forecasts:
    • Generating one-month and one-day forecasts

Today

  • CSEP V15.10.0:
    • csep-op: error from SVN about missing pristine file, seems to be bug in SVN, checked out fresh working copy, reprocessing failed Dispatcher
    • Test Japan and miniCSEP VM images
  • Read papers for the CSEP status meeting on Wednesday
  • Trac ticket #366: Adopt Pandas Python package for reading forecasts files in
  • Re-run test case using best available catalog for Canterbury experiment with zero days of delay for providing slip models to the forecasts:
    • Generating one-month and one-day forecasts

Blocked: No

Followups:

  • With John Y.:
    • csep-op: Added my email to recipients list for Oceanic Transform Faults status email, got email sent to my account, but not to the mailing list.

TODO

  • Trac ticket #353: Optimize RELM N-test by avoiding event to bin look up when masking bit is set for the whole forecast

CSEP

EEW

Responsiblities