pygeodyn merge requestshttps://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests2018-05-03T14:40:24+02:00https://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/1Reimplement forecast context2018-05-03T14:40:24+02:00Loic HuderReimplement forecast contextReimplementation was successful: ForecastContext is now a Python object with getters/setters for max degrees. The forecasts work fine with the new syntax.Reimplementation was successful: ForecastContext is now a Python object with getters/setters for max degrees. The forecasts work fine with the new syntax.Code ported in Pythonhttps://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/2Resolve "CI with git lab"2018-05-14T12:27:08+02:00Franck ThollardResolve "CI with git lab"Closes #3Closes #3https://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/3Resolve "CI: adding fun icon to README.md"2018-05-14T21:42:23+02:00Franck ThollardResolve "CI: adding fun icon to README.md"Closes #12Closes #12https://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/4Python port2018-05-16T11:00:59+02:00Loic HuderPython portpar_forecast working on the semantic side. Merge to impact the relevant files in master before test writings. Closes #6 .par_forecast working on the semantic side. Merge to impact the relevant files in master before test writings. Closes #6 .Code ported in Pythonhttps://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/5Python port2018-05-18T14:52:16+02:00Loic HuderPython portWorking version of the forecasts in par_forecast with proper init.
Implements #18 and #16.Working version of the forecasts in par_forecast with proper init.
Implements #18 and #16.Basic visualization toolsLoic HuderLoic Huderhttps://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/6Python port: par_forecast working correctly2018-05-22T10:23:22+02:00Loic HuderPython port: par_forecast working correctlypar_forecasts runs fine and gives results that are scientifically relevant (forecasts and analyses are normally computed).par_forecasts runs fine and gives results that are scientifically relevant (forecasts and analyses are normally computed).Code ported in PythonLoic HuderLoic Huderhttps://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/7Init reimplementation + file organisation2018-05-23T08:20:47+02:00Loic HuderInit reimplementation + file organisationPython implementation of init method tested with asserts in ForecastContext (Closes #4).
Reorganisation of files in relevant submodules (algo for computation functions, inout for reading/writing and the context)Python implementation of init method tested with asserts in ForecastContext (Closes #4).
Reorganisation of files in relevant submodules (algo for computation functions, inout for reading/writing and the context)Code ported in PythonLoic HuderLoic Huderhttps://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/8Impact recent work on master in test-forecastcontext2018-05-23T08:23:22+02:00Loic HuderImpact recent work on master in test-forecastcontextImportant modifications in master require a merge. Moreover, no conflicts are expected as work did not start yet on test-forecastcontext.Important modifications in master require a merge. Moreover, no conflicts are expected as work did not start yet on test-forecastcontext.CI setupLoic HuderLoic Huderhttps://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/9Implementation of Forecast in Python2018-05-23T15:41:25+02:00Loic HuderImplementation of Forecast in PythonImplements #7. Test pending.Implements #7. Test pending.Code ported in PythonLoic HuderLoic Huderhttps://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/10Fixed analysis in Python.2018-05-24T17:59:57+02:00Loic HuderFixed analysis in Python.Analysis now runs correctly in Python: master branch should now run scientifically correct algorithms with par_forecast.Analysis now runs correctly in Python: master branch should now run scientifically correct algorithms with par_forecast.Code ported in PythonLoic HuderLoic Huderhttps://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/11Merge important code modifs from master to python branch2018-05-29T09:05:01+02:00Loic HuderMerge important code modifs from master to python branchRecent work in master involved the correction of the algorithm and the implementation of several convenience methods that must be present in the python branch.Recent work in master involved the correction of the algorithm and the implementation of several convenience methods that must be present in the python branch.Loic HuderLoic Huderhttps://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/12Structural changes in the code.2018-05-30T15:46:58+02:00Loic HuderStructural changes in the code.Apply the structural changes to master branch, namely:
* The use of CoreState class as a container for core state quantities (implements #19)
* Depreciation of ComputationContext (closes #9) in favour of ComputationConfig that only store...Apply the structural changes to master branch, namely:
* The use of CoreState class as a container for core state quantities (implements #19)
* Depreciation of ComputationContext (closes #9) in favour of ComputationConfig that only stores relevant parameters of the config files (no matrices)
* The use of Forecasts and Analyses classes that inherit from ComputationStep.
* Easy choice between Python and Fortran implementation for each step
* Python implementation is now the default one and wrappers were implemented (build_legacy...) to adapt to Fortran syntax rather than the other way around.Code ported in PythonLoic HuderLoic Huderhttps://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/13Resolve "Adding sphinx comment"2018-06-11T23:07:23+02:00Franck ThollardResolve "Adding sphinx comment"Closes #27Closes #27https://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/14Resolve "Adding sphinx comment"2018-06-12T21:18:34+02:00Franck ThollardResolve "Adding sphinx comment"Closes #27Closes #27https://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/15Resolve "Parallel forecasts"2018-06-20T14:23:48+02:00Loic HuderResolve "Parallel forecasts"Closes #26 and #30.Closes #26 and #30.https://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/16Implementation of HDF5 output2018-06-27T15:50:31+02:00Loic HuderImplementation of HDF5 output* Implements #33 (hdf5 saving but while keeping ASCII saving possible)* Implements #33 (hdf5 saving but while keeping ASCII saving possible)0.2Loic HuderLoic Huderhttps://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/17CoreState changes2018-06-28T10:58:57+02:00Loic HuderCoreState changes* CoreState object is now more versatile: can be of any dimensions.
* In consequence, updated tests and added a checking decorator.* CoreState object is now more versatile: can be of any dimensions.
* In consequence, updated tests and added a checking decorator.Loic HuderLoic Huderhttps://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/18Resolve "Parallel analysis"2018-06-29T15:17:22+02:00Loic HuderResolve "Parallel analysis"Implemented #29 in augmented_Kalman_filter_with_parallel_analysis. Previous version with sequential analysis is still available for checking purposes as the algorithm is slightly different.
Tests show that:
- Scientific results seem ...Implemented #29 in augmented_Kalman_filter_with_parallel_analysis. Previous version with sequential analysis is still available for checking purposes as the algorithm is slightly different.
Tests show that:
- Scientific results seem **consistent between the two methods.**
- Due to change of algorithm, parallel_analysis is in fact **slower !**: it adds N matrix inversions and N computations of Ab with N: number of realisations.0.2Loic HuderLoic Huderhttps://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/19Scientific development2018-07-03T10:25:29+02:00Loic HuderScientific developmentWorking on v0.3: lots of structural rearranging to be able to switch algorithms.
Merge when scientific tests are OK.Working on v0.3: lots of structural rearranging to be able to switch algorithms.
Merge when scientific tests are OK.0.3: Oopservation updateLoic HuderLoic Huderhttps://gricad-gitlab.univ-grenoble-alpes.fr/Geodynamo/pygeodyn/-/merge_requests/20Scientific developement2018-07-04T16:06:36+02:00Loic HuderScientific developementCorDE algo fully implemented but scientific results are not ok.
Everything runs fine so the work can continue on the master branch.CorDE algo fully implemented but scientific results are not ok.
Everything runs fine so the work can continue on the master branch.0.3: Oopservation updateLoic HuderLoic Huder