GitLab

This commit will allow the basic CTDAS code to run, using baseclasses obs.py, observationoperator.py, pipeline.py, and optimizer.py.

The observations created now are a simple list of 10 values, which get added a random white noise by the obs operator.

Next challenge is to also bring in statevector.py from the baseclass as well, without dependencies on external files.

We changed the master project to once again run the random observation operator in python 3, using the gridded statevector and the ObsPack GlobalViewplus2 on capegrim. Next to do is to make this project also run without the specific parts but instead with baseclasses, and to program something that allows different templates on different machines.

The templates for jobs and rc-files used by start_ctdas.sh are now in a separate templates folder. This allows each user to modify their templates for running if needed, and to add templates for their own purpose.

No commit message

No commit message

Changes for South America runs: regions files in analysis directory are copied from the one specified in the system.rc file. Region pickle files are deleted in the analysis directory, to make sure the correct one is used. Option to use covariances for less than 23 transcom regions, but still use 23 for the analysis (does not work in tcfluxes.nc). Update in observation_operator.py to use restart file in first cycle also for more than one resolution (i.e. incl. zoom regions).

Added option to transition from 25 to 34 layers for OD meteo after December 2005. Added option to read initial random seed from file. Localization method is now read in from rc file. Changed terminology of mixing ratio to mole fraction.

rename *PlatForm to *Platform

renamed class names Ct- to CO2-
changed class name Observation to Observations 
small cleanup changes to template.py

reverted to r1186 (the last one from 2012), with changes in observation files from r1192-1194 preserved

original run scripts now in the example dir. Copies should go to the da/trunk/ main dir before running

changed some of the initialization logic. Now, the DaCycle object has become internal to the obsoperator, sample, and statevector objects

DaSystem
DaPlatform
ObservationOperator

are now handled before the pipeline, from the das.py contents. This replaces the situation in which they were all 
constructed inside the pipeline based on rc-file items from the da.rc script. This now only serves to identify the 
cycle information, and no longer hold info about the platform, sampling model, and da system used inside a cycle.