Validator preparator

The Validator is a web-based interactive tool for validation of ocean models at oceanographic stations.

0. How to access

This documentation focuses on the usage within the IOW intranet, which gives access to the station data of the IOW_DATABASE. The validator can be accessed at

http://10.10.10.20/radtke/validator/

Note that the IP is provided here, since some persons have problems with their DNS server when using a VPN.

Once opened you will see pre-configured model data. To add new model data, tik the box of other models on the left column and upload your model list. You will also find the option to download a sample model list that helps with the sytnax.

1. How to use

There is no common model output. Therefore, an example on the usage of the existing validator preparator can be found in prepare_for_validator_MOM5.py.

This an example suitable for validation of the MOM5 model. Please read the comments therein.

Note that if you still need a more specialized method to extract the station name from the file names you can derive a new class from the existing one like e.g.

#...
import validator_preparator_MOM5

# initialize the concrete class
class ValidatorPreparatorMOM5Special(validator_preparator_MOM5.ValidatorPreparatorMOM5):
    def __init__(self, parameters):
        # call base constructor, pass the parameters
        validator_preparator_MOM5.ValidatorPreparatorMOM5.__init__(self, parameters)
        
    # overrides base-class method
    # return value has to be one of the keys in self.station_names
    def extract_station_name(self, file_name):
        return "do a very special thing here"
		
# prepare parameters for the preparator
parameters = validator_preparator.ValidatorParameters()
#...

# create an instance of a concrete vaildator preparator (in this case your special MOM5)
# instantiate it with the parameters object
preparator = ValidatorPreparatorMOM5Special(parameters)
#...

2. How to develop

If you want to extend the validator preparator for another model, you have to create a new module, e.g. validator_preparator_XXX.py. It is intended that new modules define a class that inherits from the base class ValidatorPreparator in validator_preparator.py. For the implementation details please have a look at the comments therein.

In the above example the module could look like:

# import the base class
import validator_preparator

# define a dictionary that relates XXX variable names to validator variable names 
# the keys (left) are the model variables, the corresponding values are the validator variables
variable_names     = {
    "XXX_temp" : "MODEL_TEMP"
    }

# define a dictionary that relates XXX station names to validator station names
# the keys (left) are the model stations, the corresponding values are the validator stations
station_names = {
    "XXX_anholte"           : "anholte"
    }

# define how the depth pattern for the model looks like
depth_pattern = "XXX_depth*"

# initialize the concrete class
class ValidatorPreparatorXXX(validator_preparator.ValidatorPreparator):
    def __init__(self, parameters):
        # call base constructor, pass the concrete dictionaries defined above
        validator_preparator.ValidatorPreparator.__init__(self, parameters, variable_names, station_names, depth_pattern)
        
    # overrides base-class method
    # return value has to be one of the keys in self.station_names
    def extract_station_name(self, file_name):
        return "whatever has to be done here"

Naturally, the strings starting with XXX_ should be replaced by the actual names of your model and the dictionaries should be completed for your purposes.

Importantly, the method extract_station_name has to be implemented such that it returns one of the station_name keys.

The rest should be done by the base class and there is no need for further implementation. It would be nice if a usage example for the new module is added to the repository.