validations.md

Validations

Creating a new Entity

Product

product = Product.new(name="ExampleProduct", code="EXPRD", description="Example Product")
client.create(product)

• Product code and name should not be empty

• Product code should have between 1 and 6 characters

• If the product is already imported in the DB, the new entity creation will be skipped

• Product code and name must be unique (no duplicates in the database)

Variable

variable = Variable.new(code="Var1", 
                        name="SDK Variable 1", 
                        description="This is a variable created with the SDK", 
                        measurement_unit="g/l", 
                        variable_group="X Variables", 
                        variable_type=VariableNumeric()
                        )
client.create(variable)

• Variable code, name, measurement_unit and variable_group can't be empty

• variable_group must be valid. Use variable_groups = Variable.get_valid_variable_groups(client) to check all valid Variable Groups.

• If Variable is already imported, the new entity creation will be skipped.

• Variable code and name must be unique (no duplicates in the database)

• Measurement Unit is mandatory

If the new Variable is of type Feeds/Flows (and you're using the VariableFlow() class), this adds a few extra validations:

• References are mandatory
• measurementId must be the ID of a X Variable already imported.
• concentrationId must be ID of a Feed Concentration variable already imported.

Experiment

new_data = {
            "EXv1": {
                "timestamps": [
                    1600674350,
                    1600760750,
                    1600847150,

                ],
                "values": [
                    5.1,
                    3.5,
                    1.3,
                ]
            },
            "EXv2": {
                "timestamps": [
                    1600674350],
                "values": [
                    "A"]

            }
}

new_experiment = Experiment.new(name="Example Experiment",
                                description="new experiment example for sdk",
                                product=product,
                                variables=variables,
                                data_type="run",
                                data=new_data,
                                variant="run",
                                start_time="2020-09-21T08:45:50Z", 
                                end_time="2020-09-30T08:45:50Z")
client.create(new_experiment)

Importing a new experiment splits the validation between the new experiment metadata and the data associated with that experiment. So, following the same logic:

Experiment Metadata

• Experiment name and description can't be empty
• Experiment variant has to be one of two options (run or samples). run is for time series data and samples is for sampled data.
  • If the variant is run, then the arguments start_time and end_time are mandatory.
• Experiment data_type has to be one of two options (run or spectra). run is for cultivation experiments and spectra is for spectra experiments.
• If the Experiment is already present in the DB, the import will be skipped.
• product and all variables need to be already present in the DB.
• data needs to be in a dict format and not be empty.

Experiment data

• If one of the variable codes indicated in the argument variables is not present in data, an Error will be thrown.
• Inside each key of the data dictionary, a new dictionary must be present with timestamps and values as mandatory fields.
• If timestamps or values are missing from any variable, an Error will be thrown.
• If the data inside timestamps or values is not a list, then the data format is not valid. Even if we only have one value, it should be inside a list.
• The length of the lists inside timestamps and values must match. For each timestamp we need a value.
• The timestamps need to be in seconds in Unix time.
• The timestamps need to be sorted.
• For the run variant, all the timestamps must be between the value of start_time and end_time of the experiment.