[Mesh Python SDK] Open functionality to write/update template calculations

[AKRittinghausen]

Is your feature request related to a problem? Please describe.
Hi SDK team,

We recently came across a set of use cases where the ability to update template expressions via the Python API would have been extremely useful.

_Use case 1: xxx --> @FlorianVolue
They received a list of template expressions extracted from an old database that they now have to copy and paste manually into a new model.

_Use case 2: yyy --> @stefanosala99
They want to transfer timeseries template calculations from pre-production to production.

Use Case 3 - Python Best Practice for reading out and comparing two system environments --> @AKRittinghausen
Here, we started an initiative to promote coding best practices for consultants. The aim is to present the mesh model structure in a human-readable format, enabling consultants to track changes in template calculations and model tree structure Volue/BestPractice_TestProdCopy. If there were an option to write template calculations, it would be possible to "merge" or "synchronise" two systems without having to copy and paste between them, e.g. by using .diff files.

Describe the solution you'd like
From what we understand, there is currently no option to write template calculations into Mesh using the Python SDK. However, there is the option to write local calculations into Mesh using the update_timeseries_attribute() method.

While we recognise the potential risks of updating template expressions via the Python interface, we also recognise the significant use case for it. We understand that when adding template calculations via the calculator, an additional syntax check is performed by the 'Validate expression' button.

However, perhaps this is something that could be incorporated into the Python SDK also? And, to be fair, the "validate button" does not really prevent one from saving nonsense to calculator, either 😄

Describe alternatives you've considered
That would probably be manual copy paste

[AKRittinghausen]

@VictorEricsson, since you are the Mesh guardian, perhaps you have a view on this as well? ☺️ (after your holidays of course)

[tnoczyns-volue]

@AKRittinghausen would it suffice to enable setting template expressions on time series attribute definitions? E.g.:

def update_timeseries_attribute_definition(
        self,
        target: Union[uuid.UUID, str, AttributeBaseDefinition],
        new_template_expression: Optional[str] = None
    ) -> None:

You can get list of attribute definitions by searching for all time series attributes and getting their definition:

for obj in session.search_for_objects(target, "{*}", full_attribute_info=True):
    for attr in obj.attributes.values():
        if isinstance(attr, TimeseriesAttribute):
            definition = attr.definition
            # proceed

Depending on your use case: keep in mind that a lot of attributes can have the same attribute definition, so it is better to create first a unique list of attribute definitions (e.g. by ID or path) and then proceed.

Additionally, this won't help if some time series attribute definitions are not existing.

So, this would be helpful if you:

• want to set new template expression on existing time series attribute definition.
• have 2 model definitions with the same structure, but different template expressions and you want to sync them

See also: https://github.com/Volue/energy-mesh/issues/6737

[AKRittinghausen]

Hi @tnoczyns-volue, finally back from my holidays to reply ☺️

Thanks for your comment!
In fact, the two use cases you outlined are exactly the ones that are interesting for us - and I believe for the majority of consultants.

@AKRittinghausen would it suffice to enable setting template expressions on time series attribute definitions? E.g.:

def update_timeseries_attribute_definition(
self,
target: Union[uuid.UUID, str, AttributeBaseDefinition],
new_template_expression: Optional[str] = None
) -> None:

Answer: Yes, enabling setting template expressions would suffice. If I understand correctly, this would be similar to the way local expressions can already be set, right? (code taken from your example section ).

    # Change local expression of a time series attribute of the new object.
    new_local_expression = "## = 2\n"
    session.update_timeseries_attribute(
        new_object.attributes["TsCalcAtt"],
        new_local_expression=new_local_expression,
    )

Best regards!

[tnoczyns-volue]

If I understand correctly, this would be similar to the way local expressions can already be set, right?

Yes, exactly. But instead of update_timeseries_attribute function there will be e.g.: update_timeseries_attribute_definition:

def update_timeseries_attribute_definition(
        self,
        target: Union[uuid.UUID, str, AttributeBaseDefinition],
        new_template_expression: Optional[str] = None
    ) -> None:

Thanks for your feedback.

[AKRittinghausen]

If I understand correctly, this would be similar to the way local expressions can already be set, right?

Yes, exactly. But instead of update_timeseries_attribute function there will be e.g.: update_timeseries_attribute_definition:

Perfect, I fully agree with the naming 👍

[AKRittinghausen]

Do you think this could be a change already for the next PythonSDK release, @tnoczyns-volue? 🙃

[tnoczyns-volue]

Do you think this could be a change already for the next PythonSDK release, @tnoczyns-volue? 🙃

Unfortunately I can't say right now...

[AKRittinghausen]

Sure, no worries ☺️