[HELP] Import requirements from CSV/EXCEL to CAPELLA

[jeremy-jimmy]

Hi,
I'm currently trying to build a python function able to operate export/import of the requirements stocked in my Capella to an EXCEL file.

Using the package "capellambse" I manage quite easily to handle the export from CAPELLA to CSV/EXCEL but I'm facing difficulties to handle the other way.

My main problem is that so far I didn't acheive to custom attributes to my requirement "on the fly" with my script.

model = capellambse.MelodyModel(PATH_CAPELLA, jupyter_untrusted=True) REQ = model.sa.requirements.create(name="TEST_XTYPE", xtype="Requirements:Requirement")

How to add custom attributes fields to my object REQ ?

Thanks in advance for support,

Jérémy

[vik378]

nice use case indeed, I think we can add an example for how to create a req attribute into https://github.com/DSD-DBS/py-capellambse/blob/master/examples/06%20Introduction%20to%20Requirement%20access%20and%20management.ipynb or maybe even an example for excel import with attribute definition creation. The catch there was about having the definitions / enums pre defined or imported upfront. @ewuerger , @Wuestengecko - since you are deep in this topic now, would any of you have time to add a quick example for attr instantiation?

[jeremy-jimmy]

That would be great to add in the jupyter notebooks indeed, thanks in advance

[ewuerger]

How to add custom attributes fields to my object REQ ?

Capella (Requirements plugin) is doing that in an overengineered way. While implementing some Requirement Bots that automate maintenance we used the API to its full extent already. Just to get you going instead of waiting for a demo notebook:
In order to add AttributeValues to a Requirement you can use the same create functionality as seen in the notebook. Here it would go like this:

from capellambse.extensions import reqif

requirement = model.search(reqif.XT_REQUIREMENT)[0]
requirement.attributes.create("int", name="Example", value=10)

If you forget to handover the type hint (here "int") then the API will tell you to do so, since this is needed for creation. Possible types are:

• int, integer, integervalueattribute
• str, string, stringvalueattribute
• float, real, realvalueattribute
• date, datevalueattribute
• bool, boolvalueattribute
• enum, enumvalueattribute

Then this is all you need to create attribute values underneath your requirements. But if you want to add a requirement type and/or definition to the AttributeValue you'll need to make sure that a RequirementTypesFolder exists (can also be created via API) with a RequirementType and AttributeDefinitions underneath. It is your choice then to set these attributes on your new requirement via:

reqtfolder = model.la.requirement_types_folders[0]
reqtype = reqtfolder.requirement_types[0]
reqtype.attribute_definitions.create(reqif.XT_REQ_TYPE_ATTR_DEF, long_name="Example Definition")
new_req_attr = requirement.attributes[-1]
new_req_attr.type = reqtype
new_req_attr.definition = reqtype.attribute_definitions[0]

or handover while creation of attributes:

definition = reqtype.attribute_definitions.by_long_name("Example Definition")
requirement.attributes.create("int", name="Example", type=reqtype, definition=definition)

We like the latter more. Also this will improve the display of AttributeValues inside of Capella (since the name of the definition is displayed).
Personally I think that normal AttributeDataTypeDefinition don't have a use for now. But underneath EnumAttributeDataTypeDefinition you can define the EnumValues. These will be available for EnumAttributeValues if you set the data_type attribute on the EnumAttributeDefinition correctly. Keep that in mind.

Key takeaways:

• Creation is enabled on attributes that handback an ElementList
• When creating elements where multiple classtypes are possible the API will raise a ValueError to inform you. You'll need to find the right XTYPE string in the reqif module. We'll work on refactoring this to also take class-types right away.

We'll work on improving documentation and showcases in the future.

After that it will be a walk in the park to get your Pandas DataFrame content back into Capella. Maybe if you are invested, you can try to extent reqif to offer import from excel-file?

[jeremy-jimmy]

Thank for the fast answer, I will read this carefully and I hope I will manage to handle my problem with your help.

[jeremy-jimmy]

I managed to handle my use-case with your help, thanks again

[Luffy-cpu]

Hi,

I have a question related to this one. In the case the attribute already exists in the model, if I want to update its value, how can I do it? Which method should I call, please?

Thanks in advance,
Christian

[Luffy-cpu]

Here the result for the commands

python --version: Python 3.9.13

python -c "import capellambse; print(capellambse.version, capellambse.file)": 0.5.41 path_to_python/python3.9/site-packages/capellambse/init.py

pip show capellambse:
Name: capellambse
Version: 0.5.41
Summary: Provides access to Capella MBSE projects in Python
Home-page:
Author: DB Netz AG
Author-email:
License: Apache-2.0
Location: path_to_python/python3.9/site-packages
Requires: lxml, markupsafe, Pillow, platformdirs, svgwrite, typing-extensions
Required-by: capella-RM-Bridge

[Wuestengecko]

Ah, that's what I suspected ... you're using Python 3.9. Newer versions of capellambse use some 3.10+ specific features, so you'll have to upgrade to a newer version in order to use those. If it's an option for you, I'd suggest to go straight to the newest stable version 3.12, but 3.10 should also be fine for the time being.

[Luffy-cpu]

Okay, I understand. I upgraded to Python 3.12 and decl.FindBy works correctly now and I am able to generate the right YAML.

While saving the model, the following error occurs:
CorruptModelError(
capellambse.loader.core.CorruptModelError: Refusing to save a corrupt model without having a backup (hint: pass i_have_a_recent_backup=True)

So I tried with "model.save(i_have_a_recent_backup=True)". The error message no longer appears. But is it the recommended way to do it?

Finally, the text attribute is updated in my Capella model, but the custom attribute Rationale remains unchanged. Is there something I did wrong?

[Wuestengecko]

This error is a safeguard against making things even worse if you had to load the model with ignore_duplicate_uuids_and_void_all_warranties=True. The recommended way is to try and get rid of that :)

The duplication check used to be too strict in older versions of capellambse (in some cases, duplication is harmless), so now that you're up to date with latest stable, you might not even need that anymore anyways.

As for why the custom attribute is missing, I don't know yet, I'll try if I can reproduce that.

[Luffy-cpu]

Thank you very much.
I understand better about he safeguard.

For the custom attribute issue, I managed to import it, but I still have a bug. I have now 2 attributes for Rationale after the import.
As shown in the following screenshot, a new attribute called [String Value Attribute] is created with the value of the Rationale attribute. And the initial Rationale attribute remains with its old value.

I have one more question : in addition to attribute update, if I want to update requirement relations, how can I perfom it, please ?

Thank you again