Fix tag being undefined bug from the backend.

[rashidakanchwala]

Description

Resolves #2106.

Upon investigation based on what user did, I realised that there might be cases where tags might be iteratively added to a node as we iterate through different pipelines. In this case, the node initially had certain tags, but __default__ pipeline later added additional ones. In our GraphNodes repository, once a graph node is defined, it doesn't get updated with new tags.

This PR changes that by allowing nodes to update with any new tags even if they’re already defined. Before the namespace refactor, this issue surprisingly didn’t arise, likely because certain aspects of the refactor introduced unintended behavior that ended up compensating for each other.

See the example to understand better - the code below works fine because __default__ is defined first, so all nodes belonging to the default pipeline are added to GraphNodeRepository with the correct tags, without any errors:

working_pipes = {
    "__default__": pipeline(
        ingestion_pipeline + feature_pipeline + modelling_pipeline + reporting_pipeline,
        tags=["default_tag1", "default_tag2"]
    ),
    "Data ingestion": ingestion_pipeline,
    "Modelling stage": modelling_pipeline,
    "Feature engineering": feature_pipeline,
    "Reporting stage": reporting_pipeline,
    "Pre-modelling": ingestion_pipeline + feature_pipeline,
}
return working_pipes

However, if __default__ is defined last, nodes from ingestion_pipeline are already created. When we try to recreate these nodes with the new tags from __default__, the tags aren’t added since the nodes already exist.

working_pipes = {
    "Data ingestion": ingestion_pipeline,
    "Modelling stage": modelling_pipeline,
    "Feature engineering": feature_pipeline,
    "Reporting stage": reporting_pipeline,
    "Pre-modelling": ingestion_pipeline + feature_pipeline,
    "__default__": pipeline(
        ingestion_pipeline + feature_pipeline + modelling_pipeline + reporting_pipeline,
        tags=["default_tag1", "default_tag2"]
    ),
}
return working_pipes

This PR resolves the issue by checking if a node already exists and, if so, assigning it any additional tags:, existing nodes correctly receive any additional tags defined in subsequent pipelines, preventing skipped tag assignments.

QA notes

Checklist

• Read the contributing guidelines
• Opened this PR as a 'Draft Pull Request' if it is work-in-progress
• Updated the documentation to reflect the code changes
• Added new entries to the RELEASE.md file
• Added tests to cover my changes

[ravi-kumar-pilla]

Tested the usecase and it looks fixed. Thank you @rashidakanchwala

[SajidAlamQB]

Thank you @rashidakanchwala!