Running a specific node using tags in 0.17.2

Introduction

I’m looking to run a specific node, following instructions [here], for testing purposes. Not sure I’d need to register this pipeline to be able to do that… (https://kedro.readthedocs.io/en/stable/06_nodes_and_pipelines/01_nodes.html?highlight=tags#how-to-tag-a-node). Anyhow, I created a new environment and started a new project using kedro new pandas-iris. Next I added an example pipeline with a node I’d like to test using tags.

I add the following line to src/{{ project name }}/pipelines/data_science/pipeline.py: def test_pipeline(**kwargs): return Pipeline([node(lambda x: x,"example_train_x","example_model",name="test",tags="test")]

Next, I try to run the specific node using kedro run --tag test and get the error

Question

Is there something I’m missing here? Do I need to follow certain steps after generating pandas-iris, specific to pipeline registry? How can I run the nodes tagged with test? I think the default pipeline error may be linked to this

[WaylonWalker]

Whether this turns out as user error or not the Error message here is quite confusing and not very helpful.

If I render a new kedro template with 0l17.2 and delete my pipeline_registry.py module I get the same error. In this case the error should guide the user to create a register_pipelines function in the appropriate module

If I rename __default__ to new I get the same error. In this case the error should make its best attempt to guide the user to the module with the register_pipelines function and add a __default__ key.

[WaylonWalker]

The error you are seeing is due to improperly formatted pipelines dictionary, or a missing register_pipelines function.

Are you using pipeline_registry.py or hooks.py?

Look in either of those modules for a function called register_pipelines. It should return a dictionary of pipelines, one must be found under the key __default__