Make pipeline discovery more flexible

[jitu5]

Resolves https://github.com/kedro-org/vscode-kedro/issues/18

This PR improves the Kedro extension's ability to detect and navigate to pipeline files by supporting more flexible file naming conventions.

• Updated the references function to use glob with recursive=True, allowing it to find all files containing "pipeline" in their names, even in nested subdirectories.
• Replaced the fixed check for pipeline.py with a pattern match for pipeline\.py.

Now it Supports non-standard modular pipeline structures.

- pipelines
  - sub_pipeline
    - pipeline_data_processing.py
    - sub_pipeline_1
        - pipeline_data_processing_1.py

[jitu5]

I think this looks good. Can you add some notes to https://github.com/kedro-org/vscode-kedro/blob/main/CHANGELOG.md.

It may be helpful to also update in README.md to explains what structure does the extension support.

I have updated README.md and CHANGELOG.md please review.

Even though we don't have automated tests for this yet, I think we can also setup a sample project in this repo in test just for review purpose. It will be easier to write instructions for us what to test during review.

@noklam do you mean to add kedro project like demo-project in Kedro-Viz to test folder ? How can we run this extension over this sample kedro project ?

[noklam]

@noklam do you mean to add kedro project like demo-project in Kedro-Viz to test folder ? How can we run this extension over this sample kedro project ?

Yes, I think it's good enough to have a test project, it will include multiple pipelines (maybe filename like pipeline_abc.py). We don't have to do this now. Let's discuss next time when we catch up.

Eventually I want to create some unit tests, similar to Kedro's e2e test setup.