fix: set changed size when iterating the store's graphs

[edmondchuc]

Summary of changes

Fix: #3102

The RuntimeError: Set changed size during iteration occurs because a Dataset's default graph is lazily evaluated and only added to the store's context dict when accessed.

Since the default graph is always present in a dataset, this new change adds the default graph to the context dict on dataset creation. - this change breaks a bunch of tests unrelated to Dataset, like namespace manager etc.

We materialise the iterator using a list() call. This way, when the default graph is added to the contexts dict, we don't get the runtime error.

We also avoid calling the super class (ConjunctiveGraph)'s methods as it's redundant and we'll be removing the class hierarchy soon. This now calls self.store.contexts directly.

Checklist

• Checked that there aren't other open pull requests for
  the same change.
• Checked that all tests and type checking passes.
• If the change adds new features or changes the RDFLib public API:
  • Created an issue to discuss the change and get in-principle agreement.
  • Considered adding an example in ./examples.
• If the change has a potential impact on users of this project:
  • Added or updated tests that fail without the change.
  • Updated relevant documentation to avoid inaccuracies.
  • Considered adding additional documentation.
• Considered granting push permissions to the PR branch,
  so maintainers can fix minor issues and keep your PR up to date.

[edmondchuc]

This is the actual fix.

[edmondchuc]

Because the default graph is lazily evaluated, it makes a call here to self.graph, which adds the default graph to the store. This changes the context iterator size and throws a runtime error.

[edmondchuc]

I'm a bit puzzled as to why this only fails on the 3.8 tests in GitHub Action while it passes fine on python 3.8 locally.

[edmondchuc]

To reproduce the failing test locally, run:

./with-fuseki.sh poetry run pytest -v test/test_dataset/test_dataset.py

[edmondchuc]

The current Python 3.8 tests are extensive and include spinning up a Fuseki instance during the test run. These tests conditionally register the SPARQLUpdateStore plugin only if the Fuseki endpoint is reachable.

The issue arises from a change where the Dataset.contexts() method was modified to call the store’s contexts() method directly in 63caed5. However, not all stores guarantee that their contexts() implementation returns graph objects. For example, SPARQLUpdateStore returns graph names as URIRefs instead. To ensure contexts() always returns graph objects, we delegate to ConjunctiveGraph.contexts(), which converts graph name URIRefs into graph objects.