[wheevu]

Resolves #77

Type of change

• Bug fix
• New feature
• Breaking change
• This change requires a documentation update

Changes

• Adds convert_to_pgraph() and convert_from_pgraph() for NetworkX, edge lists, and dictionaries of dictionaries, sets, or lists.
• Moves the conversion helpers into planarity_functions.py, preserving the top-level pgraph_graph() and networkx_graph() entry points.
• Adds 14 tests covering conversions, round trips, isolated vertices, labels, drawing metadata, and errors.

Direct imports of the old helpers from planarity.express.planarity_networkx no longer work, as requested by the move.
Top-level imports remain supported.

The compatibility comments omit the sentence about future docstrings because the new converters already have full docstrings.

Testing

• Full suite: 81 passed on both Python 3.11.8 and 3.14.7, macOS ARM64.
• Examples and Sphinx documentation build passed.
• Python 3.12, 3.13, Windows, and Linux were not tested locally.

The unchanged tests/test_express_api_networkx.py passes before and after:

Before, master at 92e63d1

$ pytest -vv tests/test_express_api_networkx.py

tests/test_express_api_networkx.py::TestPlanarityNetworkX::test_is_planar PASSED [ 16%]
tests/test_express_api_networkx.py::TestPlanarityNetworkX::test_is_planar_unions PASSED [ 33%]
tests/test_express_api_networkx.py::TestPlanarityNetworkX::test_goldner_harary PASSED [ 50%]
tests/test_express_api_networkx.py::TestPlanarityNetworkX::test_kuratowski_k5 PASSED [ 66%]
tests/test_express_api_networkx.py::TestPlanarityNetworkX::test_kuratowski_k5m PASSED [ 83%]
tests/test_express_api_networkx.py::TestPlanarityNetworkX::test_networkx_graph PASSED [100%]

============================== 6 passed in 2.54s ===============================

After

$ pytest -vv tests/test_express_api_networkx.py

tests/test_express_api_networkx.py::TestPlanarityNetworkX::test_is_planar PASSED [ 16%]
tests/test_express_api_networkx.py::TestPlanarityNetworkX::test_is_planar_unions PASSED [ 33%]
tests/test_express_api_networkx.py::TestPlanarityNetworkX::test_goldner_harary PASSED [ 50%]
tests/test_express_api_networkx.py::TestPlanarityNetworkX::test_kuratowski_k5 PASSED [ 66%]
tests/test_express_api_networkx.py::TestPlanarityNetworkX::test_kuratowski_k5m PASSED [ 83%]
tests/test_express_api_networkx.py::TestPlanarityNetworkX::test_networkx_graph PASSED [100%]

============================== 6 passed in 0.13s ===============================

[wbkboyer]

Looking great so far, Josh (@wheevu)! In addition to the changes requested, could you locally run

pylint --disable="invalid-name,missing-class-docstring,missing-function-docstring,missing-module-docstring,import-error" tests/test_express_api_conversions.py

And resolve the warnings? For the global-statement, undefined-variable, and import-outside-toplevel, if there's no better way to handle testing for networkx in the test environment and skipping the test if it's not present, then you can # pylint: disable=... on those lines!

[wheevu]

Thanks, Wanda @wbkboyer! 😄 I addressed your comments in 3caf406: added dictionary-of-lists coverage for the isolated vertex, corrected cls, and resolved the Pylint warnings. Your Pylint command is clean, and all 132 tests pass locally on Python 3.11 and 3.14.

[wbkboyer]

Thanks, Wanda @wbkboyer! 😄 I addressed your comments in 3caf406: added dictionary-of-lists coverage for the isolated vertex, corrected cls, and resolved the Pylint warnings. Your Pylint command is clean, and all 132 tests pass locally on Python 3.11 and 3.14.

Right on, thank you for addressing those issues, @wheevu 🎉 I ended up making some additional changes (21ea12) to fix some of the leftover pylint warnings: I use the @pytest.mark.skipif() to selectively condition test execution on a variable that's set based on the successful import of dependencies networkx, matplotlib, pillow (is a dependency of matplotlib, so gets installed when we pip install matplotlib), or sphinx 💪🏼 I'll take one final look, but I think this is ready to rock at this point! Thank you again!

[wbkboyer]

I think this is looking good; thanks again, @wheevu ! 🎉