pyreverse: add PlantUML output
#4846
Conversation
| for arg, ann in annotations.items() | ||
| ) | ||
|
|
||
| label += fr"{func.name}({args}){return_type}\l" |
There was a problem hiding this comment.
The same logic is used by PlantUmlPrinter, so I extracted a method and put it inside the base Printer class.
| @@ -209,8 +207,7 @@ def run(self, args): | |||
| diadefs = handler.get_diadefs(project, linker) | |||
| finally: | |||
| sys.path.pop(0) | |||
| printer_class = VCGPrinter if self.config.output_format == "vcg" else DotPrinter | |||
| writer.DiagramWriter(self.config, printer_class).write(diadefs) | |||
| writer.DiagramWriter(self.config).write(diadefs) | |||
There was a problem hiding this comment.
I introduced a factory function to get the printer class for a given filetype. This supports the Open-Closed-Principle (only factory function needs to be updated if a new printer is added). Also, the DiagramWriter can now obtain the printer class itself.
| f"{arg.name}: {ann}" if ann else f"{arg.name}" | ||
| for arg, ann in annotations.items() | ||
| ] | ||
|
|
There was a problem hiding this comment.
This method is used by both DotPrinter and PlantUmlPrinter. We could probably add this feature for VCG as well if somebody needs it.
Pull Request Test Coverage Report for Build 1131107870
💛 - Coveralls |
| self.my_int = optional_int | ||
|
|
||
| def do_it(self, new_int: int) -> int: | ||
| return self.my_int + new_int |
There was a problem hiding this comment.
This new class is for checking that #4551 works correctly with PlantUML output too.
| ) | ||
| assert expected == generated, f"{files}{diff}" | ||
| os.remove(generated_file) | ||
| _assert_files_are_equal(generated_file) |
There was a problem hiding this comment.
The logic for each test case is basically the same, but the used fixtures differ. To avoid duplicated code I extracted a helper function for it.
The setup_xy are fixtures as they
a) shall perform cleanup code after the test case finishes
b) test scope is comparing the output file with expectation, not the process of generating the files itself
Is there a way to specify the fixture to use in pytest.mark.parametrize?
| @@ -136,4 +140,3 @@ def test_vcg_files(generated_file): | |||
| line for line in unified_diff(expected.splitlines(), generated.splitlines()) | |||
| ) | |||
| assert expected == generated, f"{files}{diff}" | |||
| os.remove(generated_file) | |||
There was a problem hiding this comment.
Cleanup is performed in the fixture, this os.remove is not necessary.
|
Maybe we could try covering the 3 lines not covered in the new plantuml writter with a diagram that would trigger them? |
|
Thanks for all your reviews, too! 😁 |
Pierre-Sassoulas
added
Enhancement ✨
Understood, let's merge this then :) |
doc/whatsnew/<current release.rst>.Type of Changes
Description
This MR adds output in PlantUML format to
pyreverse.I wanted to make sure that the feature from #4551 also works properly with PlantUML, so I added a new class to the test files.
This made the list of changed files a bit longer as intended.
Closes #4498