DOC: validate_docstrings cleans up leftover files; doctest +SKIP file examples #44711
Conversation
| for test in finder.find(self.raw_doc, self.name, globs=context): | ||
| f = io.StringIO() | ||
| runner.run(test, out=f.write) | ||
| error_msgs += f.getvalue() | ||
| leftovers = set(os.listdir()).difference(current_dir) |
There was a problem hiding this comment.
if you do this why do we need skips?
There was a problem hiding this comment.
This is to validate that no doctest ever creates leftover files. So the workflow is
- Check if a docstring leftover a file/folder
- If so, clean up those leftover files
- Short-circut the validation and raise that a docstring leftover a file/folder
May helps potentially avoid any ResourceWarnings or potential segfaults from the docstring validation due to created files.
There was a problem hiding this comment.
sure i get it, but you added skips to all of the doc-strings hence they are not running?
There was a problem hiding this comment.
Correct.
IMO those docstrings that show an example with interacting with files should be skipped since they are more of a demonstration. The same behavior interacting with our file should be sufficient tested in the test suite.
Also I think it's better than having os.remove everywhere in the docstrings.
And we have this precedent in the docstrings already to skip the ones that interact with files:
pandas/pandas/io/excel/_base.py
Line 292 in e992969
May help towards: #44590