[tclose]

General-purpose pieces for building modified copies of filesets that contain others, factored out of the deidentification work in fileformats-medimage and fileformats-biosig. Nothing here is specific to deidentification.

Changes

• FileSet.nested_filesets(recursive=True) now also finds filesets inside dicts and lists returned by validated properties (previously skipped). recursive=False returns only the direct children.
• New FileSet.map_nested(out_dir, func):
  • Writes a copy of the fileset, calling func(nested, dest_dir) to write each directly nested fileset in its mirrored position; returning None copies it unchanged. All other files are copied, keeping their layout relative to the common parent.
  • Nested filesets that point entirely outside the fileset are skipped, consistent with copy(). One that's partly inside and partly outside raises ValueError.
  • Recreating the type at the end checks that the layout survived.
• New staged_out_dir wrapper (exported from fileformats.core) for extras that write to an out_dir argument, used as @extra(wrapper=staged_out_dir):
  • Implementations write into a hidden sibling staging directory, which is moved into out_dir only once they return, so only complete output appears there.
  • Raises FileExistsError before moving anything if the output would overwrite existing files.
  • Checks that the implementation wrote into the directory it was given.

Tests

• New core/tests/test_map_nested.py covering:
  • nested filesets inside dicts and lists, and recursive=False;
  • mapping, copying unchanged, skipping external filesets, and the layout check;
  • staging: success, failure leaves nothing behind, and refusal to overwrite.

[codecov]

Codecov Report

❌ Patch coverage is 84.61538% with 10 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
fileformats/core/fileset.py	80.95%	5 Missing and 3 partials ⚠️
fileformats/core/extras.py	90.90%	2 Missing ⚠️

📢 Thoughts on this report? Let us know!