[experiment] validate: stream annexed content with datalad-fuse's adapter when available

[CodyCBakerPhD]

Stacked on #1933. This tries the suggestion from #1933: use datalad-fuse with its fsspec adapter, not FUSE itself, tested against datalad/datalad-fuse#131. #1933 itself keeps its pure-fsspec reader.

What changes

• dandi/support/datalad_fuse.py adds DataladFuseReadableFile, a Readable that opens an annexed file's missing content through datalad-fuse's RemoteFilesystemAdapter.
  • No FUSE mount is involved.
  • The adapter is created with caching=False, so nothing is written into the dataset's .git.
  • It uses datalad-fuse's default backend chain (remfile,fsspec), which can be changed with the datalad.fusefs.backends config.
  • Like AnnexReadableFile, its fingerprint is the annex key, so pynwb validation and metadata results are cached under the key.
• get_datalad_fuse_readable(path) returns None, so the caller falls back to the git-only reader in dandi.support.annex, when:
  • datalad-fuse isn't installed;
  • the path isn't a locked, dropped annexed file;
  • git-annex isn't initialized in the repository (otherwise DataLad would run git annex init on it);
  • or git-annex knows no URL for the key.
• validate --missing-file-content=stream tries datalad-fuse first and falls back to the git-only reader. The DANDI.FILE_CONTENT_STREAMED message names which one was used.
• What datalad-fuse adds over the git-only reader: it asks git-annex itself (whereis), so it also finds content that is reachable through a remote but has no registered URL, such as Forgejo-aneksajo or S3 exporttree remotes.

Tests

• make_git_annex_dandiset() builds a real git-annex repository: content is added, a URL is registered for it, and then it is dropped. A skipif.no_git_annex condition is added for these tests.
• New tests:
  • dandi/support/tests/test_datalad_fuse.py: streaming over a local HTTP server with range support, plus the cases that return None.
  • test_validate_stream_datalad_fuse: content-dependent results match a local copy of the file.
• A new CI job, mode: datalad-fuse (ubuntu, Python 3.12), installs git-annex from PyPI and datalad-fuse from refs/pull/131/head. Every other job runs without datalad-fuse, so they still exercise the fallback.

Open questions

• datalad-fuse brings in datalad, boto3 and fusepy, and needs git-annex. Should it become an optional extra once dandiset.yaml: should become READONLY locally, with a comment, no metadata uploaded from it #131 is released, or stay something users opt into by installing it?
• Once that release exists, switch the CI install step from the PR ref to the released version.

🤖 Generated with Claude Code

https://claude.ai/code/session_01Qn5WBSiQgoZoL4fytF6nEr

---

Generated by Claude Code

[codecov]

Codecov Report

❌ Patch coverage is 92.04545% with 14 lines in your changes missing coverage. Please review.
✅ Project coverage is 78.93%. Comparing base (d21b552) to head (25791fb).

Files with missing lines	Patch %	Lines
dandi/support/datalad_fuse.py	84.00%	12 Missing ⚠️
dandi/tests/fixtures.py	94.73%	1 Missing ⚠️
dandi/tests/skip.py	50.00%	1 Missing ⚠️

Additional details and impacted files

@@                       Coverage Diff                        @@
##           claude/brave-hamilton-ast647    #1952      +/-   ##
================================================================
+ Coverage                         78.78%   78.93%   +0.15%     
================================================================
  Files                                94       96       +2     
  Lines                             14727    14899     +172     
================================================================
+ Hits                              11603    11761     +158     
- Misses                             3124     3138      +14     

Flag	Coverage Δ
unittests	78.93% <92.04%> (+0.15%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.