fix: correctly parse headerless _motion.tsv files

[Yehuda-Bergstein]

Fixes #394

Description

This PR fixes a bug where the Deno validator incorrectly parsed the first row of data in
_motion.tsv files as column headers, resulting in false-positive
CUSTOM_COLUMN_WITHOUT_DESCRIPTION errors.

According to BEP029, _motion.tsv files are uncompressed and explicitly headerless. However,
the standard _loadTSV function was previously hardcoded to always consume the first row as
headers.

Changes Made

1. src/files/tsv.ts: Introduced a headerless boolean parameter (default false) to
2. _loadTSV and its memoized wrappers. When headerless is true, the parser safely bypasses
   duplicate header checks, auto-generates dummy columns to preserve table structure, and retains
   the first row of numerical data.
3. src/schema/context.ts: Inside BIDSContext.loadColumns(), if the file suffix is
   motion, it now explicitly calls the TSV parser with headerless: true. After structural
   validation, it dynamically clears the parsed columns map for motion files so that
   evalColumns ignores them (mirroring the legacy validator's behavior).
4. src/schema/associations.ts & Tests: Updated existing data loaders and test calls to
   explicitly pass false to maintain standard behavior for all other BIDS TSVs.

Testing

• Ran validation locally against real-world VR tracking datasets containing _motion.tsv
  files; verified the custom column errors disappeared.
• Successfully ran the full test suite (deno task test against the bids-examples submodule) with 349 passing tests and 0 failures.

[codecov]

Codecov Report

❌ Patch coverage is 86.36364% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 87.41%. Comparing base (549f03c) to head (b6566ff).
⚠️ Report is 2 commits behind head on main.

Files with missing lines	Patch %	Lines
src/schema/associations.ts	0.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #420      +/-   ##
==========================================
+ Coverage   87.38%   87.41%   +0.03%     
==========================================
  Files          65       65              
  Lines        4779     4793      +14     
  Branches      782      787       +5     
==========================================
+ Hits         4176     4190      +14     
  Misses        510      510              
  Partials       93       93              

☔ 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.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[effigies]

Thanks for this! I think the cleaner patch would actually be to the loadTSVGZ immediately above where we make the .pipeThrough(new DecompressionStream('gzip')) call optional.

I would suggest we give it a more general name, e.g., loadHeaderlessTSV, with a compression parameter. You'll also need to extract the headers from the associated channels.tsv file, which will involve adding a name field to:

bids-validator/src/schema/associations.ts

Lines 93 to 104 in f0cf0fc

	channels: async (file: BIDSFile, options: LoadOptions): Promise<Channels> => {
	const columns = await loadTSV(file, options.maxRows)
	.catch((_e) => {
	return new Map()
	})
	return {
	path: file.path,
	type: columns.get('type'),
	short_channel: columns.get('short_channel'),
	sampling_frequency: columns.get('sampling_frequency'),
	}
	},

You can then pass in this.associations.channels.name as the headers.

[effigies]

Here you undo the load. That's not going to work.