fix: validate split_overlap is less than split_length in DocumentSplitter

[i-anubhav-anand]

Related Issues

None — self-found while reviewing DocumentSplitter validation.

Proposed Changes

DocumentSplitter.__init__ validates split_length > 0 and split_overlap >= 0, but not that split_overlap < split_length. When split_overlap >= split_length, the window step (split_length - split_overlap) becomes zero or negative, and _concatenate_units calls more_itertools.windowed(..., step=step), which raises an opaque ValueError: step must be >= 1 deep in the call stack at run time — long after the misconfiguration was introduced.

This adds an early, clear validation at initialization, consistent with the two checks right above it:

if split_overlap >= split_length:
    raise ValueError("split_overlap must be less than split_length.")

How did you test it?

• Added a regression test (test_split_overlap_not_less_than_split_length) covering split_overlap == split_length and split_overlap > split_length. Verified it fails before this change (no error at init) and passes after.
• hatch run test:unit test/components/preprocessors/test_document_splitter.py → 55 passed.
• hatch run test:types (module) and hatch run fmt → clean.

Notes for the reviewer

Behavior change is limited to rejecting a configuration that was already broken (it crashed at run time); valid configs (split_overlap < split_length) are unaffected. Release note added under releasenotes/notes/.

[vercel]

@i-anubhav-anand is attempting to deploy a commit to the deepset Team on Vercel.

A member of the Team first needs to authorize it.

[CLAassistant]

All committers have signed the CLA.

[github-actions]

Coverage report

Click to see where and how coverage changed

File	Statements	Missing	Coverage	Coverage (new stmts)	Lines missing
haystack/components/preprocessors
document_splitter.py
Project Total

This report was generated by python-coverage-comment-action

[anakin87]

Makes sense. Thank you!