feat: add split_by_row feature to CSVDocumentSplitter

[Amnah199]

Related Issues

• Introduce a new splitting mode that treats each row of a CSV as a separate subtable and converts it into Haystack document.
• Original motivation from: feat: csv to document row level conversion #8916

Proposed Changes:

Add a new parameter _split_by_row to CSVDocumentSplitter. When _split_by_row=True, other split arguments won't be considered.

How did you test it?

Added new unit tests

Notes for the reviewer

If this is merged, we need to update the documentation of this component slightly.

Checklist

• I have read the contributors guidelines and the code of conduct
• I have updated the related issue with new insights and changes
• I added unit tests and updated the docstrings
• I've used one of the conventional commit types for my PR title: fix:, feat:, build:, chore:, ci:, docs:, style:, refactor:, perf:, test: and added ! in case the PR includes breaking changes.
• I documented my code
• I ran pre-commit hooks and fixed any issue

[coveralls]

Pull Request Test Coverage Report for Build 13944618786

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• 4 unchanged lines in 1 file lost coverage.
• Overall coverage decreased (-0.002%) to 90.068%

---

Files with Coverage Reduction	New Missed Lines	%
components/preprocessors/csv_document_splitter.py	4	96.26%

Totals
Change from base Build 13920793695:	-0.002%
Covered Lines:	9740
Relevant Lines:	10814

---

💛 - Coveralls

[sjrl]

@Amnah199 thanks for working on this!

I think we should take a slightly different approach to toggle between the two different split modes. Instead of using the boolean split_by_row, I think we should introduce a split_mode variable with a Literal (or Enum) type to explicitly define the two modes. An idea for the two different split mode names could be threshold and row-wise, so something like

from typing import Literal

SplitMode = Literal["threshold", "row-wise"]

• "threshold" → Uses row/column thresholds to determine splits (default behavior).
• "row-wise" → Each row becomes its own sub-table, ignoring thresholds.

I think this would improve readability and makes it immediately clear what the expected behavior is. Let me know what you think!

[sjrl]

@Amnah199 In the test_from_dict_defaults in the tests could we add an assert checking that split_mode is equal to threshold?

[sjrl]

Looks good!