fix: update meta data before initializing new Document in DocumentSplitter

[nickprock]

Related Issues

• fixes DocumentSplitter updates Document's meta data after initializing the Document #8741

Proposed Changes:

Updated code as illustrated in the issue.

How did you test it?

unit tests

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

[nickprock]

Hi @anakin87 ,
on my local environment the code pass all checks but here there are some files (I haven't worked on they) that don't pass the format checker.
I don't understand.

[anakin87]

Sometimes it's related to linter updates... I'll take a look later...

[coveralls]

Pull Request Test Coverage Report for Build 12856711277

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 91.306%

---

Totals
Change from base Build 12856629462:	0.0%
Covered Lines:	8853
Relevant Lines:	9696

---

💛 - Coveralls

[anakin87]

1. I understood the format problem, that I will fix in another PR (chore: update ruff version in pre-commit hook #8746)
2. I'm having a hard time testing this bugfix: I'm not sure that we are creating duplicate IDs with the current code, so I'll ask @julian-risch to take a look

[julian-risch]

Looks quite good to me already! Thank you @nickprock for contributing to Haystack!
There is a small bug in the code, which is also the reason why @anakin87 couldn't produce any duplicate documents. We're currently re-using meta throughout multiple iterations of the for loop, which we shouldn't. In particular, overwriting meta in meta = deepcopy(meta) is problematic.

If you do something like the following, that problem will be fixed:

copied_meta = deepcopy(meta)
copied_meta["page_number"] = splits_pages[i]
copied_meta["split_id"] = i
copied_meta["split_idx_start"] = split_idx
doc = Document(content=txt, meta=copied_meta)

Here is an example of a unit test you could please add to test/components/preprocessors/test_document_splitter.py

def test_duplicate_pages_get_different_doc_id(self):
    splitter = DocumentSplitter(split_by="page", split_length=1)
    doc1 = Document(content="This is some text.\fThis is some text.\fThis is some text.\fThis is some text.")
    splitter.warm_up()
    result = splitter.run(documents=[doc1])

    assert len({doc.id for doc in result["documents"]}) == 4 

[nickprock]

Thanks @julian-risch , I'll work on it tomorrow.

[nickprock]

Hi, @anakin87 I applied the changes requested by @julian-risch and added the test.
On my local environment everything is fine, here there is always yesterday's error about formatting.

[julian-risch]

LGTM! 👍 Thanks for adding a test too!