fix: preserve remainder rows across worker batches

[sergioballesteros-vd]

Summary

This PR fixes row loss in account and transaction generation when the requested totals do not divide evenly across workers * batches_per_worker.

The generator now splits work with exact shard planning, so every requested row is produced and no remainder is dropped.

Type of change

• fix — bug fix
• test — adding or updating tests

Linked issues

• N/A

Checklist

• Commit messages follow Conventional Commits
• Tests added or updated (pytest tests/ -v --cov=gen_fraud_graph)
• Coverage stays at or above 80%
• Lint/format/type pass locally (ruff check . && black --check . && mypy src/)
• Documentation updated (README, CHANGELOG, docstrings) if behaviour changed
• I have signed the CLA (the bot will prompt me if not)
• I agree to follow the project's Code of Conduct

Notes for reviewers

The change is intentionally small and localized:

• workload splitting was moved to an exact divmod-based helper
• the public CLI/API contract remains unchanged
• tests cover both remainder and exact-division cases

[github-actions]

All contributors have signed the CLA ✍️ ✅
_{Posted by the CLA Assistant Lite bot.}

[sergioballesteros-vd]

Closed by request.

[opensource-SantanderAI]

recheck

[opensource-SantanderAI]

Hi @sergioballesteros-vd — thanks for the fix! Before we can review/merge, please sign the CLA by posting this comment:

I have read the CLA Document and I hereby sign the CLA

The CLA Assistant check will go green automatically once signed. 🙌

[sergioballesteros-vd]

I have read the CLA Document and I hereby sign the CLA