Fix retry decorators to re-raise last exception (#110)

[Mohitlikestocode]

Summary

Fixes an edge case in the retry decorators where, when all retry attempts fail, the decorator raised a generic
MemoriError("Retry attempts exhausted") instead of re-raising the original (last) exception. This masked the
real error and made debugging difficult.

Changes

• memori/utils/helpers.py
  • Both retry_on_exception and async_retry_on_exception now re-raise the last exception (preserving traceback)
    if present; MemoriError is only raised as a fallback if no exception was captured.
• tests/test_retry_helpers.py (new)
  • Added tests for both sync and async retry decorators to verify the original exception is re-raised and retry counts are correct.
• CHANGELOG.md
  • Added an "Unreleased" note describing the bugfix.

Motivation and Context

Fixes #110 (Hacktoberfest) – consumers could not see the real exception causing retry failures because it was replaced
by a generic MemoriError. With this change, the original exception (and its traceback) will be surfaced.

How to Test

1. Run the test suite (recommended):

python -m venv .venv
.\\.venv\\Scripts\\Activate.ps1
pip install -r requirements-dev.txt
pip install pytest
pytest -q