fixed split path directory bug

[jnsLs]

fix for issue #793

[Copilot]

Pull request overview

This PR addresses issue #793 by ensuring that training runs no longer share a single persisted dataset split file just because they point at the same run.data_dir. This restores the prior behavior where separate runs can produce separate random splits by default, while also adding basic visibility into whether a split is being loaded or created.

Changes:

• Add informational logging when a split is loaded from disk or newly written.
• Change the default split_file location in configs/data/custom.yaml to be per-run (${run.path}/${run.id}) rather than shared via ${run.data_dir}.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
src/schnetpack/data/datamodule.py	Adds log.info(...) statements to make persisted split load/write behavior visible at runtime.
src/schnetpack/configs/data/custom.yaml	Moves split_file from a shared data_dir location to the per-run output directory to avoid unintended split reuse across runs.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.