Data Management

This module provides project- and version-oriented management of NEP datasets. You can organize models under projects, track versions, tag datasets, and open their working directories.

1. Concepts

  • Project: A tree of folders that groups related models. Supports nested hierarchy.

  • Model (Version): A dataset entry under a project. Stores metadata: path/URL, size, energy/force/virial metrics, tags, notes, created time.

  • Tags: Colored labels to classify or filter models. Managed centrally and applied to models.

2. UI Overview

  • Left panel: Project tree. Right panel: Model list of the selected project.

  • Context menus on both panels provide creation, modification, deletion, and utility actions.

3. Project Operations

  • New: Right‑click the project tree → New. Choose parent, name, notes.

  • Modify: Right‑click an item → Modify. Edit name, notes, parent.

  • Delete: Right‑click an item → Delete. Deleting a node removes all descendants.

4. Model (Version) Operations

  • New: Right‑click the model list → New. Fill in:

    • Name, notes, model type

    • Train path (local folder or HTTP URL)

    • Energy/Force/Virial scores (optional metadata)

    • Parent model (to build a version chain)

    • Tags

  • Modify/Delete: Right‑click an item → Modify/Delete.

  • Open Folder: Opens the train_path in your OS file manager or browser (if URL).

6. Database Location

  • The app stores management data in a SQLite database under the user config directory:

    • Windows: C:\Users\<You>\AppData\Local\NepTrainKit\mlpman.db

    • Linux: ~/.config/NepTrainKit/mlpman.db

7. Tips

  • Hierarchies: Use parent/child for simple version trees.

  • External sources: Set train_path to a Git/HTTP URL to reference online data.

  • Bulk review: Use tags to triage candidate datasets before promotion.