Update locomanipulation SDG docs with clarification on converting to lerobot dataset

[jaybdub]

Description

• Documents that convert_dataset.py's <input_dir> must contain only SDG .hdf5 outputs from generate_data.py; mixing in earlier-step datasets (e.g. dataset_annotated_g1_locomanip.hdf5, generated_dataset_g1_locomanip.hdf5) causes the converter to fail.
• Updates the example to mkdir+mv the SDG output into ./datasets/locomanip_sdg/ before running the converter.

Type of change

• Documentation update

Checklist

• I have read and understood the contribution guidelines
• I have run the pre-commit checks with ./isaaclab.sh --format
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• I have updated the changelog and the corresponding version in the extension's config/extension.toml file
• I have added my name to the CONTRIBUTORS.md or my name already exists there

[isaaclab-review-bot]

📚 Documentation Review

Thanks for this documentation update, @jaybdub!

Summary

This PR adds helpful clarification about the convert_dataset.py script's input directory requirements for the locomanipulation SDG workflow.

What Changed

1. Prerequisites section: Now explicitly warns that the input directory must contain only SDG outputs from generate_data.py, not HDF5 files from earlier pipeline steps
2. Conversion section: Updated description emphasizes the "no other HDF5 files" requirement
3. Example: Changed from using ./datasets directly to a two-step workflow:
   • Create a dedicated subdirectory (./datasets/locomanip_sdg)
   • Move only the SDG output there before conversion

Review Notes

✅ Clarity: The warning is clear and the example demonstrates the correct workflow
✅ RST formatting: Looks correct (inline code markers, code blocks)
✅ Consistency: Maintains the existing documentation style
✅ User value: This will help users avoid a confusing error when the converter encounters incompatible HDF5 files

Minor observation: The bolded only and no other HDF5 files effectively draw attention to the constraint without being excessive.

---

🤖 Reviewed by Isaac Lab Review Bot