HEED loader hardcodes HEED path but README uses heed/

[JoeScharpf]

The README tells users to put HEED data in data/heed/ and run with --base heed, but src/ecg_datasets/base/heed/heed.py hardcodes DATA_DIR/HEED/... for both glob and open_ecg. On Linux (case-sensitive filesystems) those are different paths, so preprocessing fails even when setup follows the docs.

Fix: use f"{DATA_DIR}/{self.args.base}" consistently (e.g. self.data_root), matching how the CSV path already uses args.base.

Relevant lines in heed.py: glob.glob(f"{DATA_DIR}/HEED/I*") and file_path = f"{DATA_DIR}/HEED/{row['path']}".

[willxxy]

This is not too important. The mismatch comes from our setup having the data in HEED instead of heed. Not consistent but the error that is shown when the folder name mismatches should be clear enough for users to figure it out. We will note this during the next big refactor.

[willxxy]

Reopening for future consideration and users.