[Code scan] Truncate rewritten training JSON after HDF5 conversion

[njzjz]

This issue is from a Codex global scan of the repository.

convert_training_data_to_hdf5() can leave stale bytes at the end of rewritten JSON input files.

Evidence:

dpgen/dpgen/util.py

Lines 119 to 142 in 7af5246

	with open(ii, "r+") as f:
	jinput = json.load(f)
	if "training_data" in jinput["training"]:
	# v2.0
	p_sys = jinput["training"]["training_data"]["systems"]
	else:
	# v1.x
	p_sys = jinput["training"]["systems"]
	for ii, pp in enumerate(p_sys):
	if "#" in pp:
	# HDF5 file
	p1, p2 = pp.split("#")
	ff = os.path.normpath(str((dd / p1).absolute().relative_to(cwd)))
	pp = ff + "#" + p2
	new_pp = os.path.normpath(os.path.relpath(ff, h5_dir)) + p2
	else:
	pp = os.path.normpath(str((dd / pp).absolute().relative_to(cwd)))
	new_pp = os.path.normpath(os.path.relpath(pp, h5_dir))
	p_sys[ii] = (
	os.path.normpath(os.path.relpath(h5_file, dd)) + "#/" + str(new_pp)
	)
	systems.append(pp)
	f.seek(0)
	json.dump(jinput, f, indent=4)

The function opens input JSON files with r+, updates the training system paths, calls f.seek(0), and writes the new JSON with json.dump(). It never calls f.truncate(). If the rewritten JSON is shorter than the original file, trailing bytes from the old content remain and the JSON file can become invalid.

Expected behavior: truncate the file after writing, or write to a temporary file and replace the original atomically.