NXdetector groups containing path but no data are not saved as image_data

[DanPorter]

The following files contain NXdetector groups but instead of the data field they contain "path":

• bpm | /dls/i16/data/2025/cm40634-15/1113699.nxs
• cam3 | /dls/i16/data/2025/cm40634-15/1113613.nxs
• cam4 | /dls/i16/data/2025/cm40634-15/1113615.nxs

These detectors should appear in the image_data group.

[DanPorter]

Changes could be made here to incorporate NXdetector/path:

hdfmap/src/hdfmap/nexus.py

Lines 323 to 370 in 51b0b88

	def generate_image_data_from_nxdetector(self):
	"""
	find the NXdetector group and assign the image data
	Must be called after the scannables have been defined as the scan shape is required
	"""
	self.image_data = {}
	image_ndim = len(self.scannables_shape()) + 2 if self.scannables else 3
	if NX_DETECTOR in self.classes:
	group_paths = set(self.classes[NX_DETECTOR])
	elif NX_DATA in self.classes:
	# if no detectors, check for NXdata->dataset with > 2 dimensions
	group_paths = set(self.classes[NX_DATA])
	else:
	group_paths = []
	for group_path in group_paths:
	detector_name = generate_identifier(group_path)
	# detector data is stored in NXdata in dataset 'data'
	data_path = build_hdf_path(group_path, NX_DETECTOR_DATA)
	image_data_path = build_hdf_path(group_path, NX_IMAGE_DATA)
	logger.debug(f"Looking for image_data at: '{data_path}' or '{image_data_path}'")
	if data_path in self.datasets and is_image(self.datasets[data_path].shape, image_ndim):
	logger.info(f"Adding image_data ['{detector_name}'] = '{data_path}'")
	self.image_data[detector_name] = data_path
	self.arrays[detector_name] = data_path
	# also save image_data if available
	if image_data_path in self.datasets:
	detector_name = f"{detector_name}_image_list"
	logger.info(f"Adding image_data ['{detector_name}'] = '{image_data_path}'")
	self.image_data[detector_name] = image_data_path
	self.arrays[detector_name] = image_data_path
	elif image_data_path in self.datasets:
	logger.info(f"Adding image_data ['{detector_name}'] = '{image_data_path}'")
	self.image_data[detector_name] = image_data_path
	self.arrays[detector_name] = image_data_path
	else:
	# Use first dataset with > 2 dimensions
	image_dataset = next((
	path for name in self.get_group_datasets(group_path)
	if is_image(self.datasets[path := build_hdf_path(group_path, name)].shape, image_ndim)
	), False)
	if image_dataset:
	logger.info(f"Adding image_data ['{detector_name}'] = '{image_dataset}'")
	self.image_data[detector_name] = image_dataset
	self.arrays[detector_name] = image_dataset
	if not self.image_data:
	logger.info("No NXdetector image found, image_data not populated.")

However, NXdetector/path would be a numeric array with the same shape as the scan and therefore 1D+, this makes it difficult to discriminate between this dataset and others in the file.

Also, because this is just a list of numbers, there is nothing in the in the file to point to the actual directory of the files.

In this case:

• bpm | /dls/i16/data/2025/cm40634-15/1113699.nxs -> cm40634-15/1113699-bpm1-files/00001.tif
  

• cam3 | /dls/i16/data/2025/cm40634-15/1113613.nxs  -> cm40634-15/1113699-cam3-files/00001.tif
  

• cam4 | /dls/i16/data/2025/cm40634-15/1113615.nxs  -> cm40634-15/1113699-cam4-files/00001.tif
  

[DanPorter]

Changes made in #39
image_data now contains NXdetector/path, this will return a value that points to the file number of a tif file. The file path of this tif file needs to be generated elsewhere.