Consolidate eicm.preprocessing.cellvoyager with hcs.cellvoyager

[imagejan]

Currently, we have:

For metadata

hcs

faim-ipa/src/faim_ipa/hcs/cellvoyager/acquisition.py

Lines 238 to 277 in b65f215

	def _parse_metadata(self) -> pd.DataFrame:
	mrf_file = join(self._acquisition_dir, "MeasurementDetail.mrf")
	if not exists(mrf_file):
	msg = f"MeasurementDetail.mrf not found in: {self._acquisition_dir}"
	raise ValueError(msg)
	mrf_tree = parse(mrf_file)
	mrf_root = mrf_tree.getroot()
	channels = []
	for channel in mrf_root.findall(BTS_NS + "MeasurementChannel"):
	row = {
	key.replace(BTS_NS, ""): value for key, value in channel.attrib.items()
	}
	channels.append(row)
	mes_file = join(
	self._acquisition_dir,
	mrf_root.attrib[BTS_NS + "MeasurementSettingFileName"],
	)
	if not exists(mes_file):
	msg = f"Settings file not found: {mes_file}"
	raise ValueError(msg)
	mes_tree = parse(mes_file)
	mes_root = mes_tree.getroot()
	channel_settings = []
	for channel in mes_root.find(BTS_NS + "ChannelList").findall(
	BTS_NS + "Channel"
	):
	row = {
	key.replace(BTS_NS, ""): value for key, value in channel.attrib.items()
	}
	channel_settings.append(row)
	return pd.merge(
	pd.DataFrame(channels),
	pd.DataFrame(channel_settings),
	left_on="Ch",
	right_on="Ch",
	)

and

faim-ipa/src/faim_ipa/hcs/cellvoyager/acquisition.py

Lines 189 to 211 in b65f215

	def get_channel_metadata(self) -> dict[int, ChannelMetadata]:
	metadata = self._parse_metadata()
	ch_metadata = {}
	for _i, row in metadata.iterrows():
	index = int(row["Ch"]) - 1
	ch_metadata[index] = ChannelMetadata(
	channel_index=index,
	channel_name=row["Ch"],
	display_color=row["Color"],
	spatial_calibration_x=row["HorizontalPixelDimension"],
	spatial_calibration_y=row["VerticalPixelDimension"],
	spatial_calibration_units="um",
	z_spacing=self.get_z_spacing(),
	wavelength=self.__parse_filter_wavelength(row["Acquisition"]),
	exposure_time=row["ExposureTime"],
	exposure_time_unit="ms",
	objective=row["Objective"],
	)
	assert min(ch_metadata.keys()) == 0, "Channel indices must start at 0."
	return ch_metadata

eicm

faim-ipa/src/faim_ipa/eicm/preprocessing/cellvoyager.py

Lines 13 to 61 in b65f215

	def get_metadata(input_dir: str):
	"""
	Extract acquisition date, pixel-size information, acquisition camera
	index, and channel information (laser, filter and objective) from a
	Yokogawa CV7000 or CV8000 acquisition. The information is extracted from
	mse- and mrf-files written by Yokogawa.
	:param input_dir: location of the Yokogawa acquisition
	:return: acquisition_date, pixel_size, pixel_size_unit, channel_information
	"""
	mrf_file = join(input_dir, "MeasurementDetail.mrf")
	mrf_tree = ET.parse(mrf_file)
	mrf_root = mrf_tree.getroot()
	mrf_ns = mrf_root.tag.replace("MeasurementDetail", "")
	date_format_str = mrf_root.attrib[mrf_ns + "BeginTime"]
	date = parser.parse(date_format_str)
	acquisition_date_str = date.strftime("%Y-%m-%d")
	channels = {}
	for child in mrf_root:
	ch = child.get(mrf_ns + "Ch")
	if ch is not None:
	channels[ch] = {
	"pixel_size": child.get(mrf_ns + "HorizontalPixelDimension"),
	"cam_index": child.get(mrf_ns + "CameraNumber"),
	}
	pixel_size = float(mrf_root[1].attrib.get(mrf_ns + "HorizontalPixelDimension"))
	pixel_size_unit = "micron"
	mes_file = glob(join(input_dir, "*.mes"))[0]
	mes_tree = ET.parse(mes_file)
	mes_root = mes_tree.getroot()
	mes_ns = mes_root.tag.replace("MeasurementSetting", "")
	for child in mes_root[2]:
	ch = child.get(mes_ns + "Ch")
	if ch is not None:
	channel_dict = channels[ch]
	channel_dict["objective"] = child.get(mes_ns + "Objective").replace(
	" ", "-"
	)
	channel_dict["filter"] = child.get(mes_ns + "Acquisition").replace("/", "-")
	channel_dict["laser"] = child[0].text
	return acquisition_date_str, pixel_size, pixel_size_unit, channels

For acquisition files

hcs

faim-ipa/src/faim_ipa/hcs/cellvoyager/acquisition.py

Lines 279 to 305 in b65f215

	def _parse_files(self) -> pd.DataFrame:
	mlf_file = join(self._acquisition_dir, "MeasurementData.mlf")
	if not exists(mlf_file):
	msg = f"MeasurementData.mlf not found in: {self._acquisition_dir}"
	raise ValueError(msg)
	mlf_tree = parse(mlf_file)
	mlf_root = mlf_tree.getroot()
	files = []
	for record in mlf_root.findall(BTS_NS + "MeasurementRecord"):
	row = {
	key.replace(BTS_NS, ""): value for key, value in record.attrib.items()
	}
	if row.pop("Type") == "IMG":
	row |= {
	"path": join(self._acquisition_dir, record.text),
	"well": chr(ord("@") + int(row.pop("Row")))
	+ row.pop("Column").zfill(2),
	}
	files.append(row)
	record.clear()
	files = pd.DataFrame(files)
	files["TimePoint"] = files["TimePoint"].astype(int)
	files["ZIndex"] = files["ZIndex"].astype(int)
	return files

eicm

faim-ipa/src/faim_ipa/eicm/preprocessing/cellvoyager.py

Lines 86 to 128 in b65f215

	def create_table(files: List[str], plate_name: str) -> pd.DataFrame:
	"""
	Create table of file-names with columns for plate, well, time-point,
	field, L, action, Z and channel.
	:param files: image file list
	:param plate_name: Name of the plate
	:return: table
	"""
	plate = []
	well = []
	timepoint = []
	field = []
	lines = []
	action = []
	z = []
	channel = []
	path = []
	for file in files:
	p, w, t, f, line, a, z_, c = parse_filename(file, plate_name)
	plate.append(p)
	well.append(w)
	timepoint.append(t)
	field.append(f)
	lines.append(line)
	action.append(a)
	z.append(z_)
	channel.append(c)
	path.append(file)
	return pd.DataFrame(
	{
	"plate": plate,
	"well": well,
	"timepoint": timepoint,
	"field": field,
	"line": lines,
	"action": action,
	"z": z,
	"channel": channel,
	"path": path,
	}
	)

---

We should try to consolidate those two code bases (and maybe even use the metadata model from https://github.com/tlambert03/ome-types to standardize the way we handle metadata, see #198):

• factor out into common public utility methods (e.g. in faim_ipa.hcs.cellvoyager.utils)
• use those public methods from within the StackAcquisition class