[Gotthard2][Correct] New paramater to reverse second GH2 25um module

Description

A new parameter to reverse the second(slave) module based on reverseSlaveReadOutMode control parameter.

Usually, this is not expected to be needed, but to avoid any unexpected detector read out configuration I have added this check to reverse the second module before interleaving.

https://git.xfel.eu/calibration/pycalibration/-/issues/94

How Has This Been Tested?

reverse_second_module.pdf

Relevant Documents (optional)

Types of changes

• New feature (non-breaking change which adds functionality)

Checklist:

Reviewers

@schmidtp @mramilli

notebooks/Gotthard2/Correction_Gotthard2_NBC.ipynb

 %% Cell type:markdown id:bed7bd15-21d9-4735-82c1-c27c1a5e3346 tags:
 
 # Gotthard2 Offline Correction
 
 Author: European XFEL Detector Group, Version: 1.0
 
 Offline Correction for Gotthard2 Detector.
 
 This notebook is able to correct 25um and 50um GH2 detectors using the same correction steps:
 - Convert 12bit raw data into 10bit, offset subtraction, then multiply with gain constant.
 
 | Correction | constants   | boolean to enable/disable   |
 |------------|-------------|-----------------------------|
 |   12bit to 10bit  | `LUTGotthard2` |  |
 |   Offset  | `OffsetGotthard2`|`offset_correction`|
 |   Relative gain  | `RelativeGainGotthard2` + `BadPixelsFFGotthard2` |`gain_correction`|
 
 Beside the corrected data, a mask is stored using the badpixels constant of the same parameter conditions and time.
 - `BadPixelsDarkGotthard2`
 - `BadPixelsFFGotthard2`, if relative gain correction is requested.
 
 The correction is done per sequence file. If all selected sequence files have no images to correct the notebook will fail.
 The same result would be reached in case the needed dark calibration constants were not retrieved for all modules and `offset_correction` is True.
 In case one of the gain constants were not retrieved `gain_correction` is switched to False and gain correction is disabled.
 
 The `data` datasets stored in the RECEIVER source along with the corrected image (`adc`) and `mask` are:
 
   - `gain`
 
   - `bunchId`
 
   - `memoryCell`
 
   - `frameNumber`
 
   - `timestamp`
 
   - `trainId`
 
 %% Cell type:code id:570322ed-f611-4fd1-b2ec-c12c13d55843 tags:
 
 ``` python
 in_folder = "/gpfs/exfel/exp/DETLAB/202330/p900326/raw"  # the folder to read data from, required
 out_folder = "/gpfs/exfel/data/scratch/ahmedk/test/gotthard2"  # the folder to output to, required
 metadata_folder = ""  # Directory containing calibration_metadata.yml when run by xfel-calibrate
 run = 20  # run to process, required
 sequences = [-1]  # sequences to correct, set to [-1] for all, range allowed
 sequences_per_node = 1  # number of sequence files per node if notebook executed through xfel-calibrate, set to 0 to not run SLURM parallel
 
 # Parameters used to access raw data.
 karabo_id = "DETLAB_25UM_GH2"  # karabo prefix of Gotthard-II devices
 karabo_da = [""]  # data aggregators
 receiver_template = "RECEIVER{}"  # receiver template used to read INSTRUMENT keys.
 receiver_affixes = [""]  # The affix to format into the receiver template to be able to load the correct receiver name from the data.
 control_template = "CONTROL"  # control template used to read CONTROL keys.
 ctrl_source_template = "{}/DET/{}"  # template for control source name (filled with karabo_id_control)
 karabo_id_control = ""  # Control karabo ID. Set to empty string to use the karabo-id
 corr_source_template = "{}/CORR/{}:daqOutput"  # Correction data source template. filled with karabo_id and correction receiver
 corr_receiver = ""  # The receiver name of the corrected data. Leave empty for using the same receiver name for the 50um GH2 or the first(Master) receiver for the 25um GH2.
 
 # Parameters for calibration database.
 cal_db_interface = "tcp://max-exfl-cal001:8016#8025"  # the database interface to use.
 cal_db_timeout = 180000  # timeout on caldb requests.
 creation_time = ""  # To overwrite the measured creation_time. Required Format: YYYY-MM-DD HR:MN:SC e.g. "2022-06-28 13:00:00"
 
 # Parameters affecting corrected data.
 constants_file = ""  # Use constants in given constant file path. /gpfs/exfel/data/scratch/ahmedk/dont_remove/gotthard2/constants/calibration_constants_GH2.h5
 offset_correction = True  # apply offset correction. This can be disabled to only apply LUT or apply LUT and gain correction for non-linear differential results.
 gain_correction = True  # apply gain correction.
 chunks_data = 1  # HDF chunk size for pixel data in number of frames.
 
 # Parameter conditions.
 bias_voltage = -1  # Detector bias voltage, set to -1 to use value in raw file.
 exposure_time = -1.  # Detector exposure time, set to -1 to use value in raw file.
 exposure_period = -1.  # Detector exposure period, set to -1 to use value in raw file.
 acquisition_rate = -1.  # Detector acquisition rate (1.1/4.5), set to -1 to use value in raw file.
 single_photon = -1  # Detector single photon mode (High/Low CDS), set to -1 to use value in raw file.
+reverse_second_module = False  # Reverse 25um GH2 second module before interleaving. set to -1 to use value in raw file reverse based on `CTRL/reverseSlaveReadOutMode`'s value.
 
 # Parameters for plotting
 skip_plots = False  # exit after writing corrected files
 pulse_idx_preview = 3  # pulse index to preview. The following even/odd pulse index is used for preview. # TODO: update to pulseId preview.
 
 
 def balance_sequences(in_folder, run, sequences, sequences_per_node, karabo_da):
     from xfel_calibrate.calibrate import balance_sequences as bs
     return bs(in_folder, run, sequences, sequences_per_node, karabo_da)
 ```
 
 %% Cell type:code id:6e9730d8-3908-41d7-abe2-d78e046d5de2 tags:
 
 ``` python
 import warnings
 from logging import warning
 
 import h5py
 import pasha as psh
 import numpy as np
 import matplotlib.pyplot as plt
 from IPython.display import Markdown, display
 from extra_data import RunDirectory, H5File
 from pathlib import Path
 
 import cal_tools.restful_config as rest_cfg
 from cal_tools.calcat_interface import CalCatError, GOTTHARD2_CalibrationData
 from cal_tools.files import DataFile
 from cal_tools.gotthard2 import gotthard2algs, gotthard2lib
 from cal_tools.step_timing import StepTimer
 from cal_tools.tools import (
     calcat_creation_time,
     write_constants_fragment,
     map_seq_files,
 )
 from XFELDetAna.plotting.heatmap import heatmapPlot
 
 warnings.filterwarnings('ignore')
 
 %matplotlib inline
 ```
 
 %% Cell type:code id:d7c02c48-4429-42ea-a42e-de45366d7fa3 tags:
 
 ``` python
 in_folder = Path(in_folder)
 run_folder = in_folder / f"r{run:04d}"
 out_folder = Path(out_folder)
 out_folder.mkdir(parents=True, exist_ok=True)
 
 if not karabo_id_control:
     karabo_id_control = karabo_id
 
 ctrl_src = ctrl_source_template.format(karabo_id_control, control_template)
 
 # Run's creation time:
 creation_time = calcat_creation_time(in_folder, run, creation_time)
 print(f"Creation time: {creation_time}")
 ```
 
 %% Cell type:code id:f9a8d1eb-ce6a-4ed0-abf4-4a6029734672 tags:
 
 ``` python
 step_timer = StepTimer()
 ```
 
 %% Cell type:code id:892172d8 tags:
 
 ``` python
 run_dc = RunDirectory(run_folder)
 
 # Read slow data
 g2ctrl = gotthard2lib.Gotthard2Ctrl(run_dc=run_dc, ctrl_src=ctrl_src)
 
 if bias_voltage == -1:
     bias_voltage = g2ctrl.get_bias_voltage()
 if exposure_time == -1:
     exposure_time = g2ctrl.get_exposure_time()
 if exposure_period == -1:
     exposure_period = g2ctrl.get_exposure_period()
 if acquisition_rate == -1:
     acquisition_rate = g2ctrl.get_acquisition_rate()
 if single_photon == -1:
     single_photon = g2ctrl.get_single_photon()
 
 gh2_detector = g2ctrl.get_det_type()
-
+if reverse_second_module and gh2_detector == "25um":
+    reverse_second_module = not g2ctrl.second_module_reversed()
+    print(
+        "Second module is not reversed. "
+        "Reversing second module before interleaving.")
 print("Bias Voltage:", bias_voltage)
 print("Exposure Time:", exposure_time)
 print("Exposure Period:", exposure_period)
 print("Acquisition Rate:", acquisition_rate)
 print("Single Photon:", single_photon)
 print(f"Processing {gh2_detector} Gotthard2.")
 ```
 
 %% Cell type:code id:21a8953a-8c76-475e-8f4f-b201cc25c159 tags:
 
 ``` python
 # GH2 calibration data object.
 g2_cal = GOTTHARD2_CalibrationData(
     detector_name=karabo_id,
     sensor_bias_voltage=bias_voltage,
     exposure_time=exposure_time,
     exposure_period=exposure_period,
     acquisition_rate=acquisition_rate,
     single_photon=single_photon,
     event_at=creation_time,
     client=rest_cfg.calibration_client(),
 )
 
 da_to_pdu = None
 # Keep as long as it is essential to correct
 # RAW data (FXE p003225) before the data mapping was added to CALCAT.
 try:  # in case local constants are used with old RAW data. This can be removed in the future.
     da_to_pdu = g2_cal.mod_to_pdu
 except CalCatError as e:
     print(e)
     db_modules = [None] * len(karabo_da)
 
 if da_to_pdu:
     if karabo_da == [""]:
         karabo_da = sorted(da_to_pdu.keys())
     else:
         # Exclude non selected DA from processing.
         karabo_da = [da for da in karabo_da if da in da_to_pdu]
 
     db_modules = [da_to_pdu[da] for da in karabo_da]
 
 print(f"Process modules: {db_modules} for run {run}")
 
 # Create the correction receiver name.
 receiver_names = [f"*{receiver_template.format(x)}*" for x in receiver_affixes]
 data_sources = list(run_dc.select(receiver_names).all_sources)
 
 if not corr_receiver:
     # This part assumes this data_source structure: '{karabo_id}/DET/{receiver_name}:{output_channel}'
     if gh2_detector == "25um":  # For 25um use virtual karabo_das for CALCAT data mapping.
         corr_receiver = data_sources[0].split("/")[-1].split(":")[0][:-2]
     else:
         corr_receiver = data_sources[0].split("/")[-1].split(":")[0]
     print(f"Using {corr_receiver} as a receiver name for the corrected data.")
 ```
 
 %% Cell type:code id:2551b923 tags:
 
 ``` python
 # Check the available trains to correct.
 total_trains = len(RunDirectory(run_folder).select(data_sources, require_all=True).train_ids)
 if total_trains:
     print(f"Correcting {total_trains}.")
 else:
     raise ValueError(f"No trains to correct for run {run}.")
 ```
 
 %% Cell type:markdown id:8c852392-bb19-4c40-b2ce-3b787538a92d tags:
 
 ### Retrieving calibration constants
 
 %% Cell type:code id:5717d722 tags:
 
 ``` python
 # Used for old FXE (p003225) runs before adding Gotthard2 to CALCAT
 const_data = dict()
 
 if constants_file:
     for mod in karabo_da:
         const_data[mod] = dict()
         # load constants temporarily using defined local paths.
         with h5py.File(constants_file, "r") as cfile:
             const_data[mod]["LUTGotthard2"] = cfile["LUT"][()]
             const_data[mod]["OffsetGotthard2"] = cfile["offset_map"][()].astype(np.float32)
             const_data[mod]["RelativeGainGotthard2"] = cfile["gain_map"][()].astype(np.float32)
             const_data[mod]["Mask"] = cfile["bpix_ff"][()].astype(np.uint32)
 else:
     constant_names = ["LUTGotthard2", "OffsetGotthard2", "BadPixelsDarkGotthard2"]
     if gain_correction:
         constant_names += ["RelativeGainGotthard2", "BadPixelsFFGotthard2"]
 
     g2_metadata = g2_cal.metadata(calibrations=constant_names)
     # Display retrieved calibration constants timestamps
     g2_cal.display_markdown_retrieved_constants(metadata=g2_metadata)
 
     # Validate the constants availability and raise/warn correspondingly.
     for mod, calibrations in g2_metadata.items():
 
         dark_constants = {"LUTGotthard2"}
         if offset_correction:
             dark_constants |= {"OffsetGotthard2", "BadPixelsDarkGotthard2"}
 
         missing_dark_constants = dark_constants - set(calibrations)
         if missing_dark_constants:
             karabo_da.remove(mod)
             warning(f"Dark constants {missing_dark_constants} are not available to correct {mod}.")  # noqa
 
         missing_gain_constants = {
             "BadPixelsFFGotthard2", "RelativeGainGotthard2"} - set(calibrations)
         if gain_correction and missing_gain_constants:
             warning(f"Gain constants {missing_gain_constants} are not retrieved for mod {mod}.")
 
 if not karabo_da:
     raise ValueError("Dark constants are not available for all modules.")
 ```
 
 %% Cell type:code id:ac1cdec5 tags:
 
 ``` python
 # Record constant details in YAML metadata.
 write_constants_fragment(
     out_folder=(metadata_folder or out_folder),
     det_metadata=g2_metadata,
     caldb_root=g2_cal.caldb_root)
 
 # Load constants data for all constants.
 const_data = g2_cal.ndarray_map(metadata=g2_metadata)
 
 # Prepare constant arrays.
 if not constants_file:
     for mod in karabo_da:
         # Create the mask array.
         bpix = const_data[mod].get("BadPixelsDarkGotthard2")
         if bpix is None:
             bpix = np.zeros((1280, 2, 3), dtype=np.uint32)
         if const_data[mod].get("BadPixelsFFGotthard2") is not None:
             bpix |= const_data[mod]["BadPixelsFFGotthard2"]
         const_data[mod]["Mask"] = bpix
 
         # Prepare empty arrays for missing constants.
         if const_data[mod].get("OffsetGotthard2") is None:
             const_data[mod]["OffsetGotthard2"] = np.zeros(
                 (1280, 2, 3), dtype=np.float32)
 
         if const_data[mod].get("RelativeGainGotthard2") is None:
             const_data[mod]["RelativeGainGotthard2"] = np.ones(
                 (1280, 2, 3), dtype=np.float32)
         const_data[mod]["RelativeGainGotthard2"] = const_data[mod]["RelativeGainGotthard2"].astype(  # noqa
             np.float32, copy=False)  # Old gain constants are not float32.
 ```
 
 %% Cell type:code id:2c7dd0bb tags:
 
 ``` python
 file_da = list({kda.split('/')[0] for kda in karabo_da})
 mapped_files, total_files = map_seq_files(
     run_folder,
     file_da,
     sequences,
 )
 # This notebook doesn't account for processing more
 # than one file data aggregator.
 seq_files = mapped_files[file_da[0]]
 
 if not len(seq_files):
     raise IndexError(
         "No sequence files available to correct for the selected sequences and karabo_da.")
 print(f"Processing a total of {total_files} sequence files")
 ```
 
 %% Cell type:code id:23fcf7f4-351a-4df7-8829-d8497d94fecc tags:
 
 ``` python
 context = psh.ProcessContext(num_workers=23)
 ```
 
 %% Cell type:code id:daecd662-26d2-4cb8-aa70-383a579cf9f9 tags:
 
 ``` python
 def correct_train(wid, index, d):
     g = gain[index]
     gotthard2algs.convert_to_10bit(d, const_data[mod]["LUTGotthard2"], data_corr[index, ...])
     gotthard2algs.correct_train(
         data_corr[index, ...],
         mask[index, ...],
         g,
         const_data[mod]["OffsetGotthard2"].astype(np.float32),  # PSI map is in f8
         const_data[mod]["RelativeGainGotthard2"],
         const_data[mod]["Mask"],
         apply_offset=offset_correction,
         apply_gain=gain_correction,
     )
 ```
 
 %% Cell type:code id:f88c1aa6-a735-4b72-adce-b30162f5daea tags:
 
 ``` python
 corr_data_source = corr_source_template.format(karabo_id, corr_receiver)
 
 for raw_file in seq_files:
 
     out_file = out_folder / raw_file.name.replace("RAW", "CORR")
     # Select module INSTRUMENT sources and deselect empty trains.
     dc = H5File(raw_file).select(data_sources, require_all=True)
     n_trains = len(dc.train_ids)
 
     if n_trains == 0:
         warning(f"Skipping correction. No trains to correct for this sequence file: {raw_file}.")
         continue
     else:
         print(f"Correcting {n_trains} for {raw_file}.")
 
     # Initialize GH2 data and gain arrays to store in corrected files.
     if gh2_detector == "25um":
         data_stored = np.zeros((dc[data_sources[0], "data.adc"].shape[:2] + (1280 * 2,)), dtype=np.float32)
         gain_stored = np.zeros((dc[data_sources[0], "data.adc"].shape[:2] + (1280 * 2,)), dtype=np.uint8)
     else:
         data_stored = None
         gain_stored = None
 
     for i, (src, mod) in enumerate(zip(data_sources, karabo_da)):
         step_timer.start()
         print(f"Correcting {src} for {raw_file}")
 
         data = dc[src, "data.adc"].ndarray()
         gain = dc[src, "data.gain"].ndarray()
         step_timer.done_step("Preparing raw data")
         dshape = data.shape
 
         step_timer.start()
 
         # Allocate shared arrays.
         data_corr = context.alloc(shape=dshape, dtype=np.float32)
         mask = context.alloc(shape=dshape, dtype=np.uint32)
         context.map(correct_train, data)
         step_timer.done_step(f"Correcting one receiver in one sequence file")
 
         step_timer.start()
 
         # Provided PSI gain map has 0 values. Set inf values to nan.
         # TODO: This can maybe be removed after creating XFEL gain maps.?
         data_corr[np.isinf(data_corr)] = np.nan
 
         # Create CORR files and add corrected data sections.
         image_counts = dc[src, "data.adc"].data_counts(labelled=False)
+        if reverse_second_module and i == 1:
+            data_corr = np.flip(data_corr, axis=-1)
+            gain = np.flip(gain, axis=-1)
 
         if gh2_detector == "25um":
             data_stored[..., i::2] = data_corr.copy()
             gain_stored[..., i::2] = gain.copy()
         else:  # "50um"
             data_stored = data_corr
             gain_stored = gain
 
     with DataFile(out_file, "w") as ofile:
         # Create INDEX datasets.
         ofile.create_index(dc.train_ids, from_file=dc.files[0])
         ofile.create_metadata(
             like=dc,
             sequence=dc.run_metadata()["sequenceNumber"],
             instrument_channels=(f"{corr_data_source}/data",)
         )
 
         # Create Instrument section to later add corrected datasets.
         outp_source = ofile.create_instrument_source(corr_data_source)
 
         # Create count/first datasets at INDEX source.
         outp_source.create_index(data=image_counts)
 
         # Store uncorrected trainId in the corrected file.
         outp_source.create_key(
                 f"data.trainId", data=dc.train_ids,
                 chunks=min(50, len(dc.train_ids))
             )
 
         # Create datasets with the available corrected data
         for field_name, field_data in {
             "adc": data_stored,
             "gain": gain_stored,
         }.items():
             outp_source.create_key(
                 f"data.{field_name}", data=field_data,
                 chunks=((chunks_data,) + data_corr.shape[1:])
         )
 
         # For GH2 25um, the data of the second receiver is
         # stored in the corrected file.
         for field in ["bunchId", "memoryCell", "frameNumber", "timestamp"]:
                 outp_source.create_key(
                     f"data.{field}", data=dc[src, f"data.{field}"].ndarray(),
                     chunks=(chunks_data, data_corr.shape[1])
             )
         outp_source.create_compressed_key(f"data.mask", data=mask)
         step_timer.done_step("Storing data")
 ```
 
 %% Cell type:code id:94b8e4d2-9f8c-4c23-a509-39238dd8435c tags:
 
 ``` python
 print(f"Total processing time {step_timer.timespan():.01f} s")
 step_timer.print_summary()
 ```
 
 %% Cell type:code id:0ccc7f7e-2a3f-4ac0-b854-7d505410d2fd tags:
 
 ``` python
 if skip_plots:
     print("Skipping plots")
     import sys
 
     sys.exit(0)
 ```
 
 %% Cell type:code id:ff203f77-3811-46f3-bf7d-226d2dcab13f tags:
 
 ``` python
 mod_dcs = {}
 first_seq_raw = seq_files[0]
 first_seq_corr = out_folder / first_seq_raw.name.replace("RAW", "CORR")
 mod_dcs[corr_data_source] = {}
 with H5File(first_seq_corr) as out_dc:
     tid, mod_dcs[corr_data_source]["train_corr_data"] = next(
         out_dc[corr_data_source, "data.adc"].trains()
     )
 
 if gh2_detector == "25um":
     mod_dcs[corr_data_source]["train_raw_data"] = np.zeros((data_corr.shape[1], 1280 * 2), dtype=np.float32)
     mod_dcs[corr_data_source]["train_raw_gain"] = np.zeros((data_corr.shape[1], 1280 * 2), dtype=np.uint8)
 
 for i, src in enumerate(data_sources):
     with H5File(first_seq_raw) as in_dc:
         train_dict = in_dc.train_from_id(tid)[1][src]
         if gh2_detector == "25um":
-            mod_dcs[corr_data_source]["train_raw_data"][..., i::2] = train_dict["data.adc"]
-            mod_dcs[corr_data_source]["train_raw_gain"][..., i::2] = train_dict["data.gain"]
+            if reverse_second_module and i == 1:
+                data = np.flip(train_dict["data.adc"], axis=-1)
+                gain = np.flip(train_dict["data.gain"], axis=-1)
+            else:
+                data = train_dict["data.adc"]
+                gain = train_dict["data.gain"]
+            mod_dcs[corr_data_source]["train_raw_data"][..., i::2] = data
+            mod_dcs[corr_data_source]["train_raw_gain"][..., i::2] = gain
         else:
             mod_dcs[corr_data_source]["train_raw_data"] = train_dict["data.adc"]
             mod_dcs[corr_data_source]["train_raw_gain"] = train_dict["data.gain"]
 ```
 
 %% Cell type:code id:1b379438-eb1d-42b2-ac83-eb8cf88c46db tags:
 
 ``` python
 display(Markdown("### Mean RAW and CORRECTED across pulses for one train:"))
 display(Markdown(f"Train: {tid}"))
 
 if gh2_detector == "50um":
     title = f"{{}} data for {karabo_da} ({db_modules})"
 else:
     title = f"Interleaved {{}} data for {karabo_da} ({db_modules})"
 
 step_timer.start()
 
 fig, ax = plt.subplots(figsize=(15, 15))
 raw_data = mod_dcs[corr_data_source]["train_raw_data"]
 im = ax.plot(np.mean(raw_data, axis=0))
 ax.set_title(title.format("RAW"), fontsize=20)
 ax.set_xlabel("Strip #", size=20)
 ax.set_ylabel("12-bit ADC output", size=20)
 plt.xticks(fontsize=20)
 plt.yticks(fontsize=20)
 pass
 
 fig, ax = plt.subplots(figsize=(15, 15))
 corr_data = mod_dcs[corr_data_source]["train_corr_data"]
 im = ax.plot(np.mean(corr_data, axis=0))
 ax.set_title(title.format("CORRECTED"), fontsize=20)
 ax.set_xlabel("Strip #", size=20)
 ax.set_ylabel("10-bit KeV. output", size=20)
 plt.xticks(fontsize=20)
 plt.yticks(fontsize=20)
 pass
 step_timer.done_step("Plotting mean data")
 ```
 
 %% Cell type:code id:58a6a276 tags:
 
 ``` python
 display(Markdown(f"### RAW and CORRECTED strips across pulses for train {tid}"))
 
 step_timer.start()
 for plt_data, dname in zip(
     ["train_raw_data", "train_corr_data"], ["RAW", "CORRECTED"]
 ):
     fig, ax = plt.subplots(figsize=(15, 15))
     plt.rcParams.update({"font.size": 20})
 
     heatmapPlot(
         mod_dcs[corr_data_source][plt_data],
         y_label="Pulses",
         x_label="Strips",
         title=title.format(dname),
         use_axis=ax,
         cb_pad=0.8,
     )
     pass
 step_timer.done_step("Plotting RAW and CORRECTED data for one train")
 ```
 
 %% Cell type:code id:cd8f5e08-fcee-4bff-ba63-6452b3d892a2 tags:
 
 ``` python
 # Validate given "pulse_idx_preview"
 
 if pulse_idx_preview + 1 > data.shape[1]:
     print(
         f"WARNING: selected pulse_idx_preview {pulse_idx_preview} is not available in data."
         " Previewing 1st pulse."
     )
     pulse_idx_preview = 1
 
 if data.shape[1] == 1:
     odd_pulse = 1
     even_pulse = None
 else:
     odd_pulse = pulse_idx_preview if pulse_idx_preview % 2 else pulse_idx_preview + 1
     even_pulse = (
         pulse_idx_preview if not (pulse_idx_preview % 2) else pulse_idx_preview + 1
     )
 
 if pulse_idx_preview + 1 > data.shape[1]:
     pulse_idx_preview = 1
     if data.shape[1] > 1:
         pulse_idx_preview = 2
 ```
 
 %% Cell type:code id:e5f0d4d8-e32c-4f2c-8469-4ebbfd3f644c tags:
 
 ``` python
 display(Markdown("### RAW and CORRECTED even/odd pulses for one train:"))
 display(Markdown(f"Train: {tid}"))
 fig, ax = plt.subplots(figsize=(15, 15))
 raw_data = mod_dcs[corr_data_source]["train_raw_data"]
 corr_data = mod_dcs[corr_data_source]["train_corr_data"]
 
 ax.plot(raw_data[odd_pulse], label=f"Odd Pulse {odd_pulse}")
 if even_pulse:
     ax.plot(raw_data[even_pulse], label=f"Even Pulse {even_pulse}")
 
 ax.set_title(title.format("RAW"), fontsize=20)
 ax.set_xlabel("Strip #", size=20)
 ax.set_ylabel("12-bit ADC RAW", size=20)
 plt.xticks(fontsize=20)
 plt.yticks(fontsize=20)
 ax.legend()
 pass
 
 fig, ax = plt.subplots(figsize=(15, 15))
 ax.plot(corr_data[odd_pulse], label=f"Odd Pulse {odd_pulse}")
 if even_pulse:
     ax.plot(corr_data[even_pulse], label=f"Even Pulse {even_pulse}")
 ax.set_title(title.format("CORRECTED"), fontsize=20)
 ax.set_xlabel("Strip #", size=20)
 ax.set_ylabel("10-bit KeV CORRECTED", size=20)
 plt.xticks(fontsize=20)
 plt.yticks(fontsize=20)
 ax.legend()
 pass
 step_timer.done_step("Plotting RAW and CORRECTED odd/even pulses.")
 ```