Merge branch 'fix/dssc-cellid-overflow' into 'master'

[DSSC] Allow 900 memory cells for DSSC darks to workaround appearance of cell 810 See merge request !918

Merge branch 'fix/dssc-cellid-overflow' into 'master'
[DSSC] Allow 900 memory cells for DSSC darks to workaround appearance of cell 810 See merge request !918
32e2a3e4 · Philipp Schmidt · 565d48e4 · da688d27 · 32e2a3e4
Commit 32e2a3e4 authored 1 year ago by Philipp Schmidt
--- a/notebooks/DSSC/Characterize_DSSC_Darks_NBC.ipynb
+++ b/notebooks/DSSC/Characterize_DSSC_Darks_NBC.ipynb
@@ -218,7 +218,7 @@
    "            if cells == []:\n",
    "                return\n",
    "            maxcell = np.max(cells)\n",
-    "            options = [100, 200, 400, 500, 600, 700, 800]\n",
+    "            options = [100, 200, 400, 500, 600, 700, 800, 900]\n",
    "            dists = np.array([(o-maxcell) for o in options])\n",
    "            dists[dists<0] = 10000 # assure to always go higher\n",
    "            return options[np.argmin(dists)]\n",

 %% Cell type:markdown id: tags:
 # DSSC Characterize Dark Images #
 Author: S. Hauf, Version: 0.1
 The following code analyzes a set of dark images taken with the DSSC detector to deduce detector offsets and noise. Data for the detector is presented in one run and don't acquire multiple gain stages.
 The notebook explicitely does what pyDetLib provides in its offset calculation method for streaming data.
 %% Cell type:code id: tags:
 ``` python
 cluster_profile = "noDB" # The ipcluster profile to use
 in_folder = "/gpfs/exfel/exp/SQS/202131/p900210/raw" # path to input data, required
 out_folder = "/gpfs/exfel/data/scratch/samartse/data/DSSC" # path to output to, required
 metadata_folder = ""  # Directory containing calibration_metadata.yml when run by xfel-calibrate
 sequences = [0] # sequence files to evaluate.
 modules = [-1]  # modules to run for
 run = 20 #run number in which data was recorded, required
 karabo_id = "SQS_DET_DSSC1M-1" # karabo karabo_id
 karabo_da = ['-1']  # a list of data aggregators names, Default [-1] for selecting all data aggregators
 receiver_id = "{}CH0" # inset for receiver devices
 path_template = 'RAW-R{:04d}-{}-S{:05d}.h5' # the template to use to access data
 h5path = '/INSTRUMENT/{}/DET/{}:xtdf/image' # path in the HDF5 file to images
 h5path_idx = '/INDEX/{}/DET/{}:xtdf/image' # path in the HDF5 file to images
 slow_data_pattern = 'RAW-R{}-DA{}-S00000.h5'
 use_dir_creation_date = True # use the dir creation date for determining the creation time
 cal_db_interface = "tcp://max-exfl-cal001:8020" # the database interface to use
 cal_db_timeout = 3000000 # timeout on caldb requests"
 local_output = True # output constants locally
 db_output = False # output constants to database
 mem_cells = 0 # number of memory cells used, set to 0 to automatically infer
 bias_voltage = 100 # detector bias voltage
 rawversion = 2 # RAW file format version
 thresholds_offset_sigma = 3. # thresholds in terms of n sigma noise for offset deduced bad pixels
 thresholds_offset_hard = [4, 125] # thresholds in absolute ADU terms for offset deduced bad pixels,
 # minimal threshold at 4 is set at hardware level, DSSC full range 0-511
 thresholds_noise_sigma = 3. # thresholds in terms of n sigma noise for offset deduced bad pixels
 thresholds_noise_hard = [0.001, 3] # thresholds in absolute ADU terms for offset deduced bad pixels
 offset_numpy_algorithm = "mean"
 high_res_badpix_3d = False # set this to True if you need high-resolution 3d bad pixel plots. Runtime: ~ 1h
 slow_data_aggregators = [1,1,1,1]  # quadrant/aggregator
 slow_data_path = 'SQS_NQS_DSSC/FPGA/PPT_Q'
 operation_mode = ''  # Detector operation mode, optional
 ```
 %% Cell type:code id: tags:
 ``` python
 import os
 import warnings
 # imports and things that do not usually need to be changed
 from datetime import datetime
 warnings.filterwarnings('ignore')
 from collections import OrderedDict
 import h5py
 import matplotlib
 from ipyparallel import Client
 from IPython.display import Latex, Markdown, display
 matplotlib.use('agg')
 import matplotlib.pyplot as plt
 %matplotlib inline
 import numpy as np
 import tabulate
 import yaml
 from iCalibrationDB import Conditions, Constants, Detectors, Versions
 from cal_tools.dssclib import get_dssc_ctrl_data, get_pulseid_checksum
 from cal_tools.enums import BadPixels
 from cal_tools.plotting import (
    create_constant_overview,
    plot_badpix_3d,
    show_overview,
    show_processed_modules,
 )
 from cal_tools.tools import (
    get_dir_creation_date,
    get_from_db,
    get_notebook_name,
    get_pdu_from_db,
    get_random_db_interface,
    get_report,
    map_gain_stages,
    parse_runs,
    run_prop_seq_from_path,
    save_const_to_h5,
    send_to_db,
 )
 view = Client(profile=cluster_profile)[:]
 view.use_dill()
 # make sure a cluster is running with ipcluster start --n=32, give it a while to start
 h5path = h5path.format(karabo_id, receiver_id)
 h5path_idx = h5path_idx.format(karabo_id, receiver_id)
 gain_names = ['High', 'Medium', 'Low']
 if karabo_da[0] == '-1':
    if modules[0] == -1:
        modules = list(range(16))
    karabo_da = ["DSSC{:02d}".format(i) for i in modules]
 else:
    modules = [int(x[-2:]) for x in karabo_da]
 max_cells = mem_cells
 offset_runs = OrderedDict()
 offset_runs["high"] = run
 creation_time=None
 if use_dir_creation_date:
    creation_time = get_dir_creation_date(in_folder, run)
    print(f"Using {creation_time} as creation time of constant.")
 run, prop, seq = run_prop_seq_from_path(in_folder)
 dinstance = "DSSC1M1"
 print(f"Detector in use is {karabo_id}")
 cal_db_interface = get_random_db_interface(cal_db_interface)
 ```
 %% Cell type:code id: tags:
 ``` python
 print("Parameters are:")
 print(f"Proposal: {prop}")
 print(f"Memory cells: {mem_cells}/{max_cells}")
 print("Runs: {}".format([ v for v in offset_runs.values()]))
 print(f"Sequences: {sequences}")
 print(f"Using DB: {db_output}")
 print(f"Input: {in_folder}")
 print(f"Output: {out_folder}")
 print(f"Bias voltage: {bias_voltage}V")
 file_loc = f'proposal:{prop} runs:{[ v for v in offset_runs.values()][0]}'
 report = get_report(metadata_folder)
 ```
 %% Cell type:markdown id: tags:
 The following lines will create a queue of files which will the be executed module-parallel. Distinguishing between different gains.
 %% Cell type:code id: tags:
 ``` python
 # set everything up filewise
 os.makedirs(out_folder, exist_ok=True)
 gmf = map_gain_stages(in_folder, offset_runs, path_template, karabo_da, sequences)
 gain_mapped_files, total_sequences, total_file_size = gmf
 print(f"Will process a total of {total_sequences} file.")
 ```
 %% Cell type:markdown id: tags:
 ## Calculate Offsets, Noise and Thresholds ##
 The calculation is performed per-pixel and per-memory-cell. Offsets are simply the median value for a set of dark data taken at a given gain, noise the standard deviation, and gain-bit values the medians of the gain array.
 %% Cell type:code id: tags:
 ``` python
 import copy
 from functools import partial
 def characterize_module(cells, bp_thresh, rawversion, karabo_id, h5path, h5path_idx, inp):
    import copy
    import h5py
    import numpy as np
    from cal_tools.enums import BadPixels
    def get_num_cells(fname, h5path):
        with h5py.File(fname, "r") as f:
            cells = f[f"{h5path}/cellId"][()]
            if cells == []:
                return
            maxcell = np.max(cells)
-            options = [100, 200, 400, 500, 600, 700, 800]
+            options = [100, 200, 400, 500, 600, 700, 800, 900]
            dists = np.array([(o-maxcell) for o in options])
            dists[dists<0] = 10000 # assure to always go higher
            return options[np.argmin(dists)]
    filename, channel = inp
    h5path = h5path.format(channel)
    h5path_idx = h5path_idx.format(channel)
    if cells == 0:
        cells = get_num_cells(filename, h5path)
        if cells is None:
            raise ValueError(f"ERROR! Empty image data file for channel {channel}")
    print(f"Using {cells} memory cells")
    pulseid_checksum = None
    thresholds_offset_hard, thresholds_offset_sigma, thresholds_noise_hard, thresholds_noise_sigma = bp_thresh
    infile = h5py.File(filename, "r")
    if rawversion == 2:
        count = np.squeeze(infile[f"{h5path_idx}/count"])
        first = np.squeeze(infile[f"{h5path_idx}/first"])
        last_index = int(first[count != 0][-1]+count[count != 0][-1])
        first_index = int(first[count != 0][0])
    else:
        status = np.squeeze(infile[f"{h5path_idx}/status"])
        if np.count_nonzero(status != 0) == 0:
            return
        last = np.squeeze(infile[f"{h5path_idx}/last"])
        first = np.squeeze(infile[f"{h5path_idx}/first"])
        last_index = int(last[status != 0][-1]) + 1
        first_index = int(first[status != 0][0])
    im = np.array(infile[f"{h5path}/data"][first_index:last_index,...])
    cellIds = np.squeeze(infile[f"{h5path}/cellId"][first_index:last_index,...])
    infile.close()
    pulseid_checksum = get_pulseid_checksum(filename, h5path, h5path_idx)
    im = im[:, 0, ...].astype(np.float32)
    im = np.rollaxis(im, 2)
    im = np.rollaxis(im, 2, 1)
    mcells = cells
    offset = np.zeros((im.shape[0], im.shape[1], mcells), dtype = np.float64)
    noise = np.zeros((im.shape[0], im.shape[1], mcells), dtype = np.float64)
    for cc in np.unique(cellIds[cellIds < mcells]):
        cellidx = cellIds == cc
        if offset_numpy_algorithm == "mean":
            offset[...,cc] = np.mean(im[..., cellidx], axis=2)
        else:
            offset[...,cc] = np.median(im[..., cellidx], axis=2)
        noise[...,cc] = np.std(im[..., cellidx], axis=2)
    # bad pixels
    bp = np.zeros(offset.shape, np.uint32)
    # offset related bad pixels
    offset_mn = np.nanmedian(offset, axis=(0,1))
    offset_std = np.nanstd(offset, axis=(0,1))
    bp[(offset < offset_mn-thresholds_offset_sigma*offset_std) |
       (offset > offset_mn+thresholds_offset_sigma*offset_std)] |= BadPixels.OFFSET_OUT_OF_THRESHOLD.value
    bp[(offset < thresholds_offset_hard[0]) | (offset > thresholds_offset_hard[1])] |= BadPixels.OFFSET_OUT_OF_THRESHOLD.value
    bp[~np.isfinite(offset)] |= BadPixels.OFFSET_NOISE_EVAL_ERROR.value
    # noise related bad pixels
    noise_mn = np.nanmedian(noise, axis=(0,1))
    noise_std = np.nanstd(noise, axis=(0,1))
    bp[(noise < noise_mn-thresholds_noise_sigma*noise_std) |
       (noise > noise_mn+thresholds_noise_sigma*noise_std)] |= BadPixels.NOISE_OUT_OF_THRESHOLD.value
    bp[(noise < thresholds_noise_hard[0]) | (noise > thresholds_noise_hard[1])] |= BadPixels.NOISE_OUT_OF_THRESHOLD.value
    bp[~np.isfinite(noise)] |= BadPixels.OFFSET_NOISE_EVAL_ERROR.value
    return offset, noise, bp, cells, pulseid_checksum
 offset_g = OrderedDict()
 noise_g = OrderedDict()
 gain_g = OrderedDict()
 badpix_g = OrderedDict()
 gg = 0
 start = datetime.now()
 all_cells = []
 checksums = {}
 try:
    tGain, encodedGain, operatingFreq = get_dssc_ctrl_data(in_folder + "/r{:04d}/".format(offset_runs["high"]),
                                                           slow_data_pattern,
                                                           slow_data_aggregators,
                                                           offset_runs["high"], slow_data_path)
 except IOError:
    print("ERROR: Couldn't access slow data to read tGain, encodedGain, and operatingFreq \n")
 for gain, mapped_files in gain_mapped_files.items():
    inp = []
    dones = []
    for i in modules:
        qm = "Q{}M{}".format(i//4 +1, i % 4 + 1)
        if qm in mapped_files and not mapped_files[qm].empty():
            fname_in = mapped_files[qm].get()
            print("Process file: ", fname_in)
            dones.append(mapped_files[qm].empty())
        else:
            continue
        inp.append((fname_in, i))
    p = partial(characterize_module, max_cells,
               (thresholds_offset_hard, thresholds_offset_sigma,
                thresholds_noise_hard, thresholds_noise_sigma), rawversion, karabo_id, h5path, h5path_idx)
    results = list(map(p, inp))
    for ii, r in enumerate(results):
        i = modules[ii]
        offset, noise,  bp, thiscell, pulseid_checksum = r
        all_cells.append(thiscell)
        qm = "Q{}M{}".format(i//4 +1, i % 4 + 1)
        if qm not in offset_g:
            offset_g[qm] = np.zeros((offset.shape[0], offset.shape[1], offset.shape[2]))
            noise_g[qm] = np.zeros_like(offset_g[qm])
            badpix_g[qm] = np.zeros_like(offset_g[qm], np.uint32)
            checksums[qm] = pulseid_checksum
        offset_g[qm][...] = offset
        noise_g[qm][...] = noise
        badpix_g[qm][...] = bp
    gg +=1
 if len(all_cells) > 0:
    max_cells = np.max(all_cells)
    print(f"Using {max_cells} memory cells")
 else:
    raise ValueError("0 processed memory cells. No raw data available.")
 ```
 %% Cell type:code id: tags:
 ``` python
 # TODO: add db_module when received from myMDC
 # Create the modules dict of karabo_das and PDUs
 qm_dict = OrderedDict()
 for i, k_da in zip(modules, karabo_da):
    qm = f"Q{i//4+1}M{i%4+1}"
    qm_dict[qm] = {"karabo_da": k_da,
                   "db_module": ""}
 ```
 %% Cell type:code id: tags:
 ``` python
 # Retrieve existing constants for comparison
 clist = ["Offset", "Noise"]
 old_const = {}
 old_mdata = {}
 print('Retrieve pre-existing constants for comparison.')
 for qm in offset_g.keys():
    old_const[qm] = {}
    old_mdata[qm] = {}
    qm_db = qm_dict[qm]
    karabo_da = qm_db["karabo_da"]
    for const in clist:
        dconst =getattr(Constants.DSSC, const)()
        condition = Conditions.Dark.DSSC(memory_cells=max_cells,
                                         bias_voltage=bias_voltage,
                                         pulseid_checksum=checksums[qm],
                                         acquisition_rate=operatingFreq[qm],
                                         target_gain=tGain[qm],
                                         encoded_gain=encodedGain[qm])
        # This should be used in case of running notebook
        # by a different method other than myMDC which already
        # sends CalCat info.
        # TODO: Set db_module to "" by default in the first cell
        if not qm_db["db_module"]:
            qm_db["db_module"] = get_pdu_from_db(karabo_id, karabo_da, dconst,
                                                 condition, cal_db_interface,
                                                 snapshot_at=creation_time)[0]
        data, mdata = get_from_db(karabo_id, karabo_da,
                                  dconst,
                                  condition,
                                  None,
                                  cal_db_interface, creation_time=creation_time,
                                  verbosity=2, timeout=cal_db_timeout)
        old_const[qm][const] = data
        if mdata is None or data is None:
            old_mdata[qm][const] = {
                "timestamp": "Not found",
                "filepath": None,
                "h5path": None
            }
        else:
            old_mdata[qm][const] = {
                "timestamp": mdata.calibration_constant_version.begin_at.isoformat(),
                "filepath": os.path.join(
                    mdata.calibration_constant_version.hdf5path,
                    mdata.calibration_constant_version.filename,
                ),
                "h5path": mdata.calibration_constant_version.h5path,
            }
    with open(f"{out_folder}/module_metadata_{qm}.yml", "w") as fd:
        yaml.safe_dump(
            {"module": qm, "pdu": qm_db["db_module"], "old-constants": old_mdata[qm]},
            fd,
        )
 ```
 %% Cell type:code id: tags:
 ``` python
 res = OrderedDict()
 for i in modules:
    qm = f"Q{i//4+1}M{i%4+1}"
    try:
        res[qm] = {'Offset': offset_g[qm],
                   'Noise': noise_g[qm],
                   }
    except Exception as e:
        print(f"Error: No constants for {qm}: {e}")
 ```
 %% Cell type:code id: tags:
 ``` python
 # Push the same constant two different times.
 # One with the generated pulseID check sum setting for the offline calibration.
 # And another for the online calibration as it doesn't have this pulseID checksum, yet.
 md = None
 for dont_use_pulseIds in [True, False]:
    for qm in res.keys():
        karabo_da = qm_dict[qm]["karabo_da"]
        db_module = qm_dict[qm]["db_module"]
        for const in res[qm].keys():
            dconst = getattr(Constants.DSSC, const)()
            dconst.data = res[qm][const]
            opfreq = None if dont_use_pulseIds else operatingFreq[qm]
            targetgain = None if dont_use_pulseIds else tGain[qm]
            encodedgain =  None if dont_use_pulseIds else encodedGain[qm]
            pidsum = None if dont_use_pulseIds else checksums[qm]
            # set the operating condition
            condition = Conditions.Dark.DSSC(memory_cells=max_cells,
                                             bias_voltage=bias_voltage,
                                             pulseid_checksum=pidsum,
                                             acquisition_rate=opfreq,
                                             target_gain=targetgain,
                                             encoded_gain=encodedgain)
            if db_output:
                md = send_to_db(db_module, karabo_id, dconst, condition, file_loc, report,
                                cal_db_interface, creation_time=creation_time, timeout=cal_db_timeout)
            if local_output and dont_use_pulseIds: # Don't save constant localy two times.
                md = save_const_to_h5(db_module, karabo_id, dconst, condition,
                                      dconst.data, file_loc, report,
                                      creation_time, out_folder)
                print(f"Calibration constant {const} is stored locally.\n")
        if not dont_use_pulseIds:
            print("Constants parameter conditions are:\n")
            print(f"• memory_cells: {max_cells}\n• bias_voltage: {bias_voltage}\n"
                  f"• pulseid_checksum: {pidsum}\n• acquisition_rate: {opfreq}\n"
                  f"• target_gain: {targetgain}\n• encoded_gain: {encodedgain}\n"
                  f"• creation_time: {creation_time}\n")
 ```
 %% Cell type:code id: tags:
 ``` python
 mnames = []
 for i in modules:
    qm = f"Q{i//4+1}M{i % 4+1}"
    display(Markdown(f'## Position of the module {qm} and its ASICs##'))
    mnames.append(qm)
 show_processed_modules(dinstance=dinstance, constants=None, mnames=mnames, mode="position")
 ```
 %% Cell type:markdown id: tags:
 ## Single-Cell Overviews ##
 Single cell overviews allow to identify potential effects on all memory cells, e.g. on sensor level. Additionally, they should serve as a first sanity check on expected behaviour, e.g. if structuring on the ASIC level is visible in the offsets, but otherwise no immediate artifacts are visible.
 %% Cell type:code id: tags:
 ``` python
 cell = 9
 gain = 0
 out_folder = None
 show_overview(res, cell, gain, out_folder=out_folder, infix="_{}".format(run))
 ```
 %% Cell type:code id: tags:
 ``` python
 cols = {BadPixels.NOISE_OUT_OF_THRESHOLD.value: (BadPixels.NOISE_OUT_OF_THRESHOLD.name, '#FF000080'),
        BadPixels.OFFSET_NOISE_EVAL_ERROR.value: (BadPixels.OFFSET_NOISE_EVAL_ERROR.name, '#0000FF80'),
        BadPixels.OFFSET_OUT_OF_THRESHOLD.value: (BadPixels.OFFSET_OUT_OF_THRESHOLD.name, '#00FF0080'),
        BadPixels.OFFSET_OUT_OF_THRESHOLD.value | BadPixels.NOISE_OUT_OF_THRESHOLD.value: ('MIXED', '#DD00DD80')}
 if high_res_badpix_3d:
    display(Markdown("""
    ## Global Bad Pixel Behaviour ##
    The following plots show the results of bad pixel evaluation for all evaluated memory cells.
    Cells are stacked in the Z-dimension, while pixels values in x/y are rebinned with a factor of 2.
    This excludes single bad pixels present only in disconnected pixels.
    Hence, any bad pixels spanning at least 4 pixels in the x/y-plane, or across at least two memory cells are indicated.
    Colors encode the bad pixel type, or mixed type.
    """))
    # set rebin_fac to 1 for avoiding rebining and
    # losing real values of badpixels(High resolution).
    gain = 0
    for mod, data in badpix_g.items():
        plot_badpix_3d(data, cols, title=mod, rebin_fac=2)
        plt.show()
 ```
 %% Cell type:markdown id: tags:
 ## Aggregate values, and per Cell behaviour ##
 The following tables and plots give an overview of statistical aggregates for each constant, as well as per cell behavior.
 %% Cell type:code id: tags:
 ``` python
 create_constant_overview(offset_g, "Offset (ADU)", max_cells, entries=1)
 ```
 %% Cell type:code id: tags:
 ``` python
 create_constant_overview(noise_g, "Noise (ADU)", max_cells, 0, 100, entries=1)
 ```
 %% Cell type:code id: tags:
 ``` python
 bad_pixel_aggregate_g = OrderedDict()
 for m, d in badpix_g.items():
    bad_pixel_aggregate_g[m] = d.astype(np.bool).astype(np.float)
 create_constant_overview(bad_pixel_aggregate_g, "Bad pixel fraction", max_cells, entries=1)
 ```
 %% Cell type:markdown id: tags:
 ## Summary tables ##
 The following tables show summary information for the evaluated module. Values for currently evaluated constants are compared with values for pre-existing constants retrieved from the calibration database.
 %% Cell type:code id: tags:
 ``` python
 time_summary = []
 for qm, qm_data in old_mdata.items():
    time_summary.append(f"The following pre-existing constants are used for comparison for module {qm}:")
    for const, const_data in qm_data.items():
        time_summary.append(f"- {const} created at {const_data['timestamp']}")
 display(Markdown("\n".join(time_summary)))
 ```
 %% Cell type:code id: tags:
 ``` python
 header = ['Parameter',
          "New constant", "Old constant ",
          "New constant", "Old constant ",
          "New constant", "Old constant "]
 for const in ['Offset', 'Noise']:
    table = [['','High gain', 'High gain']]
    for qm in res.keys():
        data = np.copy(res[qm][const])
        if old_const[qm][const] is not None:
            dataold = np.copy(old_const[qm][const])
        f_list = [np.nanmedian, np.nanmean, np.nanstd, np.nanmin, np.nanmax]
        n_list = ['Median', 'Mean', 'Std', 'Min', 'Max']
        for i, f in enumerate(f_list):
            line = [n_list[i]]
            line.append('{:6.1f}'.format(f(data[...,gain])))
            if old_const[qm][const] is not None:
                line.append('{:6.1f}'.format(f(dataold[...,gain])))
            else:
                line.append('-')
            table.append(line)
    display(Markdown('### {} [ADU], good and bad pixels ###'.format(const)))
    md = display(Latex(tabulate.tabulate(table, tablefmt='latex', headers=header)))
 ```