from __future__ import absolute_import, division, print_function from iotbx.phil import parse #help_message = ''' #diffBragg command line utility #''' hopper_phil = """ use_geometric_mean_Fhkl = False .type = bool .help = whether to use the geometric mean for Fhkl restraint (when betasFhkl is not None, restratin Fhkl to the mean in each res bin) Fhkl_channel_bounds = None .type = floats .help = Energy bounds for energy-dependent structure factors. Units are eV. .help = If provided refine a unique structure factor correction for each bin defined by Fhkl_channel_bins .help = 0 and infinity are implicit. Providing a single number, e.g. 8950, will refine two sets of structure .help = factors: one for energies [0-8950), and another for energies [8950-infinity] Fhkl_dspace_bins = 10 .type = int .help = number of resolution bins (out to corner of detector) for computing the average structure factor intensity .help = One can then restrain to these values when fix.Fhkl=False by using the restraint strength betas.Fhkl (default is None in which case no restraints are applied) try_strong_mask_only = False .type = bool .help = if strong spot masks are present in the input refls, then use them .help = as the trusted-flags, i.e. only run strong spot pixels through diffBragg dilate_strong_mask = None .type = int .help = if using strong mask only for refinement, then dilate it this many iterations using .help = scipy.ndimage method binary_dilation (has to be >= 1) hopper_save_freq = None .type = int .help = save the output files when the iteration number is a multiple of this argument terminate_after_n_converged_iter = None .type = int .help = optionally converge if all parameters seem converged for this many iterations. See .help = converged_param_percent_change converged_param_percent_change = 0.1 .type = float .help = if a parameter's value changes by less than this amount (percent) between iterations, for a .help = a ceterain number of iters (determined by terminate_after_n_converged_iter), then .help = optimization is considered done, and is stopped mask_highest_values = None .type = int .help = mask out the N highest-valued pixels in a shoebox when performing diffBragg refinement use_float32 = False .type = bool .help = store pixel data and background models in 32bit arrays .expert_level=10 test_gathered_file = False .type = bool .help = run a quick test to ensure the gathered data file preserves information .expert_level=10 load_data_from_refls = False .type = bool .help = load image data, background etc from reflection tables .expert_level=10 gathered_output_file = None .type = str .help = optional file for storing a new hopper input file which points to the gathered data dumps .expert_level=10 only_dump_gathers = False .type = bool .help = only reads in image data, fits background planes, and dumps .help = results to disk, writes a new exper refl file at the end .expert_level=10 gathers_dir = None .type = str .help = folder where gathered data reflection tables .help = will be writen (if dump_gathers=True) .expert_level=10 dump_gathers = False .type = bool .help = optionally dump the loaded experimental data to reflection tables .help = for portability .expert_level=10 spectrum_from_imageset = False .type = bool .help = if True, load the spectrum from the imageset in the experiment, then probably downsample it .expert_level=0 gen_gauss_spec = False .type = bool .help = If the experimental data dont include spectra, one can try generating gaussian spectra. .help = See the diffBragg/phil.py under simulator.spectrum.gauss_spec. use_perpixel_dark_rms = False .type = bool .help = some Jungfrau formats have per-pixel RMS values that change shot-to-shot with the dynamic .help = gain mode switching. If this flag is true, then the per-pixel gain modes will be extracted from the image .help = format . The image format class is expected to have a method named get_pedestal_rms(self, index=None) .help = See the method xfel/util/jungfrau/get_pedestalRMS-from_jungfrau isotropic { diffuse_gamma = False .type = bool .help = refine a single diffuse gamma parameter as opposed to 3 diffuse_sigma = False .type = bool .help = refine a single diffuse gamma parameter as opposed to 3 } downsamp_spec { skip = False .type = bool .help = if reading spectra from imageset, optionally skip the downsample portion .help = Note, if skip=True, then total flux will be determined by whats in the imageset spectrum (sum of the weights) .expert_level=10 filt_freq = 0.07 .type = float .help = low pass filter frequency in units of inverse spectrometer pixels (??) .expert_level=10 filt_order = 3 .type = int .help = order for bandpass butter filter .expert_level=10 tail = 50 .type = int .help = endpoints of the spectrum that are used in background estimation .expert_level=10 delta_en = 0.5 .type = float .help = final resolution of downsampled spectrum in eV .expert_level=0 } apply_best_crystal_model = False .type = bool .help = depending on what experiments in the exper refl file, one may want .help = to apply the optimal crystal transformations (this parameter only matters .help = if params.best_pickle is not None) .expert_level=10 filter_unpredicted_refls_in_output = True .type = bool .help = filter reflections in the output refl table for which there was no model bragg peak .help = after stage 1 termination .expert_level=10 tag = stage1 .type = str .help = output name tag .expert_level=0 ignore_existing = False .type = bool .help = experimental, ignore expts that already have optimized models in the output dir .expert_level=0 global_method = *basinhopping annealing .type = choice .help = the method of global optimization to use .expert_level=10 nelder_mead_maxfev = 60 .type = int .help = multiplied by total number of modeled pixels to get max number of iterations .expert_level=10 nelder_mead_fatol = 0.0001 .type = float .help = nelder mead functional error tolerance niter_per_J = 1 .type = int .help = if using gradient descent, compute gradients .help = every niter_per_J iterations . .expert_level=10 rescale_params = True .type = bool .help = DEPRECATED, this parameter no longer has meaning. .expert_level=10 best_pickle = None .type = str .help = path to a pandas pickle containing the best models for the experiments .expert_level=0 betas .help = variances for the restraint targets .expert_level=0 { Finit = None .type = float Friedel = None .type = float .help = set this to some value to restraint Friedel mates during refinement (ensemble mode) . Lower values are .help = tightly restrained . (Exploratory, experimental phil param) ucell_a = None .type = float .help = restraint variance for unit cell a ucell_b = None .type = float .help = restraint variance for unit cell b ucell_c = None .type = float .help = restraint variance for unit cell c ucell_alpha = None .type = float .help = restraint variance for unit cell alpha angle ucell_beta = None .type = float .help = restraint variance for unit cell beta angle ucell_gamma = None .type = float .help = restraint variance for unit cell gamma angle Nvol = 1e8 .type = float .help = tightness of the Nabc volume contraint detz_shift = 1e8 .type = float .help = restraint variance for detector shift target ucell = [1e8,1e8,1e8,1e8,1e8,1e8] .type = floats .help = DEPRECATED: use e.g. betas.ucell_a instead .help = variances for unit cell constants in order determined by unit cell manager class (see diffBragg/refiners/crystal_systems) RotXYZ = 1e8 .type = float .help = restraint factor for the rotXYZ restraint Nabc = [1e8,1e8,1e8] .type = floats(size=3) .help = restraint factor for the ncells abc Ndef = [1e8,1e8,1e8] .type = floats(size=3) .help = restraint factor for the ncells def diffuse_sigma = 1e8,1e8,1e8 .type = floats(size=3) .help = restraint factor for diffuse sigma diffuse_gamma = 1e8,1e8,1e8 .type = floats(size=3) .help = restraint factor for diffuse gamma G = 1e8 .type = float .help = restraint factor for the scale G B = 1e8 .type = float .help = restraint factor for Bfactor eta_abc = [1e8,1e8,1e8] .type = floats(size=3) .help = restrain factor for mosaic spread angles spec = [1e8,1e8] .type = floats(size=2) .help = restraint factor for spectrum coefs Fhkl = None .type = float .help = restraint factor for structure factor intensity scales } dual .help = configuration parameters for dual annealing .expert_level=10 { initial_temp = 5230 .type = float .help = init temp for dual annealing no_local_search = False .type = bool .help = whether to try local search procedure with dual annealing .help = if False, then falls back on classical simulated annealing visit = 2.62 .type = float .help = dual_annealing visit param, see scipy optimize docs accept = -5 .type = float .help = dual_annealing accept param, see scipy optimize docs } centers .help = restraint targets .expert_level=0 { ucell_a = None .type = float .help = restraint target for unit cell a (Angstrom) ucell_b = None .type = float .help = restraint target for unit cell b (Angstrom) ucell_c = None .type = float .help = restraint target for unit cell c (Angstrom) ucell_alpha = None .type = float .help = restraint target for unit cell alpha angle (deg.) ucell_beta = None .type = float .help = restraint target for unit cell beta angle (deg.) ucell_gamma = None .type = float .help = restraint target for unit cell gamma angle (deg.) Nvol = None .type = float .help = if provided, constrain the product Na*Nb*Nc to this value detz_shift = 0 .type = float .help = restraint target for detector shift along z-direction ucell = [63.66, 28.87, 35.86, 1.8425] .type = floats .help = DEPRECATED: use e.g. betas.ucell_a instead .help = centers for unit cell constants in order determined by unit cell manager class (see diffBragg/refiners/crystal_systems) RotXYZ = [0,0,0] .type = floats(size=3) .help = restraint target for Umat rotations Nabc = [100,100,100] .type = floats(size=3) .help = restraint target for Nabc Ndef = [0,0,0] .type = floats(size=3) .help = restraint target for Ndef diffuse_sigma = [1,1,1] .type = floats(size=3) .help = restraint target for diffuse sigma diffuse_gamma = [1,1,1] .type = floats(size=3) .help = restraint target for diffuse gamma G = 100 .type = float .help = restraint target for scale G B = 0 .type = float .help = restraint target for Bfactor eta_abc = [0,0,0] .type = floats(size=3) .help = restraint target for mosaic spread angles in degrees spec = [0,1] .type = floats(size=2) .help = restraint target for specturm correction (0 + 1*Lambda ) } skip = None .type = int .help = skip this many exp .expert_level=0 hess = None .type = str .help = scipy minimize hessian argument, 2-point, 3-point, cs, or None .expert_level=10 stepsize = 0.5 .type = float .help = basinhopping stepsize .expert_level=10 temp = 1 .type = float .help = temperature for basin hopping algo .expert_level=10 niter = 0 .type = int .help = number of basin hopping iterations (0 just does a gradient descent and stops at the first minima encountered) .expert_level=0 exp_ref_spec_file = None .type = str .help = path to 3 col txt file containing file names for exper, refl, spectrum (.lam) .help = Note: only single-image experiment lists are supported! Uses dials.split_experiments or diffBragg.make_input_file if necessary .expert_level=0 method = None .type = str .help = minimizer method, usually this is L-BFGS-B (gradients) or Nelder-Mead (simplex) .help = other methods are experimental (see details in hopper_utils.py) .expert_level=0 opt_det = None .type = str .help = path to experiment with optimized detector model .expert_level=0 opt_beam = None .type = str .help = path to experiment with optimized beam model .expert_level=0 number_of_xtals = 1 .type = int .help = number of crystal domains to model per shot .expert_level=10 sanity_test_input = True .type = bool .help = sanity test input .expert_level=10 outdir = None .type = str .help = output folder .expert_level=0 max_process = -1 .type = int .help = max exp to process .expert_level=0 types .help = type of target to parameter (see diffBragg.refiners.parameters.py) .expert_level=10 { G = *ranged positive .type = choice Nabc = *ranged positive .type = choice diffuse_sigma = *ranged positive .type = choice diffuse_gamma = *ranged positive .type = choice } sigmas .help = sensitivity of target to parameter (experimental) .expert_level=10 { spec = [1,1] .type=floats(size=2) .help = spectrum offset and scale factor sigmas roiPerScale = 1 .type = float detz_shift = 1 .type = float .help = sensitivity shift for the overall detector shift along z-direction Nabc = [1,1,1] .type = floats(size=3) .help = sensitivity for Nabc Ndef = [1,1,1] .type = floats(size=3) .help = sensitivity for Ndef diffuse_sigma = [1,1,1] .type = floats(size=3) .help = sensitivity for diffuse sigma diffuse_gamma = [1,1,1] .type = floats(size=3) .help = sensitivity for diffuse gamma RotXYZ = [1,1,1] .type = floats(size=3) .help = sensitivity for RotXYZ G = 1 .type = float .help = sensitivity for scale factor B = 1 .type = float .help = sensitivity for Bfactor eta_abc = [1,1,1] .type = floats(size=3) .help = sensitivity of mosaic spread parameters ucell = [1,1,1,1,1,1] .type = floats .help = sensitivity for unit cell params Fhkl = 1 .type = float .help = sensitivity for structure factors } init .help = initial value of model parameter (will be overrided if best pickle is provided) .expert_level=0 { random_Gs = None .type = floats .help = list of floats from which to select an init.G at random random_Nabcs = None .type = floats .help = list of random floats from which to select init.Nabc at random spec = [0,1] .type = floats(size=2) .help = initial offset and scale factor applied to each energy channel wavelength detz_shift = 0 .type = float .help = initial value for the detector position overall shift along z-direction in millimeters Nabc = [100,100,100] .type = floats(size=3) .help = init for Nabc Ndef = [0,0,0] .type = floats(size=3) .help = init for Ndef diffuse_sigma = [.01,.01,.01] .type = floats(size=3) .help = init diffuse sigma diffuse_gamma = [1,1,1] .type = floats(size=3) .help = init for diffuse gamma RotXYZ = [0,0,0] .type = floats(size=3) .help = init for RotXYZ G = 1 .type = float .help = init for scale factor B = 0 .type = float .help = init for B factor eta_abc = [0,0,0] .type = floats(size=3) .help = initial values (in degrees) for anisotropic mosaic spread about the 3 crystal axes a,b,c .help = Note, these can never be exactly 0 if fix.eta_abc=False } mins .help = min value allowed for parameter .expert_level = 0 { detz_shift = -10 .type = float .help = min value for detector z-shift in millimeters Nabc = [3,3,3] .type = floats(size=3) .help = min for Nabc Ndef = [-200,-200,-200] .type = floats(size=3) .help = min for Ndef diffuse_sigma = [0,0,0] .type = floats(size=3) .help = min diffuse sigma diffuse_gamma = [0,0,0] .type = floats(size=3) .help = min for diffuse gamma RotXYZ = [-1,-1,-1] .type = floats(size=3) .help = min for rotXYZ in degrees G = 0 .type = float .help = min for scale G B = 0 .type = float .help = min for Bfactor Fhkl = 0 .type = float .help = min for structure factors eta_abc = [0,0,0] .type = floats(size=3) .help = min value (in degrees) for mosaic spread angles spec = [-0.01, 0.95] .type = floats(size=2) .help = min value for spectrum correction (-0.01 + Lambda *1.05) } maxs .help = max value allowed for parameter .expert_level = 0 { detz_shift = 10 .type = float .help = max value for detector z-shift in millimeters eta = 0.1 .type = float .help = maximum mosaic spread in degrees Nabc = [300,300,300] .type = floats(size=3) .help = max for Nabc Ndef = [200,200,200] .type = floats(size=3) .help = max for Ndef diffuse_sigma = [20,20,20] .type = floats(size=3) .help = max diffuse sigma diffuse_gamma = [1000,1000,1000] .type = floats(size=3) .help = max for diffuse gamma RotXYZ = [1,1,1] .type = floats(size=3) .help = max for rotXYZ in degrees G = 1e12 .type = float .help = max for scale G B = 1e3 .type = float .help = max for Bfactor eta_abc = [10,10,10] .type = floats(size=3) .help = maximum value (in degrees) for mosaic spread angles Fhkl = 1e6 .type = float .help = max for structure factors spec = [0.01, 1.05] .type = floats(size=2) .help = max value for spectrum correction (0.01 + Lambda *1.05) } fix .help = flags for fixing parameters during refinement .expert_level = 0 { Fhkl = True .type = bool .help = fix the structure factors scales during refinement spec = True .type = bool .help = fix the spectrum. If False, a spectrum correction is refined (wavelength shift and scale) perRoiScale = True .type = bool .help = a per-roi scale factor G = False .type = bool .help = fix the Bragg spot scale during refinement B = True .type = bool .help = fix the Bfactor during refinement eta_abc = True .type = bool .help = fix the mosaic spread parameters during refinement RotXYZ = False .type = bool .help = fix the misorientation matrix during refinement Nabc = False .type = bool .help = fix the diagonal mosaic domain size parameters during refinement Ndef = True .type = bool .help = fix the diagonal mosaic domain size parameters during refinement diffuse_sigma = True .type = bool .help = fix diffuse sigma diffuse_gamma = True .type = bool .help = fix diffuse gamma ucell = False .type = bool .help = fix the unit cell during refinement detz_shift = False .type = bool .help = fix the detector distance shift during refinement } relative_tilt = False .type = bool .help = fit tilt coef relative to roi corner .expert_level = 10 ucell_edge_perc = 10 .type = float .help = precentage for allowing ucell to fluctuate during refinement .expert_level = 10 ucell_ang_abs = 5 .type = float .help = absolute angle deviation in degrees for unit cell angles to vary during refinement .expert_level = 10 no_Nabc_scale = False .type = bool .help = toggle Nabc scaling of the intensity .expert_level = 10 use_diffuse_models = False .type = bool .help = if True, let the values of init.diffuse_sigma and init.diffuse_gamma .help = be used to define the diffuse scattering. Set e.g. fix.diffuse_sigma=True in order to refine them .expert_level = 10 diffuse_stencil_size = 0 .type = int .help = Increase to add accuracy to diffuse scattering models, at the expense of longer computations .help = Best to increment by values of 1 when testing symmetrize_diffuse = True .type = bool .help = use the laue group rotation operators to symmetrize diffuse signals gamma_miller_units = False .type = bool .help = if True, let the values of init.diffuse_gamma be expressed in Miller index units .expert_level = 10 sigma_frac = None .type = float .help = sigma for Fhkl restraints will be some fraction of the starting value .expert_level = 10 sanity_test_hkl_variation = False .type = bool .help = measure the variation of each HKL within the shoebox .expert_level = 10 sanity_test_models = False .type = bool .help = make sure best models from stage 1 are reproduced at the start .expert_level = 10 sanity_test_amplitudes = False .type = bool .help = if True, then quickly run a sanity check ensuring that all h,k,l are predicted .help = and/or in the starting miller array .expert_level = 10 x_write_freq = 25 .type = int .help = save x arrays every x_write_freq iterations .expert_level = 10 percentile_cut = None .type = float .help = percentile below which pixels are masked .expert_level = 10 remove_duplicate_hkl = False .type = bool .help = for hopper, remove duplicate HKLs in input refl files space_group = None .type = str .help = space group to refine structure factors in .expert_level = 0 first_n = None .type = int .help = refine the first n shots only .expert_level = 0 maxiter = 15000 .type = int .help = stop refiner after this many iters .expert_level = 10 ftol = 1e-10 .type = float .help = ftol convergence threshold for scipys L-BFGS-B .expert_level = 10 lbfgs_maxiter = 1e5 .type = int .help = maximum number of L-BFGS-B iterations disp = False .type = bool .help = scipy minimize convergence printouts .expert_level = 10 use_restraints = True .type = bool .help = disable the parameter restraints .expert_level = 0 min_multi = 2 .type = int .help = minimum ASU multiplicity, obs that fall below this threshold .help = are removed from analysis .expert_level = 10 min_spot = 5 .type = int .help = minimum spots on a shot in order to optimize that shot .expert_level = 10 store_wavelength_images = False .type = bool .help = for simtbx.diffBragg.hopper, optionally write subimages whose value .help = is the avereage wavelength per pixels, weighted by the model logging .help = controls the logging module for hopper and stage_two .expert_level = 10 { show_params_at_minimum = True .type = bool .help = show the optimized parameters once a basinhopping minimum is reached parameters = False .type = bool .help = whether to display hopper refinement parameters at each iteration disable = False .type = bool .help = turn off logging logfiles_level = low *normal high .type = choice .help = level of the main log when writing logfiles logfiles = False .type = bool .help = write log files in the outputdir rank0_level = low *normal high .type = choice .help = console log level for rank 0, ignored if logfiles=True other_ranks_level = *low normal high .type = choice .help = console log level for all ranks > 0, ignored if logfiles=True overwrite = True .type = bool .help = overwrite the existing logfiles logname = mainLog .type = str .help = if logfiles=True, then write the log to this file, stored in the folder specified by outdir .help = if None, then defaults to main_stage1.log for hopper, main_pred.log for prediction, main_stage2.log for stage_two } profile = False .type = bool .help = profile the workhorse functions .expert_level = 0 profile_name = lineProf .type = str .help = name of the output file that stores the line-by-line profile (written to folder specified by outdir) .help = if None, defaults to prof_stage1.log, prof_pred.log, prof_stage2.log for hopper, prediction, stage_two respectively .expert_level = 10 """ simulator_phil = """ simulator { oversample = 0 .type = int .help = pixel oversample rate (0 means auto-select) device_id = 0 .type = int .help = device id for GPU simulation init_scale = 1 .type = float .help = initial scale factor for this crystal simulation total_flux = 1e12 .type = float .help = total photon flux for all energies crystal { ncells_abc = (10,10,10) .type = floats(size=3) .help = number of unit cells along each crystal axis making up a mosaic domain ncells_def = (0,0,0) .type = floats(size=3) .help = off-diagonal components for mosaic domain model (experimental) has_isotropic_ncells = False .type = bool .help = if True, ncells_abc are constrained to be the same values during refinement has_isotropic_mosaicity = False .type = bool .help = if True, eta_abc are constrained to be the same values during refinement mosaicity = 0 .type = float .help = mosaic spread in degrees anisotropic_mosaicity = None .type = floats .help = mosaic spread 3-tuple or 6-tuple specifying anisotropic mosaicity num_mosaicity_samples = 1 .type = int .help = the number of mosaic domains to use when simulating mosaic spread mos_angles_per_axis = 10 .type = int .help = if doing a uniform mosaicity sampling, use this many angles per rotation axis num_mos_axes = 10 .type = int .help = number of sampled rot axes if doing a uniform mosaicity sampling mosaicity_method = 2 .type = int .help = 1 or 2. 1 is random sampling, 2 is even sampling rotXYZ_ucell = None .type = floats(size=9) .help = three missetting angles (about X,Y,Z axes), followed by .help = unit cell parameters. The crystal will be rotated according to .help = the matrix RotX*RotY*RotZ, and then the unit cell will be updated } structure_factors { from_pdb { name = None .type = str .help = path to a pdb file add_anom = True .type = bool .help = Use the dxtbx beams wavelength to sample the henke tables .help = and add anomalous contributions from each atom k_sol = None .type = float .help = solvent component of structure factor: k_sol * exp(-b_sol*s^2/4) b_sol = None .type = float .help = solvent component of structure factor: k_sol * exp(-b_sol*s^2/4) } mtz_name = None .type = str .help = path to an MTZ file . If an mtz_name and from_pdb.name are both provided, then .help = the mtz takes precedence mtz_column = None .type = str .help = column in an MTZ file dmin = 1 .type = float .help = minimum resolution for structure factor array (not applicable when F is loaded from mtz) dmax = None .type = float .help = maximum resolution for structure factor array (not applicable when F is loaded from mtz) default_F = 0 .type = float .help = Default value for structure factor amps . MIssing structure factors will have this value .help = during simulation, for example if the mtz is incomplete. Also, if mtz_name and .help = from_pdb.name are both None, then a structure factor array will be created with this .help = value as every amplitude. } spectrum { filename = None .type = str .help = a .lam file (precognition) for inputting wavelength spectra stride = 1 .type = int .help = stride of the spectrum (e.g. set to 10 to keep every 10th value in the spectrum file data) filename_list = None .type = str .help = path to a file containing 1 .lam filename per line gauss_spec { fwhm = 10 .type = float .help = width of the gaussian in electron volts nchannels = 20 .type = int .help = total number of spectrum energies, centered on the shots nominal energy as determined from format class res = 1 .type = float .help = energy resolution of the spectrum } } beam { size_mm = 1 .type = float .help = diameter of the beam in mm } detector { thick = None .type = float .help = sensor thickness in millimeters. Overrides dxtbx detector model. .help = Note: must also provide param `atten`, otherwise this param is ineffective atten = None .type = float .help = x-ray absorption length in millimeters .help = for sensor. Overrides dxtbx detector model. .help = Note: must also provide param `thick`, otherwise this param is ineffective force_zero_thickness = False .type = bool .help = if True, then set sensor thickness to 0 thicksteps = 1 .type = int .help = number of layers within sensor where scattering .help = will be averaged over (evenly divided). This is a nanoBragg attribute } psf { use = False .type = bool .help = optionally apply a point-spread-function to the model fwhm = 100 .type = float .help = PSF full width half max in microns radius = 7 .type = int .help = PSF kernel radius (in pixels) } } """ refiner_phil = """ refiner { check_expt_format = True .type = bool .help = In some cases the expt is only used for the crystal model, in which case set check_expt_format=False. .help = If, however, the experimental data and/or spectra are to be extracted from the expt, then this .help = should remain True. refldata_trusted = *allValid fg bg .type = choice .help = If loading data from reflection table, choose which pixels are flagged as trusted/ .help = The default is allValid, meaning any shoebox mask value > 1. .help = fg is any foreground (integrated), valid pixel (mask==5). .help = bg is any valid pixel used for background fitting or foreground (integration). .help = Note, in this context, Foreground is usually the central shoebox pixels marked for integration. refldata_to_photons = False .type = bool .help = If loading data from reflection table, then optionally normalize by the refiner.adu_to_photon factor. .help = If the reflection tables were created using dials.integrate or dials.stils_process, .help = then you will need to set this flag to True. load_data_from_refl = False .type = bool .help = Rather than load data from the experiment, load data from the reflection table shoeboxes. .help = The data in shoeboxes is stored in float32. The shoebox bound boxes will determing the regions of .help = pixels used for refinement. It is assumed that shoebox background, data, and the mask are properly set. .help = See the method GatherFromReflectionTable in hopper_utils. test_gathered_file = False .type = bool gather_dir = None .type = str .help = optional dir for stashing loaded input data in refl files (mainly for tests/portability) break_signal = None .type = int .help = intended to be used to break out of a long refinement job prior to a timeout on a super computer .help = On summit, set this to 12 (SIGUSR2), at least thats what it was last I checked (July 2021) debug_pixel_panelfastslow = None .type = ints(size=3) .help = 3-tuple of panel ID, fast coord, slow coord. If set, show the state of diffBragg .help = for this pixel once refinement has finished gain_map_min_max = [.5,2] .type = floats(size=2) .help = the min, max allowed values for the gain correction terms .help = that are applied to each region (defined by region_size) refine_gain_map = False .type = bool .help = flag for refining a detector gain map, defined by the parameter region_size save_gain_freq=10 .type = int .help = after how many iterations should we save the optimized gain map region_size = [50,50] .type = ints(size=2) .help = Used for gain correction. size of subregions in each detector module in pixels. .help = Each panel region will be divided into blocks of shape region_size .help = and a unique gain correction will be applied to each subregion. .help = Note, this will usually be square shaped, but its (slowDim,fastDim). res_ranges = None .type = str .help = resolution-defining strings, where each string is .help = is comma-separated substrings, formatted according to "%f-%f,%f-%f" where the first float .help = in each substr specifies the high-resolution for the refinement trial, and the second float .help = specifies the low-resolution for the refinement trial. Should be same length as max_calls force_symbol = None .type = str .help = a space group lookup symbol used to map input miller indices to ASU force_unit_cell = None .type = ints(size=6) .help = a unit cell tuple to use randomize_devices = True .type = bool .help = randomly select a device Id num_devices = 1 .type = int .help = number of cuda devices on current node refine_Fcell = None .type = ints(size_min=1) .help = whether to refine the structure factor amplitudes refine_spot_scale = None .type = ints(size_min=1) .help = whether to refine the crystal scale factor refine_Nabc = False .type = bool .help = whether to refine the mosaic domain size tensor gain_restraint=None .type = floats(size=2) .help = if not None, apply a gain restraint to the data .help = This is two parameters, a center and a variance max_calls = [100] .type = ints(size_min=1) .help = maximum number of calls for the refinement trial panel_group_file = None .type = str .help = a text file with 2 columns, the first column is the panel_id and the second .help = column is the panel_group_id. Panels geometries in the same group are refined together update_oversample_during_refinement = False .type = bool .help = whether to update the oversample parameter as ncells changes sigma_r = 3 .type = float .help = standard deviation of the dark signal fluctuation adu_per_photon = 1 .type = float .help = how many ADUs (detector units) equal 1 photon use_curvatures_threshold = 10 .type = int .help = how many consecutiv positive curvature results before switching to curvature mode curvatures = False .type = bool .help = whether to try using curvatures start_with_curvatures = False .type = bool .help = whether to try using curvatures in the first iteration tradeps = 1e-2 .type = float .help = LBFGS termination parameter (smaller means minimize for longer) io { restart_file = None .type = str .help = output file for re-starting a simulation output_dir = None .type = str .help = optional output directory } quiet = False .type = bool .help = silence the refiner verbose = 0 .type = int .help = verbosity level (0-10) for nanoBragg num_macro_cycles = 1 .type = int .help = keep repeating the same refinement scheme over and over, this many times ncells_mask = *000 110 101 011 111 .type = choice .help = a mask specifying which ncells parameters should be the same .help = e.g. 110 specifies Na and Nb are refined together as one parameter reference_geom = None .type = str .help = path to expt list file containing a detector model stage_two { use_nominal_hkl = True .type = bool .help = use the nominal hkl as a filter for Fhkl gradients save_model_freq = 50 .type = int .help = save the model after this many iterations save_Z_freq = 25 .type = int .help = save Z-scores for all pixels after this many iterations min_multiplicity = 1 .type = int .help = structure factors whose multiplicity falls below this value .help = will not be refined Fref_mtzname = None .type = str .help = path to a reference MTZ file. if passed, this is used solely to .help = observe the R-factor and CC between it and the Fobs being optimized Fref_mtzcol = "Famp(+),Famp(-)" .type = str .help = column in the mtz file containing the data d_min = 2 .type = float .help = high res lim for binner d_max = 999 .type = float .help = low res lim for binner n_bin = 10 .type = int .help = number of binner bins } } """ roi_phil = """ roi { centroid = *obs cal .type = choice .help = Determines which refl table column contains the spot centroids .help = Shoeboxes are drawn around the centroids, and refinement uses pixels .help = within those shoeboxes. .help = obs: xyz.px.value cal: xyzcal.px trusted_range = None .type = floats(size=2) .help = optional override for detector trusted range, should be (min,max) mask_all_if_any_outside_trusted_range = True .type = bool .help = If a reflection has any pixels which are outside the detectors .help = trusted range, then mask the entire reflections and surrounding .help = pixels. If False, then only pixels outside the range are masked. .help = Note: this only takes effect if mask_outside_trusted_range=True. mask_outside_trusted_range = False .type = bool .help = Check the dxtbx detector's trusted range and use that to mask .help = out-of-range pixels on a per-image basis only_filter_zingers_above_mean = True .type = bool .help = if fitting background, theres a zinger filter step (background_threshold) .help = and typically it only applies to pixels above the mean .help = Set this to False to filter zingers below the mean, which is useful for .help = data with low background signal. cache_dir_only = False .type = bool .help = if True, create the cache folder , populate it with the roi data, then exit fit_tilt = False .type = bool .help = fit tilt plane, or else background is simply an offset force_negative_background_to_zero = False .type = bool .help = if True and the background model evaluates to a negative number .help = within an ROI, then force the background to be 0 for all pixels in that ROI background_threshold = 3.5 .type = float .help = for determining background pixels pad_shoebox_for_background_estimation = None .type = int .help = shoebox_size specifies the dimenstion of the shoebox used during refinement .help = and this parameter is used to increase that shoebox_size only during the background .help = estimation stage shoebox_size = 10 .type = int .help = roi box dimension deltaQ = None .type = float .help = roi dimension in inverse Angstrom, such that shoeboxes at wider angles are larger. .help = If this parameter is supplied, shoebox_size will be ignored. reject_edge_reflections = True .type = bool .help = whether to reject ROIs if they occur near the detector panel edge reject_roi_with_hotpix = True .type = bool .help = whether to reject an ROI if it has a bad pixel hotpixel_mask = None .type = str .help = path to a hotpixel mask (hot pixels set to True) panels = None .type = str .help = panel list for refinement as a string, e.g. "0-8,10,32-40" . The ranges are inclusive, .help = e.g. 0-8 means panels 0,1,2,3,4,5,6,7,8 fit_tilt_using_weights = True .type = bool .help = if not using robust estimation for background, and instead using tilt plane fit, .help = then this parameter will toggle the use of weights. Weights are the estimated .help = pixel variance, incuding readout and shot noises. allow_overlapping_spots = False .type = bool .help = if True, then model overlapping spots skip_roi_with_negative_bg = True .type = bool .help = If a region of interest contains negative background model, then skip entire region,else .help = mask the pixels with negative background model. } geometry { save_state_freq = 50 .type = int .help = how often to save all model parameters save_state_overwrite = True .type = bool .help = whether to overwrite model parameter files each time they are saved pandas_dir = None .type = str .help = If provided, Pandas dataframes for each shot will be written here .help = These are the same format as those saved during stage 1 .help = Also, data modelers will pickled and written here as well optimized_results_tag = None .type = str .help = optional tagname, if provided,write optimized refls/expts alongside the .help = input refls/expts however using this tag suffix refls_key = stage1_refls .type = str .help = column name for the input pickle which contains the reflection tables to be modeled optimize_method = *lbfgsb nelder .type = choice .help = lmfit optimization method (lbfgsb uses gradients, nelder is graientless) input_pkl = None .type = str .help = path to the input pickle (from hopper) containing models and experiment lists input_pkl_glob = None .type = str .help = path to a glob of input pandas pickles output by hopper or geometry_refiner optimize = False .type = bool .help = flag to specify whether to optimize the geometry save_optimized_det_freq = 1 .type = int .help = Save the optimzied detector model every X iterations optimized_detector_name = "diffBragg_detector.expt" .type = str .help = basename of the experiment which will be written, and will contain the optimized detector. Note, this should be a basename only (not to be prefixed with a directory path). If a directory path is include, it will be stripped. The file will be stored in the output folder (pandas_dir) min { panel_rotations = -1,-1,-1 .type = floats(size=3) .help = minimum value in degrees for a detector panel rotation panel_translations = -1,-1,-1 .type = floats(size=3) .help = minimum value in mm for detector panel translations in X,Y,Z } max { panel_rotations = 1,1,1 .type = floats(size=3) .help = maximum value in degrees for a detector panel rotation panel_translations = 1,1,1 .type = floats(size=3) .help = maximum value in mm for detector panel translations in X,Y,Z } center { panel_rotations = 0,0,0 .type = floats(size=3) .help = restraint target in degrees for panel rotations panel_translations = 0,0,0 .type = floats(size=3) .help = restraint target in mm for detector panel translations in X,Y,Z } betas { panel_rot = 1e6,1e6,1e6 .type = floats(size=3) .help = restraint factor for panel rotations (higher values lead to unrestrained parameters) panel_xyz = 1e6,1e6,1e6 .type = floats(size=3) .help = restraint factor in mm for detector panel translations in X,Y,Z close_distances = None .type = float .help = restraint factor for the spread of detector panel Z-distances (#TODO think about this in context of tilt) } fix { panel_rotations = 0,0,0 .type = ints(size=3) .help = refinement flags, 1 means to fix the parameter panel_translations = 0,0,0 .type = ints(size=3) .help = refinement flags, 1 means to fix the parameter } } """ predictions_phil = """ predictions { use_peak_detection = False .type = bool .help = If True, then simulations will be converted to refl tables .help = where each reflection is a single pixel corresponding .help = to the peak in the simulated spot. This is useful if one .help = models spatial overlaps with separate peaks verbose = False .type = bool .help = See more console output from prediction methods laue_mode = False .type = bool .help = if True, predict the per-pixel wavelenth from the model and use that for index assigment qcut = 0.1 .type = float .help = Label predicted reflection as a strong reflection if its within this .help = many inverse Angstromg (q=2/Lambda sin(theta)) of an observed strong spot label_weak_col = "xyzobs.px.value" .type = str weak_fraction = 0.5 .type = float .help = fraction of weak predictions to integrate threshold = 1e-3 .type = float .help = value determining the cutoff for the forward model intensity. Bragg peaks will then be determined .help = as regions of connected values greater than the threshold thicksteps_override = None .type = int .help = Use to force the number of detector thickness steps to a specific value oversample_override = None .type = int .help = force the pixel oversample rate to this value during the forward model simulation .help = for maximum speed gains, set this to 1, but inspect the output! .expert_level=10 use_diffBragg_mtz = False .type = bool .help = whether to use the mtz supplied to diffBragg for prediction Nabc_override = None .type = ints(size=3) .help = use this value of mosaic block size for every shot, useful to get more predicted spots .expert_level=10 pink_stride_override = None .type = int .help = if specified, stride through the spectrum according to this interval default_Famplitude = 1e3 .type = float .help = default structure factor amplitude for every miller index .help = this creates a flat prediction model, where the magnitude is dependent on the distance to the Ewald sphere resolution_range = [1,999] .type = floats(size=2) .help = high-res to low-res limit for prediction model symbol_override = None .type = str .help = specify the space group symbol to use in diffBragg (e,g, P43212), .help = if None, then it will be pulled from the crystal model method = *diffbragg exascale .type = choice .help = engine used for computing the forward model .help = diffbragg offers CUDA support via the DIFFBRAGG_USE_CUDA=1 environment variable specification .help = or openmp support using the OMP_NUM_THREADS flag .help = The exascale only uses CUDA (will raise error if CUDA is not confugured) } """ philz = simulator_phil + refiner_phil + roi_phil + predictions_phil phil_scope = parse(philz)