PYME.recipes.localisations module

class PYME.recipes.localisations.AddMeasurementsByLabel(parent=None, invalidate_parent=True, **kwargs)

Bases: ModuleBase

input_measurements = <PYME.recipes.traits.Input object>

input_points = <PYME.recipes.traits.Input object>

label_key = <PYME.misc.mock_traits.CStr object>

output_points = <PYME.recipes.traits.Output object>

run(input_points, input_measurements)

Propagate measurements from e.g. MeasureClusters3D back to points they were calcualted from. This is particularly useful for visualizing e.g. the gyration radius of a cluster in PYMEVis while looking at the original localization data.

Parameters

input_pointsPYME.IO.tabular.TabularBase

points used to generate the measurements

input_measurementsPYME.IO.tabular.TabularBase

measurements to propagate back to the points

Returns

PYME.IO.tabular.TabularBase

point data with new columns added, one for each scalar measurement in input_measurements, which can be accessed by ‘<label_key>_<measurement_key>’, e.g. ‘clumpIndex_gyrationRadius’.

class PYME.recipes.localisations.AutocorrelationDriftCorrection(parent=None, invalidate_parent=True, **kwargs)

Bases: ModuleBase

Perform drift correction using autocorrelation between subsets of the point data

binsize = <PYME.misc.mock_traits.Float object>

calcCorrDrift(x, y, t)

inputName = <PYME.recipes.traits.Input object>

outputName = <PYME.recipes.traits.Output object>

run(inputName)

OVERRIDE THIS in derived classes to implement module functionality. Parameters are passed in by keyword, and keyword names must be a 1:1 match to the module input traits.

The function should return either a dict (multiple outputs, keyed with output trait keys), or

step = <PYME.misc.mock_traits.CStr object>

window = <PYME.misc.mock_traits.CStr object>

class PYME.recipes.localisations.ClusterCountVsImagingTime(parent=None, invalidate_parent=True, **kwargs)

Bases: ModuleBase

WARNING: This module will likely move, dissapear, or be refactored

ClusterCountVsImagingTime iteratively filters a dictionary-like object on t, and at each step counts the number of labeled objects (e.g. DBSCAN clusters) which contain at least N-points. It does this for two N-points, so one can be set according to density with all frames included, and the other can be set for one of the earlier frame-counts.

args:

stepSize: number of frames to add in on each iteration labelsKey: key containing labels for each localization lowerMinPtsPerCluster: higherMinPtsPerCluster:

returns:

dictionary-like object with the following keys:

t: upper bound on frame number included in calculations on each iteration. N_labelsWithLowMinPoints: N_labelsWithHighMinPoints:

From wikipedia: “While minPts intuitively is the minimum cluster size, in some cases DBSCAN can produce smaller clusters. A DBSCAN cluster consists of at least one core point. As other points may be border points to more than one cluster, there is no guarantee that at least minPts points are included in every cluster.”

higherMinPtsPerCluster = <PYME.misc.mock_traits.CStr object>

inputName = <PYME.recipes.traits.Input object>

labelsKey = <PYME.misc.mock_traits.CStr object>

lowerMinPtsPerCluster = <PYME.misc.mock_traits.CStr object>

outputName = <PYME.recipes.traits.Output object>

run(inputName)

OVERRIDE THIS in derived classes to implement module functionality. Parameters are passed in by keyword, and keyword names must be a 1:1 match to the module input traits.

The function should return either a dict (multiple outputs, keyed with output trait keys), or

stepSize = <PYME.misc.mock_traits.CStr object>

class PYME.recipes.localisations.DBSCANClustering(parent=None, invalidate_parent=True, **kwargs)

Bases: ModuleBase

Performs DBSCAN clustering on input dictionary

Parameters

searchRadiusfloat

search radius for clustering [nm]

minPtsForCoreint

number of points within SearchRadius required for a given point to be considered a core point

Notes

See sklearn.cluster.dbscan for more details about the underlying algorithm and parameter meanings.

clumpColumnName = <PYME.misc.mock_traits.CStr object>

columns = <PYME.misc.mock_traits.CStr object>

property hide_in_overview

inputName = <PYME.recipes.traits.Input object>

minClumpSize = <PYME.misc.mock_traits.CStr object>

multiprocessing = <module 'multiprocessing' from '/opt/hostedtoolcache/Python/3.7.17/x64/lib/python3.7/multiprocessing/__init__.py'>

multithreaded = <PYME.misc.mock_traits.CStr object>

numberOfJobs = <PYME.misc.mock_traits.CStr object>

outputName = <PYME.recipes.traits.Output object>

run(inputName)

OVERRIDE THIS in derived classes to implement module functionality. Parameters are passed in by keyword, and keyword names must be a 1:1 match to the module input traits.

The function should return either a dict (multiple outputs, keyed with output trait keys), or

searchRadius = <PYME.misc.mock_traits.Float object>

class PYME.recipes.localisations.DensityMapping(parent=None, invalidate_parent=True, **kwargs)

Bases: ModuleBase

Use density estimation methods to generate an image from localizations

MCProbability = <PYME.misc.mock_traits.Float object>

available_renderers = ['3D Gaussian', '3D Histogram', '3D Triangularisation', 'Density Fit', 'Gaussian', 'Histogram', 'Jittered Triangulation', 'Jittered Triangulation - weighted', 'Log Likelihood', 'QuadTree', 'Voronoi']

colours = <PYME.misc.mock_traits.List object>

inputLocalizations = <PYME.recipes.traits.Input object>

jitterScale = <PYME.misc.mock_traits.Float object>

jitterScaleZ = <PYME.misc.mock_traits.Float object>

jitterVariable = <PYME.misc.mock_traits.CStr object>

jitterVariableZ = <PYME.misc.mock_traits.CStr object>

manualXYBounds = <PYME.misc.mock_traits.List object>

numSamples = <PYME.misc.mock_traits.CStr object>

outputImage = <PYME.recipes.traits.Output object>

pixelSize = <PYME.misc.mock_traits.Float object>

renderingModule = <PYME.misc.mock_traits.CStr object>

run(inputLocalizations)

OVERRIDE THIS in derived classes to implement module functionality. Parameters are passed in by keyword, and keyword names must be a 1:1 match to the module input traits.

The function should return either a dict (multiple outputs, keyed with output trait keys), or

softRender = <PYME.misc.mock_traits.CStr object>

xyBoundsMode = <PYME.misc.mock_traits.CStr object>

zBounds = <PYME.misc.mock_traits.List object>

zBoundsMode = <PYME.misc.mock_traits.CStr object>

zSliceThickness = <PYME.misc.mock_traits.Float object>

class PYME.recipes.localisations.ExtractTableChannel(parent=None, invalidate_parent=True, **kwargs)

Bases: ModuleBase

Create and return a ColourFilter which has filtered out one colour channel from a table of localizations.

channel = <PYME.misc.mock_traits.CStr object>

inputName = <PYME.recipes.traits.Input object>

outputName = <PYME.recipes.traits.Output object>

run(inputName)

OVERRIDE THIS in derived classes to implement module functionality. Parameters are passed in by keyword, and keyword names must be a 1:1 match to the module input traits.

The function should return either a dict (multiple outputs, keyed with output trait keys), or

class PYME.recipes.localisations.FiducialCorrection(parent=None, invalidate_parent=True, **kwargs)

Bases: ModuleBase

Maps each point in the input table to a pixel in a labelled image, and extracts the pixel value at that location to use as a label for the point data.

clumpRadiusMultiplier = <PYME.misc.mock_traits.Float object>

clumpRadiusVar = <PYME.misc.mock_traits.CStr object>

inputFiducials = <PYME.recipes.traits.Input object>

inputLocalizations = <PYME.recipes.traits.Input object>

outputFiducials = <PYME.recipes.traits.Output object>

outputName = <PYME.recipes.traits.Output object>

run(inputLocalizations, inputFiducials)

OVERRIDE THIS in derived classes to implement module functionality. Parameters are passed in by keyword, and keyword names must be a 1:1 match to the module input traits.

The function should return either a dict (multiple outputs, keyed with output trait keys), or

temporalFilter = <PYME.misc.mock_traits.CStr object>

temporalFilterScale = <PYME.misc.mock_traits.Float object>

timeWindow = <PYME.misc.mock_traits.CStr object>

class PYME.recipes.localisations.IDTransientFrames(parent=None, invalidate_parent=True, **kwargs)

Bases: ModuleBase

Adds an ‘isTransient’ column to the input datasource so that one can filter localizations that are from frames acquired during z-translation

framesPerStep = <PYME.misc.mock_traits.Float object>

inputEvents = <PYME.recipes.traits.Input object>

inputName = <PYME.recipes.traits.Input object>

outputName = <PYME.recipes.traits.Output object>

run(inputName, inputEvents)

OVERRIDE THIS in derived classes to implement module functionality. Parameters are passed in by keyword, and keyword names must be a 1:1 match to the module input traits.

The function should return either a dict (multiple outputs, keyed with output trait keys), or

class PYME.recipes.localisations.LabelsFromImage(parent=None, invalidate_parent=True, **kwargs)

Bases: ModuleBase

Maps each point in the input table to a pixel in a labelled image, and extracts the pixel value at that location to use as a label for the point data.

inputImage = <PYME.recipes.traits.Input object>

inputName = <PYME.recipes.traits.Input object>

label_count_key_name = <PYME.misc.mock_traits.CStr object>

label_key_name = <PYME.misc.mock_traits.CStr object>

minimum_localizations = <PYME.misc.mock_traits.CStr object>

outputName = <PYME.recipes.traits.Output object>

run(inputName, inputImage)

OVERRIDE THIS in derived classes to implement module functionality. Parameters are passed in by keyword, and keyword names must be a 1:1 match to the module input traits.

The function should return either a dict (multiple outputs, keyed with output trait keys), or

class PYME.recipes.localisations.MapAstigZ(parent=None, invalidate_parent=True, **kwargs)

Bases: ModuleBase

Uses astigmatism calibration (widths of PSF along each dimension as a function of z) to determine the z-position of localizations relative to the focal plane of the frame during which it was imaged. Note this is the single channel version and does not work for combined biplane and astigmatism (see Multiview.MapAstigZ). It also uses the more robust but less precise dsigma method.

Parameters

input_nametraits.Input

localizations as PYME.IO.Tabular types

astigmatism_calibration_locationtraits.File

file path or URL to astigmatism calibration file

Returns

output_nametraits.Output

output is a json wrapped by PYME.IO.ragged.RaggedCache

astigmatism_calibration_location = <PYME.recipes.traits.FileOrURI object>

execute(namespace)

takes a namespace (a dictionary like object) from which it reads its inputs and into which it writes outputs

NOTE: This was previously the function to define / override to make a module work. To support automatic metadata propagation and reduce the ammount of boiler plate, new modules should override the run() method instead.

input_name = <PYME.recipes.traits.Input object>

output_name = <PYME.recipes.traits.Output object>

rough_knot_spacing = <PYME.misc.mock_traits.Float object>

z_scale = <PYME.misc.mock_traits.Float object>

class PYME.recipes.localisations.MeasureClusters3D(parent=None, invalidate_parent=True, **kwargs)

Bases: ModuleBase

Measures the 3D morphology of clusters of points

Parameters

labelKey: name of column to use as a label identifying clusters

Notes

Measures calculated (to be expanded)

Count

int Number of localizations (points) in the cluster

X

float x center of mass

Y

float y center of mass

Z

float z center of mass

:gyrationRadiusfloat

root mean square displacement to center of cluster, a measure of compaction or spatial extent see also supplemental text of DOI: 10.1038/nature16496

:axis0ndarray, shape (3,)

principle axis which accounts for the largest variance of the cluster, i.e. corresponds to the largest eigenvalue

:axis1ndarray, shape (3,)

next principle axis

:axis2ndarray, shape (3,)

principle axis corresponding to the smallest eigenvalue

:sigma0float

standard deviation along axis0

:sigma1float

standard deviation along axis1

:sigma2float

standard deviation along axis2

:anisotropyfloat

metric of anisotropy based on the spread along principle axes. Standard deviations of alpha * [1, 0, 0], where alpha is a scalar, will result in an ‘anisotropy’ value of 1, i.e. maximally anisotropic. Completely isotropic clusters will have equal standard deviations, i.e. alpha * [1, 1, 1], which corresponds to an ‘anisotropy’ value of 0. Intermediate cases result in values between 0 and 1.

:thetafloat

Azimuthal angle, in radians, along which the principle axis (axis0) points

:phifloat

Zenith angle, in radians, along which the principle axis (axis0) points

inputName = <PYME.recipes.traits.Input object>

labelKey = <PYME.misc.mock_traits.CStr object>

outputName = <PYME.recipes.traits.Output object>

run(inputName)

OVERRIDE THIS in derived classes to implement module functionality. Parameters are passed in by keyword, and keyword names must be a 1:1 match to the module input traits.

The function should return either a dict (multiple outputs, keyed with output trait keys), or

class PYME.recipes.localisations.MergeClumps(parent=None, invalidate_parent=True, **kwargs)

Bases: ModuleBase

Create a new mapping object which derives mapped keys from original ones

discardTrivial = <PYME.misc.mock_traits.CStr object>

inputName = <PYME.recipes.traits.Input object>

labelKey = <PYME.misc.mock_traits.CStr object>

outputName = <PYME.recipes.traits.Output object>

run(inputName)

OVERRIDE THIS in derived classes to implement module functionality. Parameters are passed in by keyword, and keyword names must be a 1:1 match to the module input traits.

The function should return either a dict (multiple outputs, keyed with output trait keys), or

class PYME.recipes.localisations.Pipelineify(parent=None, invalidate_parent=True, **kwargs)

Bases: ModuleBase

Perform standard mappings, including those derived from acquisition events.

Parameters

inputFitResultsstring - the name of a tabular.TabularBase object

Typically the FitResults table of an h5r file

inputEventsstring - name of a tabular.TabularBase object containing acquisition events [optional]

This is not usually required as the IO methods attach .events as a datasource attribute. Use when events come from a separate file or when there are intervening processing steps between IO and this module (the .events attribute does not propagate through recipe modules). TODO - do we really want to be attaching events as an attribute or should they be there own entry in the recipe namespace TODO - should we change this to the processed events???

pixelSizeNMfloat

Scaling factor to get ‘x’ and ‘y’ into units of nanometers. Useful if handling external data input in pixel units. Defaults to 1.

Returns

outputLocalizationstabular.MappingFilter

foreshortening = <PYME.misc.mock_traits.Float object>

inputEvents = <PYME.recipes.traits.Input object>

inputFitResults = <PYME.recipes.traits.Input object>

outputLocalizations = <PYME.recipes.traits.Output object>

pixelSizeNM = <PYME.misc.mock_traits.Float object>

run(inputFitResults, inputEvents=None)

OVERRIDE THIS in derived classes to implement module functionality. Parameters are passed in by keyword, and keyword names must be a 1:1 match to the module input traits.

The function should return either a dict (multiple outputs, keyed with output trait keys), or

class PYME.recipes.localisations.ProcessColour(parent=None, invalidate_parent=True, **kwargs)

Bases: ModuleBase

input = <PYME.recipes.traits.Input object>

output = <PYME.recipes.traits.Output object>

ratios_from_metadata = <PYME.misc.mock_traits.CStr object>

run(input)

OVERRIDE THIS in derived classes to implement module functionality. Parameters are passed in by keyword, and keyword names must be a 1:1 match to the module input traits.

The function should return either a dict (multiple outputs, keyed with output trait keys), or

species_dyes = <PYME.misc.mock_traits.CStr object>

species_ratios = <PYME.misc.mock_traits.CStr object>

class PYME.recipes.localisations.TimeBlocks(parent=None, invalidate_parent=True, **kwargs)

Bases: ModuleBase

Divides series into alternating time blocks to generate 2 fake colour channels for Fourier Ring / Fourier shell correlation.

This is probably a better approach than taking random subsets as the later will tend to generate unrealistically high correlation values for repeated localizations.

Adapted from Christian Soeller’s ‘splitRender’ implementation.

block_size = <PYME.misc.mock_traits.CStr object>

input = <PYME.recipes.traits.Input object>

output = <PYME.recipes.traits.Output object>

run(input)

OVERRIDE THIS in derived classes to implement module functionality. Parameters are passed in by keyword, and keyword names must be a 1:1 match to the module input traits.

The function should return either a dict (multiple outputs, keyed with output trait keys), or