Points class

class IMTreatment.core.points.Points(xy=array([], shape=(0, 2), dtype=float64), v=[], unit_x='', unit_y='', unit_v='', name='')

Bases: object

add(pt, v=None)

Add a new point.

Parameters:

• pt (2x1 array of numbers) – Point to add.
• v (number, optional) – Value of the point (needed if other points have values).

augment_resolution(fact=2, interp='linear', inplace=False)

Augment the temporal resolution of the points. Only have sense if points are sorted to set some kind of trajectory.

Parameters:

• fact (integer) – Resolution augmentation needed (default is ‘2’, for a result profile with twice more points)
• interp (string in ['linear', 'nearest', slinear', 'quadratic, 'cubic']) – Specifies the kind of interpolation as a string (Default is ‘linear’). slinear’, ‘quadratic’ and ‘cubic’ refer to a spline interpolation of first, second or third order.
• bool (inplace) –

  .

Note

If masked values are present, they are interpolated as well, using the surrounding values.

change_unit(axe, new_unit)

Change the unit of an axe.

Parameters:

• axe (string) – ‘y’ for changing the profile y axis unit ‘x’ for changing the profile x axis unit ‘v’ for changing the profile values unit
• new_unit (Unum.unit object or string) – The new unit.

copy()

Return a copy of the Points object.

crop(intervx=None, intervy=None, intervv=None, inplace=True, ind=False)

Crop the points cloud.

Parameters:

• intervx (2x1 tuple) – Interval on x axis
• intervy (2x1 tuple) – Interval on y axis
• intervv (2x1 tuple) – Interval on v values

Returns:

tmp_pts – croped version of the point cloud.

Return type:

Points object

cut(intervx=None, intervy=None)

Return a point cloud where the given area has been removed.

Parameters:

• intervx (2x1 tuple) – Interval on x axis
• intervy (2x1 tuple) – Interval on y axis

Returns:

tmp_pts – Cutted version of the point cloud.

Return type:

Points object

decompose()

return a tuple of Points object, with only one point per object.

display(kind=None, axe_x=None, axe_y=None, axe_color=None, **plotargs)

Display the set of points.

Parameters:

• kind (string, optional) – Can be ‘plot’ (default if points have not values). or ‘scatter’ (default if points have values). or ‘colored_plot’.
• axe_y, axe_color (axe_x,) – To determine wich value has to be plotted along which axis, and whith value is used to color the scattered points. Default plot ‘y’ to ‘x’ with colors from ‘v’.
• **plotargs (dic) – Additionnal arguments sent to ‘plot’ or ‘scatter’

display3D(kind='plot', xlabel='', ylabel='', zlabel='', title='', **plotargs)

Display the points on a 3D graph.

Parameters:

• kind (string, optional) – Kind of graph to use, can be ‘plot’ or ‘surf’.
• ylabel, zlabel (xlabel,) – Label fo each axis (respectively ‘x’, ‘y’, and ‘v’)
• title (strin, optional) – Title
• **plotargs – Additional parameters feeded to matplotlib

export_to_profile(axe_x='x', axe_y='y')

Export the unsorted point object to a sorted Profile object.

Parameters:axe_y (axe_x,) – Which value used to construct the profile

fit(kind='polynomial', order=2, simplify=False)

Return the parametric coefficients of the fitting curve on the points.

Parameters:

• kind (string, optional) – The kind of fitting used. Can be ‘polynomial’ or ‘ellipse’.
• order (integer) – Approximation order for the fitting.
• Simplify (boolean or string, optional) – Can be False (default), ‘x’ or ‘y’. Perform a simplification (see Points.Siplify()) before the fitting.

Returns:

• p (array, only for polynomial fitting) – Polynomial coefficients, highest power first
• radii (array, only for ellipse fitting) – Ellipse demi-axes radii.
• center (array, only for ellipse fitting) – Ellipse center coordinates.
• alpha (number) – Angle between the x axis and the major axis.

get_clusters(eps, min_samples=5)

Perform DBSCAN clustering from vector array or distance matrix. (see sklearn.cluster.DBSCAN)

Notes

DBSCAN - Density-Based Spatial Clustering of Applications with Noise. Finds core samples of high density and expands clusters from them. Good for data which contains clusters of similar density.

get_envelope(alpha=None)

Return the convex or concave hull (if alpha specified) for the set of points.

Parameters:alpha (number) – maximum distance between two points of the hull.

Notes

Credit to mlaloux (https://github.com/mlaloux/Python–alpha-shape_concave_hull)

get_evolution_on_sf(SF, axe_x=None)

Return the evolution of the value represented by a scalar field, on the path of the trajectory.

Parameters:

• SF (ScalarField object) –
• axe_x (string, optional) – What put in the x axis (can be ‘x’, ‘y’, ‘v’). default is ‘v’ when available and ‘x’ else.

Returns:

evol

Return type:

Profile object

get_evolution_on_tsf(TSF, axe_x=None)

Return the evolution of the value represented by scalar fields, on the path of the trajectory. Timse of the TSF must be consistent with the times of the Points.

Parameters:

• TSF (tsf.TemporalScalarField object) –
• axe_x (string, optional) – What put in the x axis (can be ‘x’, ‘y’, ‘v’). default is ‘v’ (associated with time)

Returns:

evol

Return type:

Profile object

get_points_density(bw_method=None, resolution=100, output_format='normalized', raw=False)

Return a ScalarField with points density.

Parameters:

• bw_method (str, scalar or callable, optional) – The method used to calculate the estimator bandwidth. This can be ‘scott’, ‘silverman’, a scalar constant or a callable. If a scalar, this will be used as percent of the data std. If a callable, it should take a gaussian_kde instance as only parameter and return a scalar. If None (default), ‘scott’ is used.
• resolution (integer, optional) – Resolution for the resulting field.
• output_format (string, optional) –

  ‘normalized’ (default) : give position probability

  (integral egal 1).

  ’ponderated’ : give position probability ponderated by the number

  or points (integral egal number of points).

  ’concentration’ : give local concentration (in point per surface).

• raw (boolean, optional) – If ‘False’ (default), return a ScalarField object, if ‘True’, return numpy array.

Returns:

density – Return ‘None’ if there is not enough points in the cloud.

Return type:

array, ScalarField object or None

get_points_density2(res, subres=None, raw=False, ponderated=False)

Return a ScalarField with points density.

Parameters:

• res (number or 2x1 array of numbers) – fdensity field number of subdivision. Can be the same number for both axis, or one number per axis (need to give a tuple).
• raw (boolean, optional) – If ‘False’ (default), return a ScalarField object, if ‘True’, return numpy array.
• ponderated (boolean, optiona) – If ‘True’, values associated to points are used to ponderate the density field. Default is ‘False’.
• subres (odd integer, optional) – If specified, a subgrid of resolution res*subres is used to make result more accurate.

get_velocity(incr=1, smooth=0, xaxis='time')

Assuming that associated ‘v’ values are times for each points, compute the velocity of the trajectory.

Parameters:

• incr (integer, optional) – Increment use to get used points (default is 1).
• smooth (number, optional) – Cut off frequency for the lowpass filter.
• xaxis (string, optional) – Value to put in the profile x axis, can be ‘time’ (default), ‘x’ or ‘y’.

Returns:

• Vx (Profile object) – Profile of x velocity versus time.
• Vy (Profile object) – Profile of y velocity versus time.

name

order_on_line(inplace=False)

Re-order the set of points to try to create a line.

remove(ind)

Remove the point number ‘ind’ of the points cloud. In place.

Parameters:ind (integer or array of integer) –

remove_doublons(method='average', inplace=False, eps_rel=1e-06)

Replace values associated to the same ‘v’ by their average.

Parameters:method (string in {'average', 'max', 'min'}) – Method used to remove the doublons.

remove_nans(inplace=False)

Remove the points containing nans values.

reverse()

Return a Points object where x and y axis are swaped.

rotate(angle, inplace=False)

Rotate the point set.

Parameters:angle (number) – Rotation angle in radian.

scale(scalex=1.0, scaley=1.0, scalev=1.0, inplace=False)

Change the scale of the axis.

Parameters:

• scaley, scalev (scalex,) – scales along x, y and v
• inplace (boolean, optional) – If ‘True’, scaling is done in place, else, a new instance is returned.

set_origin(x=None, y=None)

Set the given point (x, y) as the new referential.

Parameters:

• x (number) –

  .

• y (number) –

  .

smooth(tos='gaussian', size=None, inplace=False, **kw)

Return a smoothed points field.

Parameters:

• tos (string, optional) – Type of smoothing, can be ‘uniform’ (default), ‘gaussian’ or ‘lowpass’.
• size (number, optional) – radius of the smoothing for ‘uniform’, radius of the smoothing for ‘gaussian’, cut off frequency for ‘lowpass’ Default are 3 for ‘uniform’, 1 for ‘gaussian’ and 0.1 for ‘ lowpass’.
• inplace (boolean) – If ‘False’, return a smoothed points field else, smooth in place.
• kw (dic) – Additional parameters for ndimage methods (See ndimage documentation)

sort(ref='x', inplace=False)

Sort the points according to the reference.

Parameters:

• ref (string or array of indice) – can be ‘x’, ‘y’ or ‘v’ to sort according those values or an array of indice
• ; boolean (inplace) – If ‘True’, sort in place, else, return an new sorted instance.

unit_v

unit_x

unit_y

v

xy