Source code for binoculars.backends.example

import sys
import os
import itertools
import numpy

from .. import backend, errors, util


'''
This example backend contains the minimal set of functions needed to construct a backend.
It consists of a child of a backend.InputBase class and a child of a backend.ProjectionBase
class. The backend.Inputbase is collects the data from the measurement. The backend.ProjectionBase
class calculates the new coordinates per pixel.

You can write as much input classes and as much projections in one backend as you prefer, provided that
the output of the inputclass is compatible with projection class. Otherwise you will be
served best by writing a new backend, for the incompatibility will create errors that break the script.
In the configuration file you specify the inputclass and projection needed for the treatment of the dataset.
'''


[docs]class QProjection(backend.ProjectionBase):
[docs] def project(self, wavelength, af, delta, omega, ai): ''' This class takes as input the tuple of coordinates returned by the process_job method in the backend.InputBase class. Here you specify how to project the coordinates that belong to every datapoint. The number of input arguments should match the second tuple returned by process_job. The shape of each returned array should match the shape of the first argument returned by process_job ''' k0 = 2 * numpy.pi / wavelength qy = k0 * (numpy.cos(af) * numpy.cos(delta) - numpy.cos(ai) * numpy.cos(omega)) qx = k0 * (numpy.cos(af) * numpy.sin(delta) - numpy.cos(ai) * numpy.sin(omega)) qz = k0 * (numpy.sin(af) + numpy.sin(ai)) return (qx.flatten(), qy.flatten(), qz.flatten()) # a tuple of numpy.arrays with the same dimension as the number of labels
[docs] def get_axis_labels(self): ''' Specify the names of the axes. The number of labels should be equal to the number of arrays returned in the project method. ''' return 'qx', 'qy', 'qz'
[docs]class Input(backend.InputBase):
[docs] def generate_jobs(self, command): ''' Command is supplied when the program is started in the terminal. This can used to differentiate between separate datasets that will be processed independently. ''' scans = util.parse_multi_range(','.join(command).replace(' ', ','))# parse the command for scanno in scans: yield backend.Job(scan=scanno)
[docs] def process_job(self, job): ''' This methods is a generator that returns the intensity, the weights and a tuple of coordinates that will be used for projection. The input is a backend.job object. This objects contains attributes that are supplied as keyword arguments in the generate_jobs method when backend.Job is instantiated. You can wet here the weights according the behaviour of your detector. To select normal averaging give the weights the value of ones. This array should be the same shape as the intensity array. This example backend simulates a random path through angular space starting at the origin. an example image will be generated using a three dimensional 10-slit interference function. The angles are with respect to the sample where af and delta are the angular coordinates of the pixels and ai and omega are the in plane and out of plane angles of the incoming beam. ''' super(Input, self).process_job(job)# call super to fix metadeta handling scan = job.scan #reflects a scan with 100 datapoints aaf = numpy.linspace(0, numpy.random.random() * 20, 100) adelta = numpy.linspace(0, numpy.random.random() * 20, 100) aai = numpy.linspace(0, numpy.random.random() * 20, 100) aomega = numpy.linspace(0, numpy.random.random() * 20, 100) for af, delta, ai, omega in zip(aaf, adelta, aai, aomega): print('af: {0}, delta: {1}, ai: {2}, omega: {3}'.format(af, delta, ai, omega)) # caculating the angles per pixel. The values specified in the configuration file # can be used for calculating these values pixelsize = numpy.array(self.config.pixelsize) sdd = self.config.sdd app = numpy.arctan(pixelsize / sdd) * 180 / numpy.pi # create an image of 100 x 100 pixels and calculate the coordinates corresponding to every pixel centralpixel = self.config.centralpixel # (column, row) = (delta, af) af_range= -app[1] * (numpy.arange(100) - centralpixel[1]) + af delta_range= app[0] * (numpy.arange(100) - centralpixel[0]) + delta #calculating the coordinates for simulating the image. This is only included #in this example for simulating of the images. It has no other use. k0 = 2 * numpy.pi / self.config.wavelength delta, af = numpy.meshgrid(delta_range, af_range) ai *= numpy.pi/180 delta *= numpy.pi/180 af *= numpy.pi/180 omega *= numpy.pi/180 qy = k0 * (numpy.cos(af) * numpy.cos(delta) - numpy.cos(ai) * numpy.cos(omega)) qx = k0 * (numpy.cos(af) * numpy.sin(delta) - numpy.cos(ai) * numpy.sin(omega)) qz = k0 * (numpy.sin(af) + numpy.sin(ai)) #simulating the image data = numpy.abs(numpy.sin(qx * 10) / numpy.sin(qx) * numpy.sin(qy * 10) / numpy.sin(qy) * numpy.sin(qz * 10) / numpy.sin(qz))**2 weights = numpy.ones_like(data) yield data, weights, (self.config.wavelength, af, delta, omega, ai)
[docs] def parse_config(self, config): ''' To collect and process data you need the values provided in the configuration file. These you can access locally through the provided config object. This is a dict with as keys the labels given in the configfile. To use them outside the parse_config method you attribute them to the self.config object which can be used throughout the input class. A warning will be generated afterwards for config values not popped out of the dict. ''' super(Input, self).parse_config(config) self.config.sdd = float(config.pop('sdd')) self.config.pixelsize = util.parse_tuple(config.pop('pixelsize'), length=2, type=float) self.config.centralpixel = util.parse_tuple(config.pop('centralpixel'), length=2, type=int) self.config.wavelength = float(config.pop('wavelength'))
[docs] def get_destination_options(self, command): ''' Creates the arguments that you can use to construct an output filename. This method returns a dict object with keys that will can be used in the configfile. In the configfile the output filename can now be described as 'destination = demo_{first}-{last}.hdf5'. This helps to organise the output automatically. ''' if not command: return False command = ','.join(command).replace(' ', ',') scans = util.parse_multi_range(command) return dict(first=min(scans), last=max(scans), range=','.join(command))