What’s new¶
A catalog of new features, improvements, and bug-fixes in each release. Follow links to the relevant GitHub issue or pull request for specific code changes and any related discussion.
v0.0.15 (2019-06-06)¶
Merged pull requests:
#537 Turn element_list to a list of strings (Ksouravsingh)
v0.0.14 (2018-09-05)¶
Closed issues:
2D SavitzkyGolay smoothing algorithm #531
Merged pull requests:
v0.0.13 (2018-01-10)¶
Closed issues:
Cythonize binned statistic? #519
Math is not rendering properly in the docs Documentation #516
bug in lazy_one_time #512
Merged pull requests:
#530 ENH: speed up binned statistic by stashing np.argsort(xy) (CJ-Wright)
#524 ENH: speed up by caching bincount and argsort (CJ-Wright)
#518 ENH: Added r_map option to RPhiBinnedStatistic (jrmlhermitte)
#517 Move some fitting related functions from pyxrf to skbeam (licode)
#513 ENH: added binmap to binned_statistic (jrmlhermitte)
v0.0.11 (2017-06-23)¶
Merged pull requests:
#494 New and improved Cross correlator (jrmlhermitte)
#489 ENH : added construct_rphi_image (jrmlhermitte)
#488 Automatically find the center and the rings (sameera2004)
#487 Spatial Correlations (jrmlhermitte)
v0.0.10 (2017-01-09)¶
Closed issues:
ENH: make
BinnedStatisticDD
statistic in call #481
Merged pull requests:
#482 Modify the computed statistic on calling BinnedStatisticDD instances (CJ-Wright)
#479 Include API docs for more of the modules (danielballan)
#478 Minor speed gains for skbeam.core.image.construct_circ_avg_image (ordirules)
#475 Make the image center a float and other associated changes (ordirules)
v0.0.9 (2016-09-23)¶
Closed issues:
ENH: save masks in Fit2D format #454
copy size-checks of pull-requests from pyqtgraph #420
Issues about fitting summarized on Monday meeting 10/27/2014 #184
Alignment algorithm for nano XRF tomography #97
Register 1-D curves: Fit, Scale & Shift #44
Automated builds with conda-forge #443
Merged pull requests:
#463 Add interpolation feature for masked circular averaged images (ordirules)
#462 Masking added to circular average as an option (ordirules)
#445 ENH: Updates to allow sequential processing (stuwilkins)
#435 ENH: Add edge and ring blur masks to scikit-beam (CJ-Wright)
#432 DOC: Fixed reference (stuwilkins)
#428 BUG: took out the repeated labels parameter (sameera2004)
#426 Roi more (sameera2004)
#422 Convert reciprocal space and 3d grinder to OpenMP (stuwilkins)
v0.0.8¶
PR #395: Compute one-time correlation from the results of two-time correlation
PR #401: Remove bad images from X-Ray Speckle Visibility Spectroscopy (XSVS)
PR #404: Implement gisaxs image conversion
PR #405: Implement user peak, so users have more flexible control of escape peak or pileup peaks.
PR #407: Clarify docstrings for correlation functions in skbeam.core.correlation
PR #408: Rework the example notebook conversion in the scikit-beam docs. The example notebooks are now grouped according to their folder in scikit-beam-examples
PR #414: Test for PEP8 compatibility on travis-ci. Enable flake8 coverage on scikit-beam. This will save code review time by allowing us to simply point at the failing flake8 test on travis-ci and say “Fix your code according to those things” instead of manually pointing out all the pep8 violations, like we do now.
PR #415: Run the test suite through the coverage run run_tests.py command. This does a better job of capturing coverage stats
PR #416: Build the docs as an extra matrix element on travis. This helps by reducing the sphinx output on travis-ci which makes it easier to find error messages in the test suite.
PR #417: Don’t compile the reciprocal space c extensions on windows. There are a multitude of errors in that code on the windows platform.
PR #419: Add functions to create box and horizontal/vertical bar regions of interest.
v0.0.7¶
Added multi-tau scheme two-time correlation. PR #391 This two-time correlator can be found at skbeam.core.correlation:two_time_corr. There is a generator version that returns its internal state after consuming each image, skbeam.core.correlation:lazy_two_time. The value that the generator yields should be passed to skbeam.core.correlation:two_time_state_to_results to produce the correlation results and the lag steps that the correlation results correspond to.
PR #400 Added removing bad images from multi-tau one time correlation.
Added new module skbeam.core.mask. It contains functions specific to mask or threshold an image basically to clean images. This module contains:
Convert the images marked as “bad” in bad list by their index in images into a np.nan array.
Set all pixels whose value is greater than threshold to 0 and yields the thresholded images out.
v0.0.6¶
Partial data, generator implementation of one-time-correlation
Rename from scikit-xray to scikit-beam and skxray->skbeam
Add cython implementation of accumulating histograms
Clean up the ctrans code
Add multiprocessing single pixel fitting
Automatically build the docs on traivs
v0.0.5¶
New Functionality¶
API Changes¶
weighted_nnls_fit()
was removed fromskbeam.core.fitting.xrf_model
. Weighted nnls fitting was combined intonnls_fit()
, which includes weights as a new argument.extract_label_indices()
is a helper function for labeled arrays and was moved to its new home in skbeam.core.roi from skbeam.core.correlation
Other updates¶
PR 316: Do a better job isolating dependencies so that our “optional” packages truly are optional
PR 319: Use latest lmfit version published to scikit-beam anaconda.org channel in travis build
PR 326: Add quick start guide and note about testing
PR 327: Pin to lmfit 0.8.3 in conda recipe
PR 332: Correct the equation in the one-time correlation docstring
PR 333: Update readme with new examples in scikit-beam-examples