- Aug 09, 2018
-
-
Joseph Weston authored
'box-forced' is deprecated as of matplotlib 2.2.
-
- Aug 08, 2018
-
-
Joseph Weston authored
Gitlab now has built-in mirroring, so we don't need to have a CI step for this now.
-
- Aug 06, 2018
-
-
Christoph Groth authored
-
- Aug 02, 2018
-
-
Anton Akhmerov authored
Closes #207. pyplot "attaches" a figure to a backend on calling pyplot.figure. Previously we implemented a hack that worked around this limitation in case the user did not import pyplot. However this hack is fragile and breaks new things with newer matplotlib versions. Now we detect whether the figure is intended for pyplot and import pyplot directly if it is. This change also allows us to simplify the interface and the implementation of output_fig, but this is permitted because it is an internal function.
-
Joseph Weston authored
clarify usage of pseudoinverse, and factor out rank checks for primitive lattice vectors Closes #209. See merge request kwant/kwant!231
-
-
Currently code that do constraints check on "prim_vecs" during lattice creation is doubled. In addition to factoring it out into separate function this commit also adds check if "prim_vecs" are linearly independent.
-
- Aug 01, 2018
-
-
Joseph Weston authored
This was always what was meant semantically; newer versions of Scipy are stricter about what they accept for row/column indices for COO matrices. Closes #211.
-
Joseph Weston authored
In the future numpy will not allow multidimensional slices that are specified as sequences other than tuples. Closes #210.
-
- Jul 31, 2018
-
-
Joseph Weston authored
Fix compatibility issues with Sympy 1.2. Closes #203. See merge request kwant/kwant!228
-
SymPy 1.2 brought changes to LambdaPrinter that broke kwant.continuum for inputs containing functions like "sin", "cos", or constant "pi". By this fix we assert that "kwant.continuum.sympify" represents in its output all functions as AppliedUndef in favour to built-in SymPy functions. We also put a hard constaint that all functions and constants in the input expression (with except to pi) must be provided through params.
-
-
Joseph Weston authored
Backport of commit 499f3d91 to stable branch. Closes #208.
-
- Jul 05, 2018
-
-
Rafal Skolasinski authored
-
- Jun 20, 2018
-
-
When passing `energy_resolution`, the number of moments is set using the yet undefined parameter `self._a`. Now the number of moments is set after defining `self._a`.
-
- Mar 08, 2018
-
-
Joseph Weston authored
-
- Mar 01, 2018
-
-
Joseph Weston authored
If 'mplot3d' is not available then plotting 2D systems fails because 'plot' calls 'set_colors', which references 'mplot3d'.
-
- Feb 28, 2018
-
-
Joseph Weston authored
The check is meant to make sure we don't use a hopping length that is smaller than a certain threshold (10E-5 previously). Here we correct the check and increase the threshold so as not to have too fine-grained gridding.
-
Christoph Groth authored
Thanks to Jörg Behrmann for reporting this problem!
-
- Feb 27, 2018
-
-
Joseph Weston authored
Add RELEASE.rst that documents the process of making a Kwant release. See merge request kwant/kwant!184
-
- Feb 26, 2018
-
-
Joseph Weston authored
Also include grammatical/stylistic changes, and a re-ordering of some sections.
-
-
-
-
-
Joseph Weston authored
-
Joseph Weston authored
-
Joseph Weston authored
Closes #188.
-
- Feb 23, 2018
-
-
Joseph Weston authored
-
- Feb 16, 2018
-
-
Running "kwant.continuum.discretize('k_x**2 + 1')" was resulting in an "AssertionError". This happend because "1 * expr = expr" in SymPy and therefore a wrong argument has been passed to "_read_offset" function.
-
Joseph Weston authored
While we *may* choose only to specify that finalize() returns a system.Finitesystem (or system.InfiniteSystem), this is probably a bad idea, as the docstring of finalize() is the only place that people can find out about builder.FiniteSystem (which contains the documentation for the 'sites' attribute, for example).
-
- Feb 08, 2018
-
-
Christoph Groth authored
-
- Feb 07, 2018
-
-
Bas Nijholt authored
-
Bas Nijholt authored
-
Bas Nijholt authored
-
- Feb 06, 2018
-
-
Christoph Groth authored
Matplotlib backends like PDF cannot display arbitrarily thin lines. That's why kwant.plotter.streamplot simulates thinner lines by blending them into the background. Previously if no lines were blended, i.e. all had the same color value 1, the background color would be used for drawing them instead of the line color.
-
- Jan 12, 2018
-
-
Closes #182
-
- Jan 11, 2018
-
-
Bas Nijholt authored
-
- Jan 05, 2018
-
-
Christoph Groth authored
-
- Dec 15, 2017
-
-
Dániel Varjas authored
This commit fixes the bug in phs_symmetrization when it fails to produce ph eigenstates with PH^2 = +1 when one of the wave functions is mapped exactly onto another one by PH, as in this case the PH symmetrized wf's from the two are identical. This is the case with wfs = np.eye(2) and particle_hole = sigma[1], the second vector returned is an eigenstate with eigenvalue -1 not +1. In general this can be even worse, if there are multiple such vector pairs some vectors may not be ph eigenstates at all. The new code uses an improved version of the matrix square root method from Applied Mathematics and Computation 234 (2014) 380-384. Tests are modified to clear up the random unitary generation process (using kwant.rmt). Tests are now guaranteed to feed orthonormal set of vectors to phs_symmetrization(). Add extra tests for the fully off-diagonal case where the old method fails.
-