gvar version 8.3

New utility gvar.dataset.svd_diagnosis for determining SVD cuts for correlation matrices estimated from a collection of random samples. See the documentation page Case Study: Correlations and SVD Cuts for more information. A couple of minor fixes.

gvar version 8.2.3

Fixes problems in BufferDict and dataset caused by changes in Python's dictionaries starting with Python3.5. The fixes have been tested in Python 2.7, 3.4, 3.5, and 3.6.

gvar version 8.2.2

Minor/rare bug fixes to SVD and GVar.str.

gvar version 8.2.1

Bug fix for gvar.gvar to remove problems caused when a GVar had exactly 0 standard deviation. Minor refinement in gvar.dataset.Dataset in its handling of hdf5 files.

gvar version 8.2

Minor additions: hdf5 format support in gvar.dataset.Dataset; del g[k] for BufferDict g; small bug fixes.

gvar version 8.1

More efficient handling of large sparse covariance matrices in gvar.svd (see gvar.evalcov_blocks()). Minor bug fixes and extensions.

gvar version 8.0

Moves PDFIntegrator to vegas; fixes integer type problem on windows machines.

gvar version 7.3.1

Optimization for PDFIntegrator that affects performance for integrals over a large (eg, 100) variables.

gvar version 7.3

New class PDFIntegrator for integrating over PDFs and computing expectation values. Bug fixes and other improvements.

gvar version 7.2

Adds correlate(g, corr) for building correlated GVars using correlation matrices. Small bug fixes to evalcov and evalcorr.