v3.0.5

Fix generation of complex wavelets with up-to-date numpy.

v3.0.4

python 3.11 support

Version 3.0.1

What's Changed

• V3.0 by @jrudoler in #248

Full Changelog: v3.0.0...v3.0.1

Version 3.0.0

Most significant PTSA update in several years. In addition to refactoring and slight changes to the API, PTSA has been updated to leverage newer versions of Python and xarray.

Version 2.0.3

• TimeSeries.to_hdf and from_hdf have been reverted to the previous
  behavior of base 64-encoding Numpy recarrays before writing to disk. This
  makes the stored files less human-readable, but offers the advantage of being
  much more robust.

Version 2.0.2

New features:

• TimeSeries now has a filter_with method to make applying filters
  slightly easier (#211)

Fixes:

• Resource warnings were resolved by ensuring files get closed (#212)
• ButterworthFilter's freq_range argument is now coerced to a list (#214)

Other:

• Filter modules have been renamed to conform with PEP8 (#213). This should not
  affect normal usage since all filters are imported to the top-level filter
  namespace by default.

Version 2.0.0

Version 2.0 of PTSA is a major update which includes several breaking changes
including the removal of several packages and modules, name changes, and
deprecations.

PTSA 2.0 is the last major version to support loading of lab-specific data. In
future releases, all of this functionality will be removed in favor of using
the cmlreaders package which includes the ability to load timeseries data and
transform into a PTSA-compatible format.

Removals

The following packages and modules have been removed:

• ptsa.plotting - see https://github.com/pennmem/ptsa_plot instead (#147)
• ptsa.stats - most functionality now exists in SciPy (#147)
• ptsa.emd, ptsa.iwasobi, ptsa.ica and ptsa.wica (#147)
• ptsa.filt - filtering is now contained primarily in ptsa.data.filters
  (#158)
• Old time series class (fb08e6c)
• All modules within ptsa.data that worked with the deprecated time series
  implementation

Renames

• TimeSeriesX has been renamed to TimeSeries and is now found in the
  ptsa.data.timeseries module (#141). TimeSeriesX still exists as an
  alias but will give a DeprecationWarning.
• Most modules have been renamed to conform to PEP8 module naming conventions.
  These renames are generally not noticed by users if loading readers from the
  ptsa.data.readers namespace.

Other changes

• Reformatted output of TalReader.read
  • The atlases field is a nested structured array, rather than an array of dictionaries
  • Most idioms for getting coordinates and atlas labels for an electrode should continue to work in the new format
  • The channels field has been changed from an object array of lists to an array of two integers
• Added IPython notebook demonstrating reading localization information with TalReader (thanks @LoganJF!)
• Replaced custom typing system with the traits package (b2f4609)
• Improved TimeSeries.to_hdf (#138)

Version 1.1.4

Summary of changes:

• Removed cerr debugging output from compiled extension modules (#93)
• Changed to use h5py instead of pytables whenever HDF5 files are
  involved (#94)
• Fixed behavior of EEG reader to warn when removing "bad" events and optionally
  disable this behavior with a keyword argument (#95)
• Updated conversion from structured arrays to avoid potential issues in numpy
  1.13 (#103)