Added
- each Group of XAFS data will have a Journal - a list of entries with (label,
value, timestamp) that will be used to record processing steps. The XAFS
processing functions will write the parameters they were called with to this
Journal. Users can add notes as Journal entries too.
- XAS Viewer allows editing of program-wide Preferences, that set the default
values and settings for the various analysis steps and program behavior.
- In many windows showing analysis results, XAS Viewer supports setting the
plot window number (1 through 9) so that more than one or two plot windows
can be shown at a time.
- Session-wide Project files (.larix extension) can now be saved and reloaded.
These files contain essentially all the arrays for each Group of data,
including Journals, processing results, and fit histories. The "processing
workspaces" for pre-edge peak fitting, PCA, Linear Combinations,
Regressions, and Feff fitting. These Session files can be saved/loaded by
plain Python commands. When loading these into XAS Viewer, the fit histories
and workspaces will be available. To allow better debugging and tracking of
provenance, these Session Files include "configuration" data about the
Session (python, larch version, etc) and a complete set of Larch commands
for the session, though session commands are not "restored". The files are
compressed and use a customized JSON encoding.
- These Session files are now auto-saved by XAS Viewer periodically (every 15
minutes of activity, by default, and rotated so that a small number (3, by
default) of the most recent auto-saved files are kept. On startup, "restore
the last session" is available (Ctrl-I).
Fixed
- XAS Viewer PCA fixed.
- XAS Viewer now displays how each Group was created, for example
"-log(i1/i0)".
- XAS Viewer panels have more uniform energy range settings, relative to E0,
per group.
- XAS Viewer has more consistent coloring and font sizes.
- XAS Viewer Quit dialog now lists the last saved Session file.
- XAS Viewer now works even harder to keep _xasgroups correct.
- pycifrw added as a dependency from conda-forge
- copying parameters from one group to another is much improved.
- rebinning of EXAFS data better avoids duplicate energies.
- a few bugs with the handling of some CIFs were improved
- Feff calculations now default to L3 edge for Z>= 60.
- Python 3.10 is better supported, Python 3.7 seems to still work.