Changed
- Filters now use pandas to keep track of filter components and switches.
- Quality of life improvements using pandas DataFrames to display filter status
- Improvements to the calculation of rolling statistics. Can choose justification window relative to data position, and rolling gradient calculation now handles nans better (excludes them).
- Improvements to `gradient_threshold` filters. Can now set `absolute_gradient=False` to select 'up' or 'down' components separately. `gplot` now shows the effect of filters directly, similar to `tplot`.
0.3.18-release