- Minor bug fix in RmarineHeatWaves.R, i.e. sort function in 'clim_only' on l.336.
- Add start-up message.
- Minor bug fixes.
- Reached end of active development cycle.
- Remove restriction to require full years for start/end points of climatology calculations in detect().
- Documentation updated accordingly.
- Fix error with smooth_percentile and smooth_percentile_width descriptions that were interchanged.
- Bug fix to the block_event() function.
- Bug fix to the event_line() function.
- Repaired event_line(), which broke in v.0.16.0 due to an edit there to enable using arbitrary column headers for the date and temperature columns (previously the defaults were 't' and 'temp').
- detect() now reurns a list with two tibbles, rather than data.frames as before.
- The make_whole(), detect() and exceedance() functions now handle any arbitrary column names, which are specified via arguments to the function.
- Remove dependence on magrittr.
- More consistent use of tidyverse packages throughout.
- Updates to examples.
- Miscellaneous other improvements troughout the function internals.
- Warnings given by exceedance() when exceedances begin/end on the first/last day of the time series have been corrected --- it will now report an NA when the exceedance occurs at the start/end of the time series.
- Some plyr routines phased out in exceedance function in favour of a dplyr approach.
- Further removed need for plyr functions --- use dplyr instead.
- Warnings given by detect() when events begin/end on the first/last day of the time series have been corrected --- it will now report an NA when the event occurs at the start/end of the time series. This bug was highlighted by Juan Bazo.
- Documentation in detect() about the use of percentiles with cold-spells has been made more clear.
- Dependency on some plyr functions have been started to be phased out; this is done to eventually ensure full compliance with the tidyverse suite of functions.
- Remove some unwanted attributes that resulted from dplyr::group_by on lines 496 and 497. The list containing 'event' and 'clim' is now uncluttered by unwanted attributes.
- Update README.Rmd to correct figure paths.
- Update to 'protoFunc' helper function within 'detect'; in the previous version it returned the error "Error: not compatible with STRSXP" under rare circumstances, the cause of which remains unknown.
- Update the calculation of proto_gaps for the same reason.
- The detect() function now also produces a 'climatological' variance.
- Improve documentation in a help file.
- The detect() function may now be told to calculate climatologies only.
- Updating README.md, README.Rmd files.
- Updating the vignette.
- 'exceedence' changed to 'exceedance' throughout.
- Final edits to 'geom_flame' and 'geom_lolli';
- specifically in geom_flame, small gaps in the corners of polygons are now being filled in correctly.
- Updated examples.
- Edits to 'geom_flame' and 'geom_lolli'.
- Edits to 'geom_flame' and 'geom_lolli'.
- Changed 'geom_event_line' and 'geom_lolli_plot' to 'geom_flame' and 'geom_lolli'
- These new functions are now proper ggplot geoms.
- Corrected plotting error in event polygons in 'event_line' function
- Added two new functions: 'geom_event_line' and 'geom_lolli_plot' which may be used with ggplot2 directly as geoms.
- Fix bug which was that was brought to my attention by Mahmoud Haouari: enable the pctile option in the detect() function, which was accidentally disabled.
- Improved error messages in exceedance().
- Bug fix.
- All changes since version 0.13.1.
- Minor edits to documentation.
- Corrected issue where the event_line graph's y-axis did not reflect the correct units for the metric selected for plotting.
- Updated the theme, i.e. removed the default ggplot2 theme in favour of a more conventional white background with nice black axes.
- Removed the legend that indicated the peak and secondary events from the event_line plot.
- Repositioned the legend in the event_line plot: it is now int he bottom right corner as here it has less chance of plotting over the peak events when the graph is rescaled.
- Minor update to exceedance() function to improve usability.
- Usage of exceedance() shown in README file
- The additions to the detect() function madde in the previous update have been rolled back in favour of creating an independent function to calculate exceedances.
- Added 'threshold' variable to detect() function.
- This now allows the function to detect when temperatures are over (under) a static threshold supplied by the user.
- An example is given in the README.md file
- Expanded example included as package vignette.
- Added a marine cold spell example.
- Correcting a few typos.
- The new development version.
- Edit to README.md file.
- Significant changes to detection algorithm:
- ... more extensive use of magrittr pipe operators %>% and %<>%
- ... simplification of protoFunc helper function
- ... removal of automatic start/end day calculation if these are not provided
- ... climatology start/end dates must now always be explicitely set
- ... finally enabled join_across_gaps option
- ... fixed a bug that caused the function to fail when none of the gaps between events (i.e. proto_gaps) were less than max_gap
- Update examples to accommodate these changes.
- Add README.md file.
- Various minor changes to documentation of several functions.
- Fix Windows build dependency warnings.
- Added lolliplot functionality -- lolli_plot().
- Simplify make_whole() -- it should accept dates as class POSIXct or Date without the need for unneccesary 'if' logic.
- All example data dates (t) changed to class Date.
- Some minor rewording to documentation.
- Minor edits to event_line() as per Robert Schlegel (allows broader selection of metrics for plotting.)
- block_average() rewritten -- it is now based upon dplyr functions so it is faster and more stream-lined.
- Completely removed the use of reshape2 in favour of tidyr.
- All comments removed inside of functions.
- Minor changes to the version change info provided for v0.10.1.
- Fixed a bug that caused detect() to fail whenever it encountered fewer than two non-NAs in the period doy 59 to doy 61 when it was asked to interpolate over the non-existent day-60 during non-leap years. This is specific to versions of 'zoo' (required for na.approx()) up to 1.7-12; from 1.7-13 it works fine. A few extra lines of code were added to fix this.
- Added the block_average() function.
- Expanded documentation.
- Changes to event_line() documentation.
- Renamed 'metric' options to function as 'mean', 'maximum', 'cumulative'.
- Added more TODOs at the bottom of this file.
- Minor refinements to the documentation (i.e. package description).
- Replace 'mhw' with 'event'.
- Add basic plotting functionality a-la Robert Schlegel in the form of the event_line() function (plus edits to make Rob's code produce a clean build process, thereby avoiding throwing 'notes' that might be frowned upon by the CRAN people.)
- Renamed some things: 'make_whole.R' -> 'makeWhole.R'; 'marineHeatWaves-package.r' -> 'RmarineHeatWaves-package.r'; 'marineHeatWaves.R' -> 'RmarineHeatWaves.R'.
- Include CITATION file.
- Update DESCRIPTION file.
- Update author, creator and contributor roles.
- Add dates to NEWS.md file.
- Marine cold spell option enabled.
- Hacky changes to suppress some notes during package compilation, needed for acceptance to CRAN, apparently.
- Internal variable name changes (replace more camelCase).
- Less problematic handling of NA.
- Improved/expanded documentation.
- Internal variable name changes (replace camelCase).
- Split out make_whole() function.
- Add make_whole() example.
- Added data and a functional example.
- Slightly improved documentation.
- Robert Schlegel added as co-author for work on graphing functions.
- Changed 'eventNo' calculation to properly reflect the actual events, and not one of the proto-event types.
- Changed ID variable in 'mhw' output to 'eventNo', thereby following the same naming convention.
- Removed ', .id = NULL' from the event metrics 'ldply' calculations, enabling compatibility with plyr (version < 1.8.3).
- Add 'eventNo' to climatology output to identify each unique event and to fascilitate plotting of filled polygons using 'geom_polygon' in ggplot2, as per Robert Schlegel's suggestion.