Skip to content

NEWS.md

Latest commit

 

History

History
175 lines (109 loc) · 8.13 KB

NEWS.md

File metadata and controls

175 lines (109 loc) · 8.13 KB

fplot 1.1.1

New features

  • export_graph now creates the folder receiving the exported file if it does not exist

  • export_graph now accepts the arguments: margin, margin.left, margin.right, margin.top, margin.bottom, margin.unit, box, col.bg, col.default, ylab.horiz, lwd, outermargin, outermargin.left, outermargin.right, outermargin.top, outermargin.bottom, outermargin.unit, square_plot, nrow, ncol, byrow, title.size, title.col, title.bold, title.italic, axis.size, axis.col, axis.bold, axis.italic, label.size, label.col, label.bold, label.italic.

    These arguments modify the graphical parameters before resetting them when export_graph_end is called (or a new instance of export_graph_start).

Bug fixes

  • fix bug in plot_distr when it was applied to continuous data under certain conditions.

  • fix bug partial matching of the label argument in the text function

Other changes

  • remove data.table in depends

  • the dictionary of fplot becomes common with the dictionary of the fixest package

fplot 1.1.0

New functions

  • new pair of functions export_graph_start and export_graph_end to export graphs. This is similar to the family of functions pdf_fit, png_fit, etc, and fit.off(), but possibly more intuitive.

Bug fixes

  • Bug in plot_box when a call of the form plot_box(var ~ 1, base) was used.
  • Bug in plot_box when a call of the form plot_box(. ~ 1, base) was used.

Other

  • update the maintainer's email address
  • remove warning from pdftools::pdf_convert popping when exporting graphs.
  • when exporting graphs, there is a preview of the exported graph. Now the image is embedded in the HTML file via URI to avoid file path issues (this led to a bug in the VSCode viewer).
  • the documentation is updated to add clear guidance for ggplot2 graph exportation
  • improved documentation

fplot 1.0.0 (2020-05-08)

New set of functions to export graphs

  • A new set of functions has been created to easily export figures that should fit in a document.

  • The five functions are pdf_fit, png_fit, jpeg_fit, tiff_fit and bmp_fit.

  • First, if needed, you set the geometry of your document with setFplot_page. Defaults to US-letter and normal margins => this is used to deduce the size the figures will take in the document.

  • When using pdf_fit, instead of providing the height and width of the figure in absolute terms, you provide the fraction of the text-width the figure should take, and the target font-size at which the plotting text should be rendered. The size of the plotting text, once the figure is in the final document, is guaranteed.

  • Use fit.off instead of dev.off to close the device and you will get the exported graph displayed in the viewer pane! No need to open your file to see whether the export is proper, you have it automatically done!

New function: setting default values for plot_distr

  • New function setFplot_distr to set most default values of the function plot_distr.

User visible changes in plot_distr

  • MAJOR CHANGE NOT RETRO COMPATIBLE: complete overhaul of the formula. Now the main argument 'fml' is of the form weights ~ var | moderator (before it was var ~ moderator | weights). Allows to use one-sided formulas.

  • Argument tick_5 becomes at_5 and can take three values: FALSE, "roman" or "line". It is used to inform on the index of the bars by adding a small Roman number under every 5 bars (at_5 = "roman"), or drawing a thick axis line every 5 bars (at_5 = "line"). The default depends on the types of the data.

  • When fractions are displayed on top of the bars, very low numbers are now trucated (avoids having to display too small numbers).

  • The name of the moderator is now displayed in the yaxis when appropriate for within distributions.

  • Legend titles are now always displayed by default.

  • You can display the distribution for several weights at the same time.

  • Argument mod.method has been revamped. It's meaning has been split: mod.method takes values in "side", "split" or "stack" and two new arguments within and total have been created to define whether the distribution should be within each moderator or over the total distribution.

  • The argument trunc.method now takes the values "auto", "right" and "mid" (instead of "auto", "trimRight" and "trimMid").

  • The size of the bottom and left margins are now automatically adjusted so the text always fit.

  • Argument mod.select now supports regular expressions.

  • New argument mod.NA which display the distribution of missing values.

User visible changes in plot_lines

  • You can display the trend for several variables at the same time => they will be treated as moderators.
  • One sided formulas are accepted.
  • Argument mod.select now supports regular expressions.
  • New argument mod.NA which display the distribution of missing values.

User visible changes in plot_box

  • You can display several variables at once (long overdue!).

Error-handling

  • Better error-handling with dreamerr's functions.

Bug correction

  • plot_lines: Bug when plotting frequencies with moderators.
  • plot_box: Bug regarding the display of the outliers when the cases are of type factor.
  • plot_distr: Bug when the variable was of type logical.
  • fplot: Bug when the moderator was loigical.

Arguments: name change

  • max_line => line.max
  • tick_5 => at_5 (and is much better implemented)
  • maxFirst => sorted
  • onTop => top
  • addOther => other
  • maxBins => nbins
  • toLog => log

Changes in version 0.2.1 (2020-02-12)

User visible changes

  • Better adjustment of the legend height.

Bug fixes

  • plot_distr: Bug when the weight contained NAs. Now fixed.

Other

  • Package fixest is removed from help files and called conditionnaly in the vignette.

Changes in version 0.2.0 (2019-11-21)

Major improvements

  • plot_distr: A new method for displaying moderators: mod.method = "stack".
  • plot_distr: A new argument "cumul" to represent cumulative distributions.
  • mod.method = "splitXX" is now supported with logarithmic data
  • mod.method = "splitXX" + toLog=TRUE: now supports the "other" column.
  • labels.tilted=TRUE now possible with logarithmic data.

Other improvements to plot_distr

  • when onTop != "none", very small numbers are not displayed any more (could lead to cluttering).
  • when onTop can also be equal to FALSE.
  • Now discrete ticks appear when plotting categorical data. They appear every 5/10 bins, in order to help counting the bins.
  • better labeling of the y-axis when yaxis.num = TRUE.
  • better automatic finding of the number of bins when argument bin was not missing.
  • better algorithm to find the default bin.size.
  • when bin.size is given, toLog's default is automatically set off.

Improvements to plot_lines

  • Argument mod.select introduced to easily select moderator cases.
  • When smoothing_window > 1, supplemental years are removed.

Data sets

  • Data on US publications in biology is removed.
  • Data on US publications in economics is added (much lighter than biology publications).

Bug correction

  • yaxis.num = TRUE in the presence of moderators could cause problems in special cases.
  • Display problem when mod.method = "splitxx" and factor variables.
  • Display problem when mod.method = "sidexx" and with the other columns (problem in the bar cutting).
  • plot_distr: with moderators and mod.method = "sideWithin": rare bug corrected (identical values could be considered different).

Relabeling of arguments

  • Argument bin becomes bin.size.
  • in argument mod.method, the value within and total become sideWithin and sideTotal.

Changes in version 0.1.0 (2019-11-01)

First version!

  • This package is an effort to provide a simple class of plotting functions which: i) perform common operations in exploratory analyses, ii) have a compact syntax thanks to formulas, allowing aggregate/conditional/weighted graphs with minimum effort, iii) provide automatic options setting to best fit the data, iv) have a nice looking layout, and v) always take advantage of the full plotting region (without changing the graphical parameters) to export beautiful graphs with no/minimum tweak.