Skip to content

TODO.md

Latest commit

 

History

History
175 lines (121 loc) · 7.21 KB

TODO.md

File metadata and controls

175 lines (121 loc) · 7.21 KB

General issues

  • should not depend on rapport, try to only import (against that: depending on rapport's options?)

Documentation

README.md

  • add installation details + Pandoc dependency
  • add example files and html/pdf/docx exports
  • add absolute path to examples (as view from two locations)
  • remove rapport dependency
  • add ESS functions
  • add: short documentation (or at least mention!): evals
  • comparison with knitr
  • add: options + p options + evals.messages option

Github pages

  • add absolute path to examples (as view from two locations)

Examples

  • do not forget about switching I.have.time to TRUE before exporting to Github pages

Vignette

  • TODO won't fix: we have a homepage

Coding

General tasks

  • add controllable, global options pander.option
  • rename pander.options & eval.options not to conflict with S3 methods** panderOptions and evalsOptions
  • update package functions to use new naming scheme for options
  • add caption/other tweaks (alignment, sign. stars, emphasis etc.) to images/tables using attrs

Evals

  • cleanup code from repeating code (like: suppressMessages(suppressWarnings(...))))

  • remove evaluate dependency

  • grab multiple returned object/images in one chunk

  • add option: grab.images (render those to disk or return unprinted) use eval.msgs instead

  • caching

  • add: global options

  • add: option to copy images

  • check if img file referenced in cached result really exists

  • add: evals.messages option passed to pander.evals to possibly suppress messages

  • add: optionally save img's R object (lattice & ggplot2) to RDA file

  • What if img file referenced in cached result was altered? Should not we try to check for existing recordedPlot and just rerender the img (without actual eval) to overcome this issue?: resolved with cache.copy.images and cache.mode == "disk"

  • run twice with evalsOptions('cache.time', 0): evals('x<-1:21;histogram(x)'): resolved by updating objects (while returning from cache) changed by cached code~~

  • add option to run code in sandbox:

Image options:

Global options for: lattice, ggplot2 and base plots

  • plot margins

  • theme: font (family, base size, color)

  • theme: plotting area background colors

    • known issues: not in base plots (just global backgroung)

  • theme: foreground colors (discrete, continuous)

    • known issues: base plot solution is really ugly (overwriting calls' col parameter)

  • grid options:

    • enabled/disabled
    • split enable/disable: major/minor
    • grid color
    • grid line type

  • axes:

    • color
    • angle
    • split long labels

  • title: split if too long (strwidth, strheight)

  • others:

    • border/box around plot/strip
    • transparent border of histogram
    • symbols

  • remove dependencies: ggplot2, latticeExtra and build only on grid

Add demo to GH page!: inst/examples/graph.brew

Helper functions

  • "p" from rapport: modify defaults, fork it
  • check strings before applying formatting (to prevent e.g. "foo")
  • table: implement multi-line syntax for tables (as Pandoc does not support alignment with the current grid approach)
  • table: support cells with line breaks
  • table: split too wide tables into multiple tables (global solution which can be handled in LaTeX, docx etc. separately)
  • table: add strong emphasis to row names
  • table: fails with one-column tables. E.g.: pander(mtcars[1,])
  • table: issue with rounding numbers, just check: pander(mtcars$mpg) vs. pander(mtcars$mpg) -> this is based on low digits option
  • table: add more styling options (e.g. (strong) emphasizing custom cells, alignment)
  • table: add option to draw significance starts in cells
  • table: @idea wrap header of tables to minimal (even in words!) to keep the table's width minimal

Required helper functions

  • add significance stars
  • indent concatenated strings: pandoc.indent

Calling Pandoc (converting docs)

  • update footer (currently shows: rapport)

  • revert colorbox to slimbox2 with fullscreen images

  • add templates for different formats

  • add option to change rendering back-end's name (like: pander -> rapport with version number)

  • Q: include Pandoc somehow in the package not to ask users to install it?

    A: INSTALL file updated + if Pandoc is not found it's shown to the user. Might try to auto-download binary for Mac/Windows? Linux users would deal with that problem, right? :) UPDATE: see installr package

Pandoc/pander methods

  • logical
  • default
  • list
  • density
  • CrossTable
  • What to do with summary classes? Leaving out, implement in Pandoc methods or what?

Leaving out! As most Pandoc methods return a table and also trying to add a chatty caption, in most cases the full object information is needed which can (and mostly would) be truncated inside the Pandoc method with summary.

Brew

  • check out image directory (should be getwd()/images)
  • remove image absolute path
  • prettify image names to output + index no.
  • exporting features?
  • option to convert document to multiple formats at once (although with caching it's no so bad even now)
  • remove parse (to deal with syntax errors)

R5

  • migrate to Pandoc.convert
  • remove multiple line breaks, see: remove.extra.newlines()
  • open exported docs
  • tweak evals like in Pandoc.brew

Add support for knitr

  • implement it: Pandoc.knitr
  • create knitr hooks which would apply pander to each R object

Tests

  • Do that at last!
  • evals
  • brew
  • helper functions

HTML/JS/CSS issues

  • move errors (footnote) above of upper footer and add a header tag (if found)

ESS

Great ideas

  • Michael Lawrence @ [ESS] mail list: Another thing I've been wanting is a way to output objects in multiple ways, separate from the report itself. Typical side effects would be: generation of an R data package with the result as a dataset, or storage of the result in some other database. It would be cool to be able to specify, on a per-block basis, the output driver (or a list of them). For example, your pander() function could be a dual-dispatch S4 generic, dispatching on both the object to export, and an object representing the target. -> hook could depend on outer parameters in evals?