Add framework for animation options (fixed)

And no need to filter out NULL elements

Leads to better documentation where it is easier to understand how to use the anim_opts parameter

- Collected in animate_options.R
- Sets global defaults in plot_settings environment
- Adds ability to set animation options, similar to theme_set() from ggplot2
- Adds getters for current (if set) or default animation options

Merge branch 'pkg-add-anim-options-merge' into pkg-add-anim-options

# Conflicts:
#	NAMESPACE
#	R/animate_tidyr.R
#	R/plot_helpers.R
#	man/anim_options.Rd
#	man/animate_gather.Rd
#	man/animate_spread.Rd
#	man/static_plot.Rd

ggproto objects like those created by enter_fade() or exit_appear() evaluate to enviroments making it impossible to compare two ggproto objects to determine if they are the same (both will be different environments). To get around this, I store the expression used to create the enter/exit arguments so that we can later compare two objects created by anim_options().

Set option "tidyexplain.strict_dots" to TRUE for checking for issues with extra arguments being passed in ... to the _plot() functions.

And update documentation for both options

And demonstrates usage with the tidyr gather/spread animators. In essence, user-facing functions take an argument `anim_opts = anim_options()` as a last argument (after ... if applicable). These options should be passed through to lower level functions and should be given to `animate_plot()` and `static_plot()`.

This implementation actually doesn't break previous functions, but the join and set families will need to be updated to match.

Calls to tidyr and dplyr are now prefixed with package name. This lets us use the README as a test to make sure all namespace issues are fixed. We can roll this back when the package is more stable.

- Use quasiquotation to get variable names
- Unified documentation
- Use match.arg to guard for correct arguments in type and export
- Removed animate_helpers.R

Adds get_input_text() that can be used to get the literal variable name and make_named_data() that creates a list with the input data frames. This second one is necessary because the inputs may be quosures and we need to get the underlying data frames. Default names are $x and $y.

And change Long/Wide to lowercase to reflect that they are both variables names and a description of the state of the data.

animate_join() now lives in animate_join.R with the other animate_join_ functions. These all use rlang::enquo() to capture their arguments so that animate_join() can print the variable name in the plot title.

Changed name of set_text_color() to choose_text_color() to avoid confusion with the global option setter functions, e.g. set_text_size(). Updated tests to reflect modified threshold.

Sets figure output to man/figures/tidyexplain- and ensures that all R chunks that output figures are named. This gives a collection of animations and static images with reasonable names ready for download and use.

Currently only used for setting global text and title font sizes, but can be extended to other options that we may wish to set for more than one plot. For fonts, we now provide `set_font_size()` that can be used to set text and title sizes across all subsequent animations. Includes internal getter functions get_{text|title}_size().

Adds a few additional dplyr functions to the package imports, in particular: slice, data_frame, and row_number.

Uses the algorithm mentioned at https://stackoverflow.com/a/3943023/2022615 based on the luminance of the background color to determine black or white text. The threshold favored black text in a way that didn't look good with our plots, so I increased the threshold from 0.179 to 0.333.

if printed it will still call animate, but gives the user more flexibility about animation factors (such as times etc)

- Import packages from tidyverse
- Set global imports in zzzz-package.R
- Import pipe and tidyeval
- Update authors
- Downgrade version to 0.0.1.9000
- Change license to MIT
- TODO: Internal package qualifications still needed, e.g. dplyr::row_numbers()

-id1, -id2 work, id1:id2 does not (yet)

this makes it easier to differentiate between the join-related and the (coming) tidyr-related functions that have similar goals