Merge branch 'develop' into master

Author: Chester Ismay

.travis.yml

@@ -33,8 +33,11 @@ deploy:
   keep-history: true
   local-dir: docs
   github_token: $GITHUBTRAVIS
+  target-branch: gh-pages-dev
+# target-branch: gh-pages
   on:
-    branch: master
+   branch: develop
+#  branch: master
 
 after_success:
   - Rscript -e 'covr::codecov()'

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: infer
 Type: Package
 Title: Tidy Statistical Inference
-Version: 0.3.1
+Version: 0.3.1.9000
 Authors@R: c(
     person("Andrew", "Bray", email = "[email protected]", role = c("aut", "cre")),
     person("Chester", "Ismay", email = "[email protected]", role = "aut"),
@@ -40,5 +40,6 @@ Suggests:
     covr
 URL: https://github.com/tidymodels/infer
 BugReports: https://github.com/tidymodels/infer/issues
-RoxygenNote: 6.0.1.9000
+Roxygen: list(markdown = TRUE)
+RoxygenNote: 6.1.0
 VignetteBuilder: knitr

NAMESPACE

@@ -17,10 +17,7 @@ export(specify)
 export(t_stat)
 export(t_test)
 export(visualize)
-importFrom(dplyr,as.tbl)
-importFrom(dplyr,as_tibble)
 importFrom(dplyr,bind_rows)
-importFrom(dplyr,data_frame)
 importFrom(dplyr,group_by)
 importFrom(dplyr,inner_join)
 importFrom(dplyr,mutate_if)

NEWS.md

@@ -1,3 +1,10 @@
+# infer 0.3.1.9000
+
+- Account for `NULL` value in left hand side of formula in `specify()` (#156) and `type` in `generate()` (#157).
+- Update documentation code to follow tidyverse style guide (#159).
+- Remove help page for internal `set_params()` (#165).
+- Fully use {tibble} (#166).
+
 # infer 0.3.1
 
 - Stop using package {assertive} in favor of custom type checks (#149)

R/calculate.R

@@ -1,31 +1,32 @@
 #' Calculate summary statistics
-#' @param x the output from \code{\link{generate}} for computation-based
-#' inference or the output from \code{\link{hypothesize}}
-#' piped in to here for theory-based inference.
-#' @param stat a string giving the type of the statistic to calculate. Current
-#' options include "mean", "median", "sd", "prop", "diff in means",
-#' "diff in medians", "diff in props", "Chisq", "F", "t", "z", "slope",
-#' and "correlation".
-#' @param order a string vector of specifying the order in which the levels of
-#' the explanatory variable should be ordered for subtraction, where
-#' \code{order = c("first", "second")} means \code{("first" - "second")}
-#' Needed for inference on difference in means, medians, or proportions and
-#' t and z statistics.
-#' @param ... to pass options like \code{na.rm = TRUE} into functions like
-#'mean, sd, etc.
-#' @return A tibble containing a \code{stat} column of calculated statistics
+#' 
+#' @param x The output from [generate()] for computation-based inference or the
+#'   output from [hypothesize()] piped in to here for theory-based inference.
+#' @param stat A string giving the type of the statistic to calculate. Current
+#'   options include `"mean"`, `"median"`, `"sd"`, `"prop"`, `"diff in means"`,
+#'   `"diff in medians"`, `"diff in props"`, `"Chisq"`, `"F"`, `"t"`, `"z"`,
+#'   `"slope"`, and `"correlation"`.
+#' @param order A string vector of specifying the order in which the levels of
+#'   the explanatory variable should be ordered for subtraction, where `order =
+#'   c("first", "second")` means `("first" - "second")` Needed for inference on
+#'   difference in means, medians, or proportions and t and z statistics.
+#' @param ... To pass options like `na.rm = TRUE` into functions like
+#'   [mean()][base::mean()], [sd()][stats::sd()], etc.
+#'   
+#' @return A tibble containing a `stat` column of calculated statistics.
+#'
+#' @examples
+#' # Permutation test for two binary variables
+#' mtcars %>%
+#'   dplyr::mutate(am = factor(am), vs = factor(vs)) %>%
+#'   specify(am ~ vs, success = "1") %>%
+#'   hypothesize(null = "independence") %>%
+#'   generate(reps = 100, type = "permute") %>%
+#'   calculate(stat = "diff in props", order = c("1", "0"))
+#' 
 #' @importFrom dplyr group_by summarize n
 #' @importFrom rlang !! sym quo enquo eval_tidy
 #' @export
-#' @examples
-#' # Permutation test for two binary variables
-#'   mtcars %>%
-#'     dplyr::mutate(am = factor(am), vs = factor(vs)) %>%
-#'     specify(am ~ vs, success = "1") %>%
-#'     hypothesize(null = "independence") %>%
-#'     generate(reps = 100, type = "permute") %>%
-#'     calculate(stat = "diff in props", order = c("1", "0"))
-
 calculate <- function(x,
                       stat = c(
                         "mean",

R/conf_int.R

@@ -1,28 +1,40 @@
-#' Compute the confidence interval for (currently only) simulation-based methods
+#' Compute confidence interval
 #' 
-#' \code{get_confidence_interval()} and \code{get_ci()} are both aliases of \code{conf_int()}
-#' @param x data frame of calculated statistics or containing attributes
-#' of theoretical distribution values. Currently, dependent on statistics being stored in \code{stat} column as created in \code{calculate()} function.
-#' @param level a numerical value between 0 and 1 giving the confidence level. Default value is 0.95.
-#' @param type a string giving which method should be used for creating the confidence interval. The default is \code{"percentile"} with \code{"se"} corresponding to (multiplier * standard error) as the other option.
-#' @param point_estimate a numeric value or a 1x1 data frame set to NULL by default. Needed to be provided if \code{type = "se"}.
+#' Only simulation-based methods are (currently only) supported.
+#' `get_confidence_interval()` and `get_ci()` are both aliases of `conf_int()`.
+#' 
+#' @param x Data frame of calculated statistics or containing attributes of
+#'   theoretical distribution values. Currently, dependent on statistics being
+#'   stored in `stat` column as created in [calculate()] function.
+#' @param level A numerical value between 0 and 1 giving the confidence level.
+#'   Default value is 0.95.
+#' @param type A string giving which method should be used for creating the
+#'   confidence interval. The default is `"percentile"` with `"se"`
+#'   corresponding to (multiplier * standard error) as the other option.
+#' @param point_estimate A numeric value or a 1x1 data frame set to `NULL` by
+#'   default. Needed to be provided if `type = "se"`.
 #'
-#' @return a 2 x 1 tibble with values corresponding to lower and upper values in the confidence interval
-#' @export
-#' @rdname get_ci
+#' @return A 1 x 2 tibble with values corresponding to lower and upper values in
+#'   the confidence interval.
+#' 
 #' @examples
 #' mtcars_df <- mtcars %>%
-#'     dplyr::mutate(am = factor(am))
+#'   dplyr::mutate(am = factor(am))
 #' d_hat <- mtcars_df %>%
-#'     specify(mpg ~ am) %>%
-#'     calculate(stat = "diff in means", order = c("1", "0"))
+#'   specify(mpg ~ am) %>%
+#'   calculate(stat = "diff in means", order = c("1", "0"))
 #' bootstrap_distn <- mtcars_df %>%
-#'     specify(mpg ~ am) %>%
-#'     generate(reps = 100) %>%
-#'    calculate(stat = "diff in means", order = c("1", "0"))
+#'   specify(mpg ~ am) %>%
+#'   generate(reps = 100) %>%
+#'   calculate(stat = "diff in means", order = c("1", "0"))
 #' bootstrap_distn %>% conf_int(level = 0.9)
 #' bootstrap_distn %>% conf_int(type = "se", point_estimate = d_hat)
+#' 
+#' @name get_ci
+NULL
 
+#' @rdname get_ci
+#' @export
 conf_int <- function(x, level = 0.95, type = "percentile", 
                      point_estimate = NULL){
 
@@ -72,12 +84,10 @@ check_ci_args <- function(x, level, type, point_estimate){
 }
 
 
-#' @export
 #' @rdname get_ci
-
+#' @export
 get_ci <- conf_int
 
-#' @export
 #' @rdname get_ci
-
+#' @export
 get_confidence_interval <- conf_int

R/generate.R

@@ -1,27 +1,35 @@
-#' Generate resamples, permutations, or simulations based on
-#' `specify` and (if needed) `hypothesize` inputs
-#' @param x a data frame that can be coerced into a \code{\link[dplyr]{tbl_df}}
-#' @param reps the number of resamples to generate
-#' @param type currently either \code{bootstrap}, \code{permute}, 
-#' or \code{simulate}
-#' @param ... currently ignored
-#' @return A tibble containing \code{rep} generated datasets, indicated by the
-#' \code{replicate} column.
-#' @importFrom dplyr group_by
-#' @export
+#' Generate resamples, permutations, or simulations
+#' 
+#' Generation is done based on [specify()] and (if needed) [hypothesize()]
+#' inputs.
+#' 
+#' @param x A data frame that can be coerced into a [tibble][tibble::tibble].
+#' @param reps The number of resamples to generate.
+#' @param type Currently either `bootstrap`, `permute`, or `simulate`.
+#' @param ... Currently ignored.
+#' 
+#' @return A tibble containing `rep` generated datasets, indicated by the
+#'   `replicate` column.
+#' 
 #' @examples
 #' # Permutation test for two binary variables
-#'   mtcars %>%
-#'     dplyr::mutate(am = factor(am), vs = factor(vs)) %>%
-#'     specify(am ~ vs, success = "1") %>%
-#'     hypothesize(null = "independence") %>%
-#'     generate(reps = 100, type = "permute")
-
+#' mtcars %>%
+#'   dplyr::mutate(am = factor(am), vs = factor(vs)) %>%
+#'   specify(am ~ vs, success = "1") %>%
+#'   hypothesize(null = "independence") %>%
+#'   generate(reps = 100, type = "permute")
+#' 
+#' @importFrom dplyr group_by
+#' @export
 generate <- function(x, reps = 1, type = attr(x, "type"), ...) {
 
   auto_type <- attr(x, "type")
 
   if(!is.null(auto_type)){
+    if (is.null(type)) {
+      stop_glue("Supply not `NULL` value of `type`.")
+    }
+    
     if(auto_type != type)
       stop_glue(
         "You have specified `type = \"{type}\"`, but `type` is expected to be ",
@@ -148,7 +156,6 @@ permute_once <- function(x, ...) {
 #' @importFrom dplyr pull
 #' @importFrom tibble tibble
 #' @importFrom rlang :=
-
 simulate <- function(x, reps = 1, ...) {
   fct_levels <- as.character(unique(dplyr::pull(x, !! attr(x, "response"))))
 
@@ -158,8 +165,10 @@ simulate <- function(x, reps = 1, ...) {
                                               prob = format_params(x)),
                                  simplify = FALSE))
 
-  rep_tbl <- tibble(!! attr(x, "response") := as.factor(col_simmed),
-                   replicate = as.factor(rep(1:reps, rep(nrow(x), reps))))
+  rep_tbl <- tibble::tibble(
+    !! attr(x, "response") := as.factor(col_simmed),
+    replicate = as.factor(rep(1:reps, rep(nrow(x), reps)))
+  )
 
   rep_tbl <- set_attributes(to = rep_tbl, from = x)
 

R/hypothesize.R

@@ -1,21 +1,23 @@
 #' Declare a null hypothesis
-#' @param x a data frame that can be coerced into a \code{\link[dplyr]{tbl_df}}
-#' @param null the null hypothesis. Options include "independence" and "point"
-#' @param ... arguments passed to downstream functions
-#' @return A tibble containing the response (and explanatory, if specified) 
-#' variable data with parameter information stored as well
-#' @importFrom dplyr as.tbl
-#' @return a data frame with attributes set
-#' @export
+#' 
+#' @param x A data frame that can be coerced into a [tibble][tibble::tibble].
+#' @param null The null hypothesis. Options include `"independence"` and
+#'   `"point"`.
+#' @param ... Arguments passed to downstream functions.
+#' 
+#' @return A tibble containing the response (and explanatory, if specified)
+#'   variable data with parameter information stored as well.
+#' 
 #' @examples
 #' # Permutation test similar to ANOVA
-#'   mtcars %>%
-#'     dplyr::mutate(cyl = factor(cyl)) %>%
-#'     specify(mpg ~ cyl) %>%
-#'     hypothesize(null = "independence") %>%
-#'     generate(reps = 100, type = "permute") %>%
-#'     calculate(stat = "F")
-
+#' mtcars %>%
+#'   dplyr::mutate(cyl = factor(cyl)) %>%
+#'   specify(mpg ~ cyl) %>%
+#'   hypothesize(null = "independence") %>%
+#'   generate(reps = 100, type = "permute") %>%
+#'   calculate(stat = "F")
+#' 
+#' @export
 hypothesize <- function(x, null, ...) {
 
   hypothesize_checks(x, null)
@@ -69,6 +71,6 @@ hypothesize <- function(x, null, ...) {
   #               '`mu`, `med`, or `sd` to be used as a parameter.')
   # }
 
-  return(as.tbl(x))
+  return(tibble::as_tibble(x))
 }
 

R/infer.R

@@ -1,14 +1,15 @@
 #' infer: a grammar for statistical inference
 #' 
-#' The objective of this package is to perform statistical inference 
-#' using a grammar that illustrates the underlying
-#' concepts and a format that coheres with the tidyverse.
+#' The objective of this package is to perform statistical inference using a
+#' grammar that illustrates the underlying concepts and a format that coheres
+#' with the tidyverse.
 #'
-#' @docType package
-#' @name infer
 #' @examples
 #' # Example usage:
 #' library(infer)
+#' 
+#' @docType package
+#' @name infer
 NULL
 
 ## quiets concerns of R CMD check re: the .'s that appear in pipelines

R/p_value.R

@@ -1,28 +1,36 @@
-#' Compute the p-value for (currently only) simulation-based methods
-#' \code{get_pvalue()} is an alias of \code{p_value}
-#' @param x data frame of calculated statistics or containing attributes
-#' of theoretical distribution values
-#' @param obs_stat a numeric value or a 1x1 data frame (as extreme or more extreme than this)
-#' @param direction a character string. Options are "less", "greater", or "two_sided".
-#' Can also specify "left", "right", or "both".
+#' Compute p-value
+#' 
+#' Only simulation-based methods are (currently only) supported. `get_pvalue()`
+#' is an alias of `p_value`.
+#' 
+#' @param x Data frame of calculated statistics or containing attributes of
+#'   theoretical distribution values.
+#' @param obs_stat A numeric value or a 1x1 data frame (as extreme or more
+#'   extreme than this).
+#' @param direction A character string. Options are `"less"`, `"greater"`, or
+#'   `"two_sided"`. Can also specify `"left"`, `"right"`, or `"both"`.
 #'
-#' @return a 1x1 data frame with value between 0 and 1
-#' @export
-#' @rdname get_pvalue
+#' @return A 1x1 data frame with value between 0 and 1.
+#' 
 #' @examples
 #' mtcars_df <- mtcars %>%
-#'     dplyr::mutate(am = factor(am))
+#'   dplyr::mutate(am = factor(am))
 #' d_hat <- mtcars_df %>%
-#'     specify(mpg ~ am) %>%
-#'     calculate(stat = "diff in means", order = c("1", "0"))
+#'   specify(mpg ~ am) %>%
+#'   calculate(stat = "diff in means", order = c("1", "0"))
 #' null_distn <- mtcars_df %>%
-#'     specify(mpg ~ am) %>% 
-#'     hypothesize(null = "independence") %>%
-#'     generate(reps = 100) %>%
-#'     calculate(stat = "diff in means", order = c("1", "0"))
+#'   specify(mpg ~ am) %>% 
+#'   hypothesize(null = "independence") %>%
+#'   generate(reps = 100) %>%
+#'   calculate(stat = "diff in means", order = c("1", "0"))
 #' null_distn %>% 
-#'     p_value(obs_stat = d_hat, direction = "right")
+#'   p_value(obs_stat = d_hat, direction = "right")
+#' 
+#' @name get_pvalue
+NULL
 
+#' @rdname get_pvalue
+#' @export
 p_value <- function(x, obs_stat, direction){
 
   check_type(x, is.data.frame)
@@ -91,8 +99,8 @@ two_sided_p_value <- function(x, obs_stat){
     return(tibble::tibble(p_value = basic_p_value))
 }
 
-#' @export
 #' @rdname get_pvalue
+#' @export
 get_pvalue <- p_value
 
 # which_distribution <- function(x, theory_type, obs_stat, direction){