Merge pull request #1 from tidymodels/master

Merge from master

Author: juliasilge

DESCRIPTION

@@ -1,5 +1,5 @@
 Package: parsnip
-Version: 0.0.5.9000
+Version: 0.0.5.9001
 Title: A Common API to Modeling and Analysis Functions
 Description: A common interface is provided to allow users to specify a model without having to remember the different argument names across different functions or computational engines (e.g. 'R', 'Spark', 'Stan', etc). 
 Authors@R: c(
@@ -31,7 +31,7 @@ Imports:
     prettyunits,
     vctrs (>= 0.2.0)
 Roxygen: list(markdown = TRUE)
-RoxygenNote: 7.0.2.9000
+RoxygenNote: 7.1.0
 Suggests: 
     testthat,
     knitr,

NAMESPACE

@@ -25,12 +25,18 @@ S3method(predict,model_spec)
 S3method(predict,nullmodel)
 S3method(predict_class,"_lognet")
 S3method(predict_class,"_multnet")
+S3method(predict_class,model_fit)
 S3method(predict_classprob,"_lognet")
 S3method(predict_classprob,"_multnet")
+S3method(predict_classprob,model_fit)
+S3method(predict_confint,model_fit)
 S3method(predict_numeric,"_elnet")
+S3method(predict_numeric,model_fit)
+S3method(predict_quantile,model_fit)
 S3method(predict_raw,"_elnet")
 S3method(predict_raw,"_lognet")
 S3method(predict_raw,"_multnet")
+S3method(predict_raw,model_fit)
 S3method(print,boost_tree)
 S3method(print,control_parsnip)
 S3method(print,decision_tree)
@@ -92,7 +98,10 @@ export(boost_tree)
 export(check_empty_ellipse)
 export(check_final_param)
 export(control_parsnip)
+export(convert_args)
+export(convert_stan_interval)
 export(decision_tree)
+export(eval_args)
 export(fit)
 export(fit.model_spec)
 export(fit_control)
@@ -122,6 +131,13 @@ export(null_value)
 export(nullmodel)
 export(pred_value_template)
 export(predict.model_fit)
+export(predict_class.model_fit)
+export(predict_classprob.model_fit)
+export(predict_confint.model_fit)
+export(predict_numeric)
+export(predict_numeric.model_fit)
+export(predict_quantile.model_fit)
+export(predict_raw.model_fit)
 export(rand_forest)
 export(rpart_train)
 export(set_args)
@@ -144,6 +160,7 @@ export(svm_poly)
 export(svm_rbf)
 export(tidy.model_fit)
 export(translate)
+export(translate.default)
 export(update_dot_check)
 export(update_main_parameters)
 export(varying)

R/aaa.R

@@ -13,7 +13,13 @@ maybe_multivariate <- function(results, object) {
   results
 }
 
+#' Convenience function for intervals
 #' @importFrom stats quantile
+#' @export
+#' @keywords internal
+#' @param x A fitted model object
+#' @param level Level of uncertainty for intervals
+#' @param lower Is `level` the lower level?
 convert_stan_interval <- function(x, level = 0.95, lower = TRUE) {
   alpha <- (1 - level) / 2
   if (!lower) {
@@ -24,6 +30,33 @@ convert_stan_interval <- function(x, level = 0.95, lower = TRUE) {
   res
 }
 
+#' Make a table of arguments
+#' @param model_name A character string for the model
+#' @keywords internal
+#' @export
+convert_args <- function(model_name) {
+  envir <- get_model_env()
+
+  args <-
+    ls(envir) %>%
+    tibble::tibble(name = .) %>%
+    dplyr::filter(grepl("args", name)) %>%
+    dplyr::mutate(model = sub("_args", "", name),
+                  args  = purrr::map(name, ~envir[[.x]])) %>%
+    tidyr::unnest(args) %>%
+    dplyr::select(model:original)
+
+  convert_df <- args %>%
+    dplyr::filter(grepl(model_name, model)) %>%
+    dplyr::select(-model) %>%
+    tidyr::pivot_wider(names_from = engine, values_from = original)
+
+  convert_df %>%
+    knitr::kable(col.names = paste0("**", colnames(convert_df), "**"))
+
+}
+
+
 # ------------------------------------------------------------------------------
 # nocov
 
@@ -32,8 +65,8 @@ utils::globalVariables(
   c('.', '.label', '.pred', '.row', 'data', 'engine', 'engine2', 'group',
     'lab', 'original', 'predicted_label', 'prediction', 'value', 'type',
     "neighbors", ".submodels", "has_submodel", "max_neighbor", "max_penalty",
-    "max_terms", "max_tree", "name", "num_terms", "penalty", "trees",
+    "max_terms", "max_tree", "model", "name", "num_terms", "penalty", "trees",
     "sub_neighbors", ".pred_class")
-  )
+)
 
 # nocov end

R/aaa_multi_predict.R

@@ -117,18 +117,18 @@ multi_predict_args.default <- function(object, ...) {
 #' @export
 #' @rdname has_multi_predict
 multi_predict_args.model_fit <- function(object, ...) {
-  existing_mthds <- methods("multi_predict")
-  cls <- class(object)
-  tst <- paste0("multi_predict.", cls)
-  .fn <- tst[tst %in% existing_mthds]
-  if (length(.fn) == 0) {
-    return(NA_character_)
+  model_type <- class(object$spec)[1]
+  arg_info <- get_from_env(paste0(model_type, "_args"))
+  arg_info <- arg_info[arg_info$engine == object$spec$engine,]
+  arg_info <- arg_info[arg_info$has_submodel,]
+
+  if (nrow(arg_info) == 0) {
+    res <- NA_character_
+  } else {
+    res <- arg_info[["parsnip"]]
   }
 
-  .fn <- getFromNamespace(.fn, ns = "parsnip")
-  omit <- c('object', 'new_data', 'type', '...')
-  args <- names(formals(.fn))
-  args[!(args %in% omit)]
+  res
 }
 
 #' @export

R/arguments.R

@@ -103,6 +103,11 @@ maybe_eval <- function(x) {
   y
 }
 
+#' Evaluate parsnip model arguments
+#' @export
+#' @keywords internal
+#' @param spec A model specification
+#' @param ... Not used.
 eval_args <- function(spec, ...) {
   spec$args   <- purrr::map(spec$args,   maybe_eval)
   spec$eng_args <- purrr::map(spec$eng_args, maybe_eval)

R/boost_tree.R

@@ -61,6 +61,13 @@
 #'
 #' @section Engine Details:
 #'
+#' The standardized parameter names in parsnip can be mapped to their original
+#' names in each engine:
+#'
+#' ```{r echo = FALSE}
+#' convert_args("boost_tree")
+#' ```
+#'
 #' Engines may have pre-set default arguments when executing the
 #'  model fit call.  For this type of model, the template of the
 #'  fit calls are:

R/decision_tree.R

@@ -46,6 +46,13 @@
 #'
 #' @section Engine Details:
 #'
+#' The standardized parameter names in parsnip can be mapped to their original
+#' names in each engine:
+#'
+#' ```{r echo = FALSE}
+#' convert_args("decision_tree")
+#' ```
+#'
 #' Engines may have pre-set default arguments when executing the
 #'  model fit call. For this type of
 #'  model, the template of the fit calls are::

R/linear_reg.R

@@ -44,6 +44,13 @@
 #'
 #' @section Engine Details:
 #'
+#' The standardized parameter names in parsnip can be mapped to their original
+#' names in each engine:
+#'
+#' ```{r echo = FALSE}
+#' convert_args("linear_reg")
+#' ```
+#'
 #' Engines may have pre-set default arguments when executing the
 #'  model fit call. For this type of
 #'  model, the template of the fit calls are:
@@ -60,6 +67,8 @@
 #'
 #' \Sexpr[results=rd]{parsnip:::show_fit(parsnip:::linear_reg(), "stan")}
 #'
+#' (note that the `refresh` default prevents logging of the estimation process. Change this value in `set_engine()` will show the logs)
+#'
 #' \pkg{spark}
 #'
 #' \Sexpr[results=rd]{parsnip:::show_fit(parsnip:::linear_reg(), "spark")}
@@ -98,7 +107,7 @@
 #'  separately saved to disk. In a new session, the object can be
 #'  reloaded and reattached to the `parsnip` object.
 #'
-#' @seealso [[fit()], [set_engine()]
+#' @seealso [fit()], [set_engine()]
 #' @examples
 #' linear_reg()
 #' # Parameters can be represented by a placeholder:

R/linear_reg_data.R

@@ -179,7 +179,7 @@ set_fit(
     interface = "formula",
     protect = c("formula", "data", "weights"),
     func = c(pkg = "rstanarm", fun = "stan_glm"),
-    defaults = list(family = expr(stats::gaussian))
+    defaults = list(family = expr(stats::gaussian), refresh = 0)
   )
 )
 
@@ -293,7 +293,7 @@ set_model_arg(
   parsnip = "penalty",
   original = "reg_param",
   func = list(pkg = "dials", fun = "penalty"),
-  has_submodel = TRUE
+  has_submodel = FALSE
 )
 
 set_model_arg(

R/logistic_reg.R

@@ -42,6 +42,13 @@
 #'
 #' @section Engine Details:
 #'
+#' The standardized parameter names in parsnip can be mapped to their original
+#' names in each engine:
+#'
+#' ```{r echo = FALSE}
+#' convert_args("logistic_reg")
+#' ```
+#'
 #' Engines may have pre-set default arguments when executing the
 #'  model fit call.  For this type of
 #'  model, the template of the fit calls are:
@@ -58,6 +65,8 @@
 #'
 #' \Sexpr[results=rd]{parsnip:::show_fit(parsnip:::logistic_reg(), "stan")}
 #'
+#' (note that the `refresh` default prevents logging of the estimation process. Change this value in `set_engine()` will show the logs)
+#'
 #' \pkg{spark}
 #'
 #' \Sexpr[results=rd]{parsnip:::show_fit(parsnip:::logistic_reg(), "spark")}
@@ -97,7 +106,7 @@
 #'  separately saved to disk. In a new session, the object can be
 #'  reloaded and reattached to the `parsnip` object.
 #'
-#' @seealso [[fit()]
+#' @seealso [fit()]
 #' @examples
 #' logistic_reg()
 #' # Parameters can be represented by a placeholder:

R/logistic_reg_data.R

@@ -359,7 +359,7 @@ set_fit(
     interface = "formula",
     protect = c("formula", "data", "weights"),
     func = c(pkg = "rstanarm", fun = "stan_glm"),
-    defaults = list(family = expr(stats::binomial))
+    defaults = list(family = expr(stats::binomial), refresh = 0)
   )
 )
 

R/mars.R

@@ -38,6 +38,13 @@
 #'
 #' @section Engine Details:
 #'
+#' The standardized parameter names in parsnip can be mapped to their original
+#' names in each engine:
+#'
+#' ```{r echo = FALSE}
+#' convert_args("mars")
+#' ```
+#'
 #' Engines may have pre-set default arguments when executing the
 #'  model fit call.  For this type of
 #'  model, the template of the fit calls are:
@@ -55,7 +62,7 @@
 #'  attached.
 #'
 #' @importFrom purrr map_lgl
-#' @seealso [[fit()]
+#' @seealso [fit()]
 #' @examples
 #' mars(mode = "regression", num_terms = 5)
 #' @export

R/mlp.R

@@ -56,6 +56,13 @@
 #'
 #' @section Engine Details:
 #'
+#' The standardized parameter names in parsnip can be mapped to their original
+#' names in each engine:
+#'
+#' ```{r echo = FALSE}
+#' convert_args("mlp")
+#' ```
+#'
 #' Engines may have pre-set default arguments when executing the
 #'  model fit call. For this type of
 #'  model, the template of the fit calls are:
@@ -77,7 +84,7 @@
 #' \Sexpr[results=rd]{parsnip:::show_fit(parsnip:::mlp(mode = "regression"), "nnet")}
 #'
 #' @importFrom purrr map_lgl
-#' @seealso [[fit()]
+#' @seealso [fit()]
 #' @examples
 #' mlp(mode = "classification", penalty = 0.01)
 #' # Parameters can be represented by a placeholder:

R/multinom_reg.R

@@ -41,6 +41,13 @@
 #'
 #' @section Engine Details:
 #'
+#' The standardized parameter names in parsnip can be mapped to their original
+#' names in each engine:
+#'
+#' ```{r echo = FALSE}
+#' convert_args("multinom_reg")
+#' ```
+#'
 #' Engines may have pre-set default arguments when executing the
 #'  model fit call.  For this type of
 #'  model, the template of the fit calls are:
@@ -84,7 +91,7 @@
 #'  separately saved to disk. In a new session, the object can be
 #'  reloaded and reattached to the `parsnip` object.
 #'
-#' @seealso [[fit()]
+#' @seealso [fit()]
 #' @examples
 #' multinom_reg()
 #' # Parameters can be represented by a placeholder:

R/nearest_neighbor.R

@@ -48,6 +48,13 @@
 #'
 #' @section Engine Details:
 #'
+#' The standardized parameter names in parsnip can be mapped to their original
+#' names in each engine:
+#'
+#' ```{r echo = FALSE}
+#' convert_args("nearest_neighbor")
+#' ```
+#'
 #' Engines may have pre-set default arguments when executing the
 #'  model fit call. For this type of
 #'  model, the template of the fit calls are:
@@ -63,7 +70,7 @@
 #' on new data. This also means that a single value of that function's
 #' `kernel` argument (a.k.a `weight_func` here) can be supplied
 #'
-#' @seealso [[fit()]
+#' @seealso [fit()]
 #'
 #' @examples
 #' nearest_neighbor(neighbors = 11)