R/input_check.R

Defines functions new_error new_warning is.numeric_integer ensure_distances ensure_indicators ensure_scclust coerce_args coerce_character coerce_cluster_labels coerce_counts coerce_data_point_indices coerce_scalar_indicator coerce_radius coerce_size_constraint coerce_total_size_constraint coerce_type_constraints coerce_type_labels

# ==============================================================================
# scclust for R -- R wrapper for the scclust library
# https://github.com/fsavje/scclust-R
#
# Copyright (C) 2016-2017  Fredrik Savje -- http://fredriksavje.com
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see http://www.gnu.org/licenses/
# ==============================================================================


# ==============================================================================
# Helper functions
# ==============================================================================

# Throw error
new_error <- function(...) {
  stop(structure(list(message = paste0(...),
                      call = match.call(definition = sys.function(-2),
                                        call = sys.call(which = -2),
                                        expand.dots = TRUE,
                                        envir = sys.frame(-3))),
                 class = c("error", "condition")))
}


# Throw warning
new_warning <- function(...) {
  warning(structure(list(message = paste0(...),
                         call = match.call(definition = sys.function(-2),
                                           call = sys.call(which = -2),
                                           expand.dots = TRUE,
                                           envir = sys.frame(-3))),
                    class = c("warning", "condition")))
}


# Is `x` a numeric that can be coerced into integer without loss of information?
is.numeric_integer <- function(x) {
  is.numeric(x) &&
    !any(is.nan(x)) &&
    !any(is.infinite(x)) &&
    all(is.na(x) | as.integer(x) == x)
}


# ==============================================================================
# Ensure functions
# ==============================================================================

# Ensure that `distances` is `distances` object
ensure_distances <- function(distances,
                             req_length = NULL) {
  if (!distances::is.distances(distances)) {
    new_error("`", match.call()$distances, "` is not a `distances` object.")
  }
  if (!is.null(req_length) && (length(distances) != req_length)) {
    new_error("`", match.call()$distances, "` does not contain `", match.call()$req_length, "` data points.")
  }
}


# Ensure that `indicators` are non-NA logicals
ensure_indicators <- function(indicators,
                              req_length = NULL,
                              any_true = FALSE) {
  if (!is.logical(indicators)) {
    new_error("`", match.call()$indicators, "` must be logical.")
  }
  if (any(is.na(indicators))) {
    new_error("`", match.call()$indicators, "` may not contain NAs.")
  }
  if (!is.null(req_length) && (length(indicators) != req_length)) {
    new_error("`", match.call()$indicators, "` is not of length `", match.call()$req_length, "`.")
  }
  if (any_true) {
    if (!any(indicators)) {
      new_error("`", match.call()$indicators, "` cannot be all `FALSE`.")
    }
  }
}


# Ensure that `clustering` is a `scclust` object
ensure_scclust <- function(clustering,
                           req_length = NULL) {
  if (!is.scclust(clustering)) {
    new_error("`", match.call()$clustering, "` is not a `scclust` object.")
  }
  if (!is.null(req_length) && (length(clustering) != req_length)) {
    new_error("`", match.call()$clustering, "` does not contain `", match.call()$req_length, "` data points.")
  }
}


# ==============================================================================
# Coerce functions
# ==============================================================================

# Similar to `match.arg` but with custom error message
coerce_args <- function(arg,
                        choices) {
  stopifnot(is.character(choices),
            length(choices) > 0)
  if (!is.character(arg) || (length(arg) != 1L)) {
    new_error("`", match.call()$arg, "` must be character scalar.")
  }
  i <- pmatch(arg, choices, nomatch = 0L)
  if (i == 0) {
    new_error("`", match.call()$arg, "` must be one of ", paste0(paste0("\"", choices, "\""), collapse = ", "), ".")
  }
  choices[i]
}


# Coerce `x` to character vector
coerce_character <- function(x,
                             req_length = NULL) {
  x <- as.character(x)
  if (!is.null(req_length) && (length(x) != req_length)) {
    new_error("`", match.call()$x, "` is not of length `", match.call()$req_length, "`.")
  }
  x
}


# Coerce `cluster_labels` to factor with appropriate NAs
coerce_cluster_labels <- function(cluster_labels,
                                  unassigned_labels = NULL) {
  if (!is.factor(cluster_labels) && !is.vector(cluster_labels)) {
    new_error("`", match.call()$cluster_labels, "` must be factor or vector.")
  }
  if (!is.null(unassigned_labels) && !all(is.na(unassigned_labels))) {
    if (!all(unassigned_labels %in% cluster_labels)) {
      new_error("`", match.call()$unassigned_labels, "` contains entries not in `", match.call()$cluster_labels, "`.")
    }
    cluster_labels[cluster_labels %in% unassigned_labels] <- NA
  }
  if (is.vector(cluster_labels) || !is.null(unassigned_labels)) {
    cluster_labels <- factor(cluster_labels)
  }
  cluster_labels
}


# Coerce `counts` to non-NA, non-negative integers
coerce_counts <- function(counts,
                          req_length = NULL) {
  if (!is.integer(counts)) {
    if (is.numeric_integer(counts)) {
      storage.mode(counts) <- "integer"
    } else {
      new_error("`", match.call()$counts, "` must be integer.")
    }
  }
  if (any(is.na(counts))) {
    new_error("`", match.call()$counts, "` may not contain NAs.")
  }
  if (any(counts < 0L)) {
    new_error("`", match.call()$counts, "` must be non-negative.")
  }
  if (!is.null(req_length) && (length(counts) != req_length)) {
    new_error("`", match.call()$counts, "` is not of length `", match.call()$req_length, "`.")
  }
  counts
}


# Coerce data point indices
coerce_data_point_indices <- function(indices,
                                      num_data_points) {
  if (is.null(indices)) {
    # do nothing
  } else if (is.logical(indices)) {
    if (any(is.na(indices))) {
      new_error("`", match.call()$indices, "` may not contain NAs.")
    }
    if (!any(indices)) {
      new_error("`", match.call()$indices, "` cannot be all `FALSE`.")
    }
    if (length(indices) != num_data_points) {
      new_error("`", match.call()$indices, "` is not of length `", match.call()$num_data_points, "`.")
    }
  } else {
    if (!is.integer(indices)) {
      if (is.numeric_integer(indices)) {
        storage.mode(indices) <- "integer"
      } else {
        new_error("`", match.call()$indices, "` must be integer, logical or NULL.")
      }
    }
    if (any(is.na(indices))) {
      new_error("`", match.call()$indices, "` may not contain NAs.")
    }
    if (any(indices < 1L)) {
      new_error("`", match.call()$indices, "` must be positive.")
    }
    if (length(indices) == 0) {
      new_error("`", match.call()$indices, "` cannot be empty.")
    }
  }

  indices
}


coerce_scalar_indicator <- function(x) {
  if (is.null(x)) {
    x <- FALSE
  }
  x <- as.logical(x)[1]
  if (is.na(x)) {
    new_error("`", match.call()$x, "` must be TRUE or FALSE.")
  }
  x
}


# Coerce `radius` to NULL or a scalar, positive, non-na, numeric
coerce_radius <- function(radius,
                          is_seed = FALSE) {
  if (!is.null(radius)) {
    if (length(radius) != 1L) {
      new_error("`", match.call()$radius, "` must be scalar.")
    }
    if (is.na(radius)) {
      new_error("`", match.call()$radius, "` may not be NA.")
    }
    if (is.character(radius) && !is_seed) {
      choices <- c("no_radius", "seed_radius", "estimated_radius")
      i <- pmatch(radius, choices, nomatch = 0L)
      if (i == 0) {
        new_error("`", match.call()$radius, "` must be one of ", paste0(paste0("\"", choices, "\""), collapse = ", "), ".")
      }
      radius <- choices[i]
    } else if (is.numeric(radius)) {
      if (radius <= 0.0) {
        new_error("`", match.call()$radius, "` must be positive.")
      }
      # If `radius` is integer
      radius <- as.numeric(radius)
    } else if (is_seed) {
      new_error("`", match.call()$radius, "` must be numeric or `NULL`.")
    } else {
      new_error("`", match.call()$radius, "` must be numeric, character or `NULL`.")
    }
  }
  radius
}


# Coerce `size_constraint` to scalar, non-NA integer
coerce_size_constraint <- function(size_constraint,
                                   num_data_points) {
  if (length(size_constraint) != 1L) {
    new_error("`", match.call()$size_constraint, "` must be scalar.")
  }
  if (!is.integer(size_constraint)) {
    if (is.numeric_integer(size_constraint)) {
      storage.mode(size_constraint) <- "integer"
    } else {
      new_error("`", match.call()$size_constraint, "` must be integer.")
    }
  }
  if (is.na(size_constraint)) {
    new_error("`", match.call()$size_constraint, "` may not be NA.")
  }
  if (size_constraint < 2L) {
    new_error("`", match.call()$size_constraint, "` must be greater or equal to two.")
  }
  if (size_constraint > num_data_points) {
    new_error("`", match.call()$size_constraint, "` may not be great than the number of data points.")
  }
  size_constraint
}


# Coerce `total_size_constraint` to scalar, non-NA integer with default as `sum(type_constraints)`
coerce_total_size_constraint <- function(total_size_constraint,
                                         type_constraints,
                                         num_data_points) {
  sum_type_constraints <- sum(type_constraints)
  stopifnot(is.integer(type_constraints),
            sum_type_constraints >= 0L)

  if (is.null(total_size_constraint)) {
    total_size_constraint <- sum_type_constraints
  }
  if (length(total_size_constraint) != 1L) {
    new_error("`", match.call()$total_size_constraint, "` must be scalar.")
  }
  if (!is.integer(total_size_constraint)) {
    if (is.numeric_integer(total_size_constraint)) {
      storage.mode(total_size_constraint) <- "integer"
    } else {
      new_error("`", match.call()$total_size_constraint, "` must be integer.")
    }
  }
  if (is.na(total_size_constraint)) {
    new_error("`", match.call()$total_size_constraint, "` may not be NA.")
  }
  if (total_size_constraint < 2L) {
    new_error("`", match.call()$total_size_constraint, "` must be greater or equal to two.")
  }
  if (total_size_constraint < sum_type_constraints) {
    new_error("`", match.call()$total_size_constraint, "` must be greater or equal to the sum of the type constraints.")
  }
  if (total_size_constraint > num_data_points) {
    new_error("`", match.call()$total_size_constraint, "` may not be great than the number of data points.")
  }
  total_size_constraint
}


# Coerce `type_constraints` to valid type constraints
coerce_type_constraints <- function(type_constraints) {
  if (is.null(names(type_constraints))) {
    new_error("`", match.call()$type_constraints, "` must be named.")
  }
  if (anyDuplicated(names(type_constraints))) {
    new_error("`", match.call()$type_constraints, "` may not contain duplicate names.")
  }
  if (!is.integer(type_constraints)) {
    if (is.numeric_integer(type_constraints)) {
      storage.mode(type_constraints) <- "integer"
    } else {
      new_error("`", match.call()$type_constraints, "` must be integer.")
    }
  }
  if (any(is.na(type_constraints))) {
    new_error("`", match.call()$type_constraints, "` may not contain NAs.")
  }
  if (any(type_constraints < 0L)) {
    new_error("`", match.call()$type_constraints, "` must be non-negative.")
  }
  type_constraints
}


# Coerce `type_labels` to non-NA factor or non-NA, non-negative integer
coerce_type_labels <- function(type_labels,
                               req_length = NULL) {
  if (!is.factor(type_labels) && !is.integer(type_labels)) {
    if (is.numeric(type_labels)) {
      if (is.numeric_integer(type_labels)) {
        type_labels <- as.integer(type_labels)
      } else {
        new_error("`", match.call()$type_labels, "` must be integer or factor.")
      }
    } else if (is.character(type_labels)) {
      type_labels <- as.factor(type_labels)
    } else {
      new_warning("Coercing `", match.call()$type_labels, "` to factor.")
      type_labels <- as.factor(type_labels)
    }
  }
  if (any(is.na(type_labels))) {
    new_error("`", match.call()$type_labels, "` may not contain NAs.")
  }
  if (is.integer(type_labels) && any(type_labels < 0L)) {
    new_error("`", match.call()$type_labels, "` may not contain negtive entries.")
  }
  if (!is.null(req_length) && (length(type_labels) != req_length)) {
    new_error("`", match.call()$type_labels, "` is not of length `", match.call()$req_length, "`.")
  }
  type_labels
}

Try the scclust package in your browser

Any scripts or data that you put into this service are public.

scclust documentation built on May 2, 2019, 4:04 p.m.