davidtedfordholt

obj_size <- function(obj, units = 'Mb') {
    dimensions <- dim(obj)
    data.frame(
        rows = format(dimensions[1], big.mark = ","),
        columns = format(dimensions[2], big.mark = ","),
        size = format(object.size(obj), units = units, big.mark = ",")
    )
}

davidtedfordholt

DECLARE dataset_names ARRAY<STRING>;
DECLARE batch ARRAY<STRING>;
DECLARE batch_size INT64 DEFAULT 25;

CREATE TEMP TABLE results (
    project_id STRING,
    dataset_id STRING,
    last_modified DATE,
    table_id STRING,
    row_count INT64,

davidtedfordholt

snippet ggstuff
	labs(title = "${1:title}",
		 subtitle = "${2:subtitle}",
	     x = "${3:xlab}", 
	     y = "${4:ylab}") +
	theme_minimal()

snippet gglabs
	labs(title = "${1:title}", 
	     x = "${2:xlab}", 

davidtedfordholt

#' Sample groups randomly in a grouped data frame
#'
#' @param .data dataframe
#' @param ... names of key variables to define groups. If unspecified, dataframe must be grouped.
#' @param n if integer, number of unique groups to keep. If between 1 and 0, proportion of unique groups to keep.
#'
#' @return tibble of all rows for sampled groups
#' @export
#'
#' @examples

davidtedfordholt

#' Scale a Variable Between a Given Min and Max
#'
#' @param x values to be scaled
#' @param new_min new minimum value
#' @param new_max new maximum value
#'
#' @return an object of the same class as `x`
#' @export
#'
#' @examples

davidtedfordholt

#' symlog transformation
#'
#' `symlog_trans()` transforms data using `log(x)` for `abs(x) > thr`, where
#' `thr` is a tuneable threshold, but leaves the data linear for `abs(x) < thr`.
#' (credit for base code to https://stackoverflow.com/users/1320535/julius-vainora)
#'
#'
#' @param base base of logarithm
#' @param thr numeric threshold for transitioning from log to linear
#' @param scale numeric scaling factor for data

davidtedfordholt

-- implementation
WITH input AS (
  SELECT *
  FROM UNNEST([100,100,100,100,100,100,100,100,100,100]) x
  JOIN UNNEST([.4,.4,.4,.4,.4,.4,.4,.4,.4,.4]) p
)
SELECT x * (1 + (-p + (RAND() * (2 * p)))) AS rand_within_p_percent_of_x
FROM input;

-- verification of concept

davidtedfordholt

sapply(list.files("R/", full.names = T), source)

davidtedfordholt

copy_to_gcp <- function(.data, filename_prefix, gcp_folder = NULL, gcp_bucket) {
    
    if (!startsWith(gcp_bucket, "gs://")) {
        gcp_bucket <- paste0("gs://", gcp_bucket)
    }
    source <- paste0(filename_prefix, "_", format(Sys.Date(), "%Y%m%d"), ".csv")
    destination <- ifelse(is.null(gcp_folder), gcp_bucket, paste0(gcp_bucket, "/", gcp_folder))
    
    packageStartupMessage("Writing file")
    write.csv(.data, source)

davidtedfordholt

## Plumber-verse Base Image #####################
FROM rocker/verse

RUN apt-get update -qq && apt-get install -y \
  git-core \
  libssl-dev \
  libcurl4-gnutls-dev

## Install Plumber