talegari

# dataset: https://zenodo.org/records/2594012
df = arrow::read_parquet("personal/Avazu/test.parquet") |> 
  tibble::as_tibble()

dim(df) # 4,218,938 X 24

res = 
  bench::mark(
    # using `duckplyr`
    duckplyr = {

talegari

-------------------------------------------------------------------
Instructions to make a R script/program executable on *nix machines
-------------------------------------------------------------------

Author: Srikanth KS
Date  : 12th February 2018

- We assume that R, Rscript and required R packages are installed.
- 'Rscript' is a program to execute scripts/programs written in R language.
- Add this shebang(without quotes) as the first line of the script: '#!/usr/bin/env Rscript'.

talegari

# Read 20newsgroups data as a datatable (dataframe)
# Author: Srikanth KS
# license: GPL-3
#
# download data from here:
# https://archive.ics.uci.edu/ml/machine-learning-databases/20newsgroups-mld/20_newsgroups.tar.gz
# extract it and provide its location to `baseDir` on line 9

baseDir        = "Downloads/20_newsgroups"
newsGroupNames = list.files(baseDir, full.names = TRUE)

talegari

#' @title cutq
#' @description Discretize a numeric vector along quantiles
#' @param vec numeric/integer vector
#' @param n number of buckets (atleast two)
#' @param ... extra named arguments passed to `cut`
#' @return A factor
#' @details By passing extra arguments to `cut`, output can be styled

cutq = function(vec, n = 10, ...){
  stopifnot(inherits(vec, "numeric") || inherits(vec, "integer"))

talegari

###############################################################################
#
# cor2
# -- Compute correlations of columns of a dataframe of mixed types
#
###############################################################################
#
# author  : Srikanth KS (talegari)
# license : GNU AGPLv3 (http://choosealicense.com/licenses/agpl-3.0/)
#

talegari

Installing dlib to access via python and R

Author: Srikanth KS

You might need minor modifications

For ubuntu 16.04:

sudo apt-get install build-essential cmake
sudo apt-get install libgtk-3-dev

talegari

#' @title require2
#'
#' @author Srikanth KS (talegari), gmail at sri dot teach GNU AGPLv3
#'   (http://choosealicense.com/licenses/agpl-3.0/)
#'
#' @param pkgname a string (character vector of length 1)
#' @param similar an positive integer indicating the number of similar package
#'   names to be suggested, if the match is not found
#'
#' @description The function attaches and loads a R library, if present and

talegari

Guide to install rcppmlpack2 on ubuntu trusty

until its released on CRAN

• sudo add-apt-repository ppa:edd/misc
• sudo apt-get update
• sudo apt-get install libmlpack-dev libboost-all-dev libboost-program-options-dev libboost-serialization-dev libarmadillo-dev r-cran-rcpp r-cran-rcpparmadillo

in R: devtools::install_github("eddelbuettel/rcppmlpack2")

talegari

R libraries for recommender systems

A list of R libraries for Recommender systems. Most of the libraries are good for quick prototyping.

Maintainer: Srikanth KS(talegari) Email: gmail me at sri dot teach (do write to me about packages ommited)

Package	Dev page	Description
recommenderlab	github	Provides a research infrastructure to test and develop recommender algorithms including UBCF, IBCF, FunkSVD and association rule-based algorithms
rrecsys	github	Implementations of several popular recommendation systems like Global/Item/User-Average baselines, Item-Based KNN, FunkSVD, BPR and weighted ALS for rapid prototyping

talegari

# how much memory did an execution use in MBs

mem_usage <- function(...){
  exprs <- as.list(match.call(expand.dots = FALSE)$...)
  invisible(gc(reset = TRUE))
  start_mem <- sum(gc()[,2])
  
  lapply(exprs, eval, parent.frame())
  
  max_mem   <- sum(gc()[,6])