core_commons.R 14 KB
Newer Older
Holger Brandl's avatar
Holger Brandl committed
1 2 3 4
########################################################################################################################
## set a default cran r mirror  and customize environment

#cat(".Rprofile: Setting Goettingen repository\n")
Holger Brandl's avatar
Holger Brandl committed
5 6
#todo consider to use chooseCRANmirror(graphics=FALSE, ind=10) instead

Holger Brandl's avatar
Holger Brandl committed
7 8 9 10 11 12 13 14 15 16 17 18
r = getOption("repos") # hard code the UK repo for CRAN

r["CRAN"] = "http://ftp5.gwdg.de/pub/misc/cran/"
options(repos = r)
rm(r)


## user browser for help
options(help_type="html")

## plot more characters per line
options(width=150)

19 20 21 22 23
## adjust dplyr printing settings
## http://stackoverflow.com/questions/22471256/overriding-variables-not-shown-in-dplyr-to-display-all-columns-from-df
## http://stackoverflow.com/questions/29396111/dplyrtbl-df-fill-whole-screen
options(dplyr.print_max = 20) ## num rows
options(dplyr.width = 130) ## width
Holger Brandl's avatar
Holger Brandl committed
24
#options(dplyr.width = 280); options(width=280) ## width
25

Holger Brandl's avatar
Holger Brandl committed
26 27 28
# for sqldf to avoid the use of tckl
options(gsubfn.engine = "R")

Holger Brandl's avatar
Holger Brandl committed
29

Holger Brandl's avatar
Holger Brandl committed
30
########################################################################################################################
Holger Brandl's avatar
Holger Brandl committed
31
## automatic package installation
Holger Brandl's avatar
Holger Brandl committed
32

Holger Brandl's avatar
Holger Brandl committed
33

34
## externalized installer to also allow for installation without loading
35
install_package <- function(x){
36 37 38 39 40 41 42 43 44 45 46 47 48 49
    if(!isTRUE(x %in% .packages(all.available=TRUE)) && any(available.packages()[,1]==x)) {
        # update.packages(ask=F) # update dependencies, if any.
        eval(parse(text=paste("install.packages('", x, "')", sep="")))
    }

    ## if it's still missing check if it's on bioconductor
    if(!isTRUE(x %in% .packages(all.available=TRUE))) {
        bcPackages <- as.vector(read.dcf(url("https://bioconductor.org/packages/3.3/bioc/src/contrib/PACKAGES"), "Package"))

        if(any(bcPackages==x)){
            source("http://bioconductor.org/biocLite.R")
            eval(parse(text=paste("biocLite('", x, "', ask=FALSE)", sep="")))
        }
    }
50 51
}

Holger Brandl's avatar
Holger Brandl committed
52 53 54
load_pack <-  function(x, warn_conflicts=T){
    x <- as.character(substitute(x));

55
   install_package(x)
Holger Brandl's avatar
Holger Brandl committed
56

57
    ## load it using a library function so that load_pack errors if package is still not ins
Holger Brandl's avatar
Holger Brandl committed
58 59 60
    eval(parse(text=paste("library(", x, ",  quietly=T, warn.conflicts=", warn_conflicts, ")", sep="")))
}

61

62 63 64 65 66 67
check_version = function(pkg_name, min_version) {
    cur_version = packageVersion(pkg_name)
    if(cur_version < min_version) stop(sprintf("Package %s needs a newer version,
               found %s, need at least %s", pkg_name, cur_version, min_version))
}
#check_version("dplyr", "0.4-1")
Holger Brandl's avatar
Holger Brandl committed
68

Holger Brandl's avatar
Holger Brandl committed
69
########################################################################################################################
Holger Brandl's avatar
Holger Brandl committed
70
## load core packages
Holger Brandl's avatar
Holger Brandl committed
71

Holger Brandl's avatar
Holger Brandl committed
72 73
#load_pack(plyr)
#load_pack(reshape2)
74
#load_pack(reshape2, quietly=T, warn_conflicts=F)
Holger Brandl's avatar
Holger Brandl committed
75

Holger Brandl's avatar
Holger Brandl committed
76
## load on purpose after plyr
77 78 79
load_pack(dplyr, warn_conflicts=F)
load_pack(magrittr, warn_conflicts=F)
load_pack(tidyr, warn_conflicts=F)
Holger Brandl's avatar
Holger Brandl committed
80 81 82 83
load_pack(stringr)
load_pack(purrr)
load_pack(readr)
load_pack(readxl) ## supress differring build number
84

Holger Brandl's avatar
Holger Brandl committed
85
## needed for caching
86
load_pack(digest)
Holger Brandl's avatar
Holger Brandl committed
87 88 89

#suppressWarnings(load_pack(readxl)) ## supress differring build number

90
#load_pack(readxl) ## supress differring build number
91 92 93


## common plotting requirements since they are omnipresent
94 95 96
load_pack(ggplot2)
load_pack(scales, warn_conflicts=F)
load_pack(grid)
97

Holger Brandl's avatar
Holger Brandl committed
98
## for table exploration without using Rstudio
99
load_pack(DT)
Holger Brandl's avatar
Holger Brandl committed
100

Holger Brandl's avatar
Holger Brandl committed
101 102

## moved into datatable_commons because replaced almost everywhere with dplyr
103
#load_pack(data.table)
Holger Brandl's avatar
Holger Brandl committed
104 105 106 107 108




########################################################################################################################
Holger Brandl's avatar
Holger Brandl committed
109
#### Convenience aliases
Holger Brandl's avatar
Holger Brandl committed
110 111


Holger Brandl's avatar
Holger Brandl committed
112
echo <- function(...) cat(paste(...), fill=T)
Holger Brandl's avatar
Holger Brandl committed
113 114 115 116

ac <- function(...) as.character(...)

# string concatenation without space gaps (could/should use paste0 instead)
Holger Brandl's avatar
Holger Brandl committed
117 118
## Deprecated: use paste0 instead
#concat <- function(...) paste(..., sep="")
Holger Brandl's avatar
Holger Brandl committed
119 120 121 122 123

unlen <- function(x) length(unique(x))

pp <- function(dat) page(dat, method = "print")

Holger Brandl's avatar
Holger Brandl committed
124 125
as.df <- function(dt){ warning("DEPRECATED: use as_df instead of as.df"); as.data.frame(dt) }
as_df <- function(dt) as.data.frame(dt)
Holger Brandl's avatar
Holger Brandl committed
126

Holger Brandl's avatar
Holger Brandl committed
127

Holger Brandl's avatar
Holger Brandl committed
128 129 130 131 132
install_package("tibble")

## restore pre-tibble-v1.2 naming to creating data-frame in place
frame_data = function(...) tibble::tribble(...)

133 134 135

add_rownames = function(...) tibble::rownames_to_column(...)

Holger Brandl's avatar
Holger Brandl committed
136 137 138 139 140 141 142 143 144 145 146 147 148

## redefine dply::splat to allow for more streamlined rowwise df-processing
## see https://groups.google.com/forum/?utm_source=digest&utm_medium=email#!topic/manipulatr/E6WOcHlRJcw
#splat = function (flat) {
#    function(args, ...) {
#        do.call(flat, c(args, list(...)))
#    }
#}
## for now simply import just splat from plyr namespace
install_package("plyr")
splat = plyr::splat


Holger Brandl's avatar
Holger Brandl committed
149 150 151 152 153 154 155 156 157 158
########################################################################################################################
#### data.frame manipulation


shuffle <- function(df) df[sample(nrow(df)),]

first <- function(x, n=1) head(x,n)



Holger Brandl's avatar
Holger Brandl committed
159 160 161 162 163 164 165 166
vec2df <- function(namedVec){
    warning("DEPRECATED use vec_as_df instead of vec2df")
    namedVec %>% {data.frame(name=names(.), value=., row.names=NULL)}
}

vec_as_df <- function(namedVec, row_name="name", value_name="value"){
    data_frame(name=names(namedVec), value=namedVec) %>% set_names(row_name, value_name)
}
Holger Brandl's avatar
Holger Brandl committed
167

Holger Brandl's avatar
Holger Brandl committed
168 169

rownames2column <- function(df, colname){
170 171
    warning("DEPRECATED Use dplyr::add_rownames directly")
    add_rownames(df, var=colname)
Holger Brandl's avatar
Holger Brandl committed
172 173 174 175
}


column2rownames<- function(df, colname){
176
    #browser()
Holger Brandl's avatar
Holger Brandl committed
177
    ## force into df to avoid dplyr problems
Holger Brandl's avatar
Holger Brandl committed
178
    df <- as_df(df)
Holger Brandl's avatar
Holger Brandl committed
179

Holger Brandl's avatar
Holger Brandl committed
180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195
    rownames(df) <- ac(df[,colname])
    df[colname] <- NULL
    return(df)
}

## pushing some columns to the right end of a data.frame
push_right <- function(df, pushColNames){
    df[,c(setdiff(names(df),pushColNames), pushColNames)]
}


push_left <- function(df, pushColNames){
    df[,c(pushColNames, setdiff(names(df),pushColNames))]
}


Holger Brandl's avatar
Holger Brandl committed
196 197
#http://astrostatistics.psu.edu/datasets/R/html/base/html/formals.html
set_names <- function(df, ...){
198
    #browser()
Holger Brandl's avatar
Holger Brandl committed
199
    newnames <- as.character(unlist(list(...)))
200 201 202 203

    ## automatically convert matrices to data.frames (otherwise the names set would fail
    if(is.matrix(df)) df %<>% as.data.frame()

204
    names(df) <- newnames;
Holger Brandl's avatar
Holger Brandl committed
205 206
    return(df)
}
Holger Brandl's avatar
Holger Brandl committed
207 208
#iris %>% set_names(c("setosa", "hallo")) %>% head
#iris %>% set_names("setosa", "hallo") %>% head
Holger Brandl's avatar
Holger Brandl committed
209 210


211 212
rify_names  <- function(df){
    warning("DEPRECATED: use pretty_columns instead");
Holger Brandl's avatar
Holger Brandl committed
213
    names(df) <- names(df) %>% str_replace_all(c(" "="_", "-"="_", "/"="_"));
214 215 216
    df
}

217 218
pretty_columns  <- function(df){
    names(df) <- names(df) %>%
Holger Brandl's avatar
Holger Brandl committed
219
        str_replace_all("[#=.()/ -]+", "_") %>%
220 221 222 223 224
        str_replace("[_]+$", "") %>%
        str_replace("^[_]+", "") %>%
        tolower;
    df
}
Holger Brandl's avatar
Holger Brandl committed
225

226 227
# http://stackoverflow.com/questions/23188900/view-entire-dataframe-when-wrapped-in-tbl-df
print_all <- function(df) df %>% tbl_df %>% print(n = nrow(.))
Holger Brandl's avatar
Holger Brandl committed
228

Holger Brandl's avatar
Holger Brandl committed
229 230
head_html <- function(df, n=5) head(df, n) %>% knitr::kable(format="html") %>% print()

Holger Brandl's avatar
Holger Brandl committed
231 232
print_head <- function(df, desc=NULL){
    print(head(df))
Holger Brandl's avatar
Holger Brandl committed
233
    print(nrow(df))
Holger Brandl's avatar
Holger Brandl committed
234 235 236 237 238
    return(df)
}


fac2char <- function(mydata, convert=names(mydata)[sapply(mydata, is.factor)]){
Holger Brandl's avatar
Holger Brandl committed
239 240 241
    if(length(convert)==0){
        return(mydata)
    }
Holger Brandl's avatar
Holger Brandl committed
242 243 244 245 246 247 248 249 250 251 252 253 254

    inputColOrder <- names(mydata)

    convertData <- subset(mydata, select= names(mydata)%in%convert)
    convertData <- as.data.frame(lapply(convertData, as.character), stringsAsFactors = FALSE)

    keepData <-  subset(mydata, select=!(names(mydata)%in%convert))
    newdata <- cbind(convertData, keepData)
    newdata <- newdata[,inputColOrder]

    return(newdata)
}

Holger Brandl's avatar
Holger Brandl committed
255
## replace R within pipe change just use ... %>% do(replaceNA(0)) %>% ...
Holger Brandl's avatar
Holger Brandl committed
256
replaceNA <- function(x, withValue) { x[is.na(x)] <- withValue; x }
Holger Brandl's avatar
Holger Brandl committed
257

Holger Brandl's avatar
Holger Brandl committed
258

Holger Brandl's avatar
Holger Brandl committed
259

Holger Brandl's avatar
Holger Brandl committed
260
## todo still needed since there's if_else in dplyr now
Holger Brandl's avatar
Holger Brandl committed
261
safe_ifelse <- function(cond, yes, no) {
Holger Brandl's avatar
Holger Brandl committed
262

Holger Brandl's avatar
Holger Brandl committed
263 264 265 266 267 268
#browser()
    isfacOrChar <- function(x) class(x)  %in% c("factor", "character")

    if (isfacOrChar(yes) | isfacOrChar(no)) {
        yes <- ac(yes)
        no <- ac(no)
269
    }
Holger Brandl's avatar
Holger Brandl committed
270 271

    ifelse(cond,yes,no)
Holger Brandl's avatar
Holger Brandl committed
272 273
}

Holger Brandl's avatar
Holger Brandl committed
274

Holger Brandl's avatar
Holger Brandl committed
275
## for na instead use mutate_each with:
Holger Brandl's avatar
Holger Brandl committed
276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303
#empty_as_na <- function(x) safe_ifelse(x=="", NA, x)
#empty_as_na <- function(x) ifelse(class(x)  %in% c("factor", "character") & x=="", NA, x)
empty_as_na <- function(x){
    if("factor" %in% class(x)) x <- as.character(x) ## since ifelse wont work with factors
    ifelse(as.character(x)!="", x, NA)
}

#if(F){ ## DEBUG of empty_as_na
#cond <- allJobs %>% head %$% submit_time %>% c("")
#empty_as_na( cond)
#cond <- allJobs %>% head %$% resubmission_of
#empty_as_na( cond)
#
#empty_as_na( c(1, 2, NA))
#empty_as_na( c("sdf", "sdf2", NA))
#empty_as_na( c("sdf", "sdf2", ""))
#
#myFac <- as.factor(c("sdf", "sdf2", NA))
#empty_as_na( myFac)
#ifelse(as.character(myFac)!="", myFac, NA)
#
#empty_as_na( c("sdf", "sdf2", ""))
#
#iris[1,1] <- ""
#apply(iris, 2, function(x) gsub("^$|^ $", NA, x))
#}


Holger Brandl's avatar
Holger Brandl committed
304 305 306
## see http://stackoverflow.com/questions/24172111/change-the-blank-cells-to-na/33952598#33952598


Holger Brandl's avatar
Holger Brandl committed
307 308 309 310 311 312
## apply dplyr::filter to df but use filter criterions for cross-tabulation beforehand
filter_count <- function(df, ...){
    print(count(df, ...))
    filter(df, ...)
}

Holger Brandl's avatar
Holger Brandl committed
313 314 315 316 317 318
# Example: publications %>% count(journal) %>% as("num_pubs")
as = function(df, name){
    names(df)[length(names(df))] = name
    df
}

Holger Brandl's avatar
Holger Brandl committed
319

Holger Brandl's avatar
Holger Brandl committed
320 321
distinct_all = function (x, ...) distinct(x, ..., .keep_all=T)

Holger Brandl's avatar
Holger Brandl committed
322

323 324 325 326 327 328 329
reload_dplyr <- function(){
    unloadNamespace('tidyr')
    unloadNamespace('dplyr')
    require(tidyr);require(dplyr)
}


330
## from http://stackoverflow.com/questions/7505547/detach-all-packages-while-working-in-r
331
unload_packages <- function() {
332 333 334 335 336
    basic.packages <- c("package:stats","package:graphics","package:grDevices","package:utils","package:datasets","package:methods","package:base")
    package.list <- search()[ifelse(unlist(gregexpr("package:",search()))==1,TRUE,FALSE)]
    package.list <- setdiff(package.list,basic.packages)
    if (length(package.list)>0)  for (package in package.list) detach(package, character.only=TRUE)
}
Holger Brandl's avatar
Holger Brandl committed
337 338


Holger Brandl's avatar
Holger Brandl committed
339 340 341 342
## workaround for biomart
## Deprecated: load dplyr after biomart to avoid this problem
#dselect <- function(...) dplyr::select(...)

Holger Brandl's avatar
Holger Brandl committed
343

Holger Brandl's avatar
Holger Brandl committed
344
########################################################################################################################
Holger Brandl's avatar
Holger Brandl committed
345
#### Result Caching for long running tasks
Holger Brandl's avatar
Holger Brandl committed
346

Holger Brandl's avatar
Holger Brandl committed
347 348
## related: http://cran.r-project.org/web/packages/R.cache/R.cache.pdf

349
cache_it <- function(expr, filePrefix="cache"){
350
    cacheFile <- paste0(filePrefix, "_", substr(digest::digest(deparse(expr)), 1,6)) %>% paste0(".", ., ".RData")
Holger Brandl's avatar
Holger Brandl committed
351

Holger Brandl's avatar
Holger Brandl committed
352 353 354 355 356 357 358 359 360 361 362
    if(file.exists(cacheFile)){
        local(get(load(cacheFile)))
    } else {
        result <- eval(expr)
        save(result, file=cacheFile)
        result
    }
}

## Examples
#mydata <- quote(iris %>% filter(Species=="setosa")) %>% cache_it("tt")
Holger Brandl's avatar
Holger Brandl committed
363 364
#mydata <- quote(iris %>% filter(Species=="setosa")) %>% cache_it()
#mydata <- quote( { print("evaluate expr"); iris %>% filter(Species=="setosa") } ) %>% cache_it()
Holger Brandl's avatar
Holger Brandl committed
365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382

########################################################################################################################
#### File System

is.directory <- function(dirname) !is.na(file.info(dirname)$isdir)


mcdir <- function(dirname){
    if(!file.exists(dirname)){
        dir.create(dirname)
    }

    setwd(dirname)
}

locload <- function(fileName) local(get(load(fileName)))


Holger Brandl's avatar
Holger Brandl committed
383
## tbd: it would be more efficient to use Reduce here (see http://stackoverflow.com/questions/34344214/how-to-join-multiple-data-frames-using-dplyr)
Holger Brandl's avatar
Holger Brandl committed
384 385 386 387 388 389 390 391 392 393
rmerge <- function(LDF, by, ...){
    DF <- LDF[[1]]
    for (i in 2:length(LDF)) {
        DF <- merge(DF, LDF[[i]], by=by)
    }
    DF
}



Holger Brandl's avatar
Holger Brandl committed
394 395
## Deprecated: use trim_ext instead
trimEnd <- function(fileNames, ...) trim_ext(fileNames, ...)
Holger Brandl's avatar
Holger Brandl committed
396

Holger Brandl's avatar
Holger Brandl committed
397
trim_ext <-function(fileNames, exts=c()){
Holger Brandl's avatar
Holger Brandl committed
398 399 400 401 402 403 404
    for(fileExt in exts){
        fileNames <- str_replace(fileNames, paste(fileExt, "$",sep=""), "")
    }

    fileNames
}

Holger Brandl's avatar
Holger Brandl committed
405
## DEPRECATED Use write_tsv instead
Holger Brandl's avatar
Holger Brandl committed
406 407
write.delim <- function(df, file, header=TRUE,...){
    write.table(df, file, row.names=FALSE, col.names=header, sep="\t", ...)
Holger Brandl's avatar
Holger Brandl committed
408 409 410 411 412 413 414 415 416 417
}

rmSomeElements <- function(vec, toDel) vec[!(vec %in% toDel)]

rmLastElement <- function(vec) vec[-length(vec)]


########################################################################################################################
## Memory management

Holger Brandl's avatar
Holger Brandl committed
418

Holger Brandl's avatar
Holger Brandl committed
419
# improved list of objects
Holger Brandl's avatar
Holger Brandl committed
420 421
lsos <- function (pos = 1, pattern, order.by, decreasing=FALSE, head=FALSE, n=5) {
    napply <- function(names, fn) sapply(names, function(x) fn(get(x, pos = pos)))
Holger Brandl's avatar
Holger Brandl committed
422
    names <- ls(pos = pos, pattern = pattern)
Holger Brandl's avatar
Holger Brandl committed
423

Holger Brandl's avatar
Holger Brandl committed
424 425 426 427 428
    obj.class <- napply(names, function(x) as.character(class(x))[1])
    obj.mode <- napply(names, mode)
    obj.type <- ifelse(is.na(obj.class), obj.mode, obj.class)
    obj.size <- napply(names, object.size)/1000000
    obj.dim <- t(napply(names, function(x)
Holger Brandl's avatar
Holger Brandl committed
429

430
    as.numeric(dim(x))[1:2]))
Holger Brandl's avatar
Holger Brandl committed
431 432
    vec <- is.na(obj.dim)[, 1] & (obj.type != "function")
    obj.dim[vec, 1] <- napply(names, length)[vec]
Holger Brandl's avatar
Holger Brandl committed
433

Holger Brandl's avatar
Holger Brandl committed
434 435
    out <- data.frame(obj.type, obj.size, obj.dim)
    names(out) <- c("Type", "Size", "Rows", "Columns")
Holger Brandl's avatar
Holger Brandl committed
436

Holger Brandl's avatar
Holger Brandl committed
437
    if (!missing(order.by))
438
    out <- out[order(out[[order.by]], decreasing=decreasing), ]
Holger Brandl's avatar
Holger Brandl committed
439 440 441

    if (head) out <- head(out, n)

Holger Brandl's avatar
Holger Brandl committed
442 443 444 445 446
    out <- transform(out, var_name=rownames(out))
    rownames(out) <- NULL
    arrange(out, Size)
}

Holger Brandl's avatar
Holger Brandl committed
447
# shorthand that just shows top 1 results
Holger Brandl's avatar
Holger Brandl committed
448 449 450 451
lsosh <- function(..., n=10) {
    lsos(..., order.by="Size", decreasing=TRUE, head=TRUE, n=n)
}

Holger Brandl's avatar
Holger Brandl committed
452 453 454 455 456 457 458 459 460 461
########################################################################################################################
### Statistics


## outlier handling
trim_outliers <- function(values, range=quantile(values, c(0.05, 0.95)))  pmax(range[1], pmin(range[2], values))

## use trim_outliers instead
#limit_range <- function(values, range)  pmax(range[1], pmin(range[2], values))

Holger Brandl's avatar
Holger Brandl committed
462
se <- function(x) sd(x, na.rm=TRUE) / sqrt(sum(!is.na(x)))
463 464 465 466 467 468 469 470


########################################################################################################################
### Misc

## inspired by http://stackoverflow.com/questions/8343509/better-error-message-for-stopifnot
assert <- function (expr, error) {
    if (! expr) stop(error, call. = FALSE)
471 472 473
}

### table rendering
Holger Brandl's avatar
Holger Brandl committed
474 475
table_browser <- function(df, caption=deparse(substitute(df)), ...){
    datatable(df, filter = "bottom", extensions = 'Buttons', options = list( dom = 'Bfrtip', buttons = c('copy', 'csv', 'excel')), caption=caption,...)
476
}