core_commons.R 13.9 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

33
require_auto <-  function(x){
34 35
    warning("require_auto is deprecated, use loadpack() instead")

Holger Brandl's avatar
Holger Brandl committed
36 37 38 39 40
    x <- as.character(substitute(x))

    if(isTRUE(x %in% .packages(all.available=TRUE))) {
        eval(parse(text=paste("require(", x, ",  quietly=T)", sep="")))
    } else {
41
        #        update.packages(ask=F) # update dependencies, if any.
Holger Brandl's avatar
Holger Brandl committed
42 43 44 45 46 47 48
        eval(parse(text=paste("install.packages('", x, "')", sep="")))
    }

    if(isTRUE(x %in% .packages(all.available=TRUE))) {
        eval(parse(text=paste("require(", x, ",  quietly=T)", sep="")))
    } else {
        source("http://bioconductor.org/biocLite.R")
49
        #        biocLite(character(), ask=FALSE) # update dependencies, if any.
50
        eval(parse(text=paste("biocLite('", x, "', ask=FALSE)", sep="")))
Holger Brandl's avatar
Holger Brandl committed
51 52 53 54
        eval(parse(text=paste("require(", x, ",  quietly=T)", sep="")))
    }
}

55
## externalized installer to also allow for installation without loading
56
install_package <- function(x){
57 58 59 60 61 62 63 64 65 66 67 68 69 70
    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="")))
        }
    }
71 72
}

Holger Brandl's avatar
Holger Brandl committed
73 74 75
load_pack <-  function(x, warn_conflicts=T){
    x <- as.character(substitute(x));

76
   install_package(x)
Holger Brandl's avatar
Holger Brandl committed
77 78 79 80 81

    ## load it using a library function so that loadpack errors if package is still not ins
    eval(parse(text=paste("library(", x, ",  quietly=T, warn.conflicts=", warn_conflicts, ")", sep="")))
}

82
loadpack <-  function(x, warn_conflicts=T){
Holger Brandl's avatar
Holger Brandl committed
83 84
    warning("DEPRECATED Use load_pack() instead")

85 86
    x <- as.character(substitute(x));

87
   install_package(x)
88 89

    ## load it using a library function so that loadpack errors if package is still not ins
Holger Brandl's avatar
Holger Brandl committed
90
    eval(parse(text=paste("library(", x, ",  quietly=T, warn.conflicts=", warn_conflicts, ")", sep="")))
91 92
}

93 94 95 96 97 98
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
99

Holger Brandl's avatar
Holger Brandl committed
100
########################################################################################################################
Holger Brandl's avatar
Holger Brandl committed
101
## load core packages
Holger Brandl's avatar
Holger Brandl committed
102

103 104 105
loadpack(plyr)
loadpack(stringr)
loadpack(reshape2)
106
#loadpack(reshape2, quietly=T, warn_conflicts=F)
Holger Brandl's avatar
Holger Brandl committed
107

Holger Brandl's avatar
Holger Brandl committed
108
## load on purpose after plyr
109 110 111
loadpack(dplyr, warn_conflicts=F)
loadpack(magrittr, warn_conflicts=F)
loadpack(tidyr, warn_conflicts=F)
112

Holger Brandl's avatar
Holger Brandl committed
113
## needed for caching
114 115 116 117
loadpack(digest)
loadpack(readr)
suppressWarnings(loadpack(readxl)) ## supress differring build number
#loadpack(readxl) ## supress differring build number
118 119 120


## common plotting requirements since they are omnipresent
121
loadpack(ggplot2)
122
loadpack(scales, warn_conflicts=F)
123
loadpack(grid)
124

Holger Brandl's avatar
Holger Brandl committed
125
## for table exploration without using Rstudio
126
loadpack(DT)
Holger Brandl's avatar
Holger Brandl committed
127

Holger Brandl's avatar
Holger Brandl committed
128 129

## moved into datatable_commons because replaced almost everywhere with dplyr
130
#loadpack(data.table)
Holger Brandl's avatar
Holger Brandl committed
131 132 133 134 135




########################################################################################################################
Holger Brandl's avatar
Holger Brandl committed
136
#### Convenience aliases
Holger Brandl's avatar
Holger Brandl committed
137 138


Holger Brandl's avatar
Holger Brandl committed
139
echo <- function(...) cat(paste(...), fill=T)
Holger Brandl's avatar
Holger Brandl committed
140 141 142 143

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

# string concatenation without space gaps (could/should use paste0 instead)
Holger Brandl's avatar
Holger Brandl committed
144 145
## Deprecated: use paste0 instead
#concat <- function(...) paste(..., sep="")
Holger Brandl's avatar
Holger Brandl committed
146 147 148 149 150

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

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

Holger Brandl's avatar
Holger Brandl committed
151 152
as.df <- function(dt) as.data.frame(dt)

Holger Brandl's avatar
Holger Brandl committed
153

Holger Brandl's avatar
Holger Brandl committed
154 155 156 157 158 159 160 161 162 163
########################################################################################################################
#### 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
164 165 166
#vec2df <- function(namedVec) namedVec %>% {data.frame(name=names(.), value=., row.names=NULL)}
vec2df <- function(namedVec) data.frame(name=names(namedVec), value=namedVec, row.names=NULL)

Holger Brandl's avatar
Holger Brandl committed
167 168

rownames2column <- function(df, colname){
169 170
    warning("DEPRECATED Use dplyr::add_rownames directly")
    add_rownames(df, var=colname)
Holger Brandl's avatar
Holger Brandl committed
171 172 173 174
}


column2rownames<- function(df, colname){
175
    #browser()
Holger Brandl's avatar
Holger Brandl committed
176 177 178
    ## force into df to avoid dplyr problems
    df <- as.df(df)

Holger Brandl's avatar
Holger Brandl committed
179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194
    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
195 196
#http://astrostatistics.psu.edu/datasets/R/html/base/html/formals.html
set_names <- function(df, ...){
197
    #browser()
Holger Brandl's avatar
Holger Brandl committed
198
    newnames <- as.character(unlist(list(...)))
199 200 201 202

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

203
    names(df) <- newnames;
Holger Brandl's avatar
Holger Brandl committed
204 205
    return(df)
}
Holger Brandl's avatar
Holger Brandl committed
206 207
#iris %>% set_names(c("setosa", "hallo")) %>% head
#iris %>% set_names("setosa", "hallo") %>% head
Holger Brandl's avatar
Holger Brandl committed
208 209


210 211 212 213 214 215
rify_names  <- function(df){
    warning("DEPRECATED: use pretty_columns instead");
    names(df) <- names(df) %>% str_replace_all(c(" "="_", "-"="_"));
    df
}

216 217 218 219 220 221 222 223
pretty_columns  <- function(df){
    names(df) <- names(df) %>%
        str_replace_all("[#=.()-]+", "_") %>%
        str_replace("[_]+$", "") %>%
        str_replace("^[_]+", "") %>%
        tolower;
    df
}
Holger Brandl's avatar
Holger Brandl committed
224

225 226
# 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
227

Holger Brandl's avatar
Holger Brandl committed
228 229
head_html <- function(df, n=5) head(df, n) %>% knitr::kable(format="html") %>% print()

Holger Brandl's avatar
Holger Brandl committed
230 231
print_head <- function(df, desc=NULL){
    print(head(df))
Holger Brandl's avatar
Holger Brandl committed
232
    print(nrow(df))
Holger Brandl's avatar
Holger Brandl committed
233 234 235 236 237
    return(df)
}


fac2char <- function(mydata, convert=names(mydata)[sapply(mydata, is.factor)]){
Holger Brandl's avatar
Holger Brandl committed
238 239 240
    if(length(convert)==0){
        return(mydata)
    }
Holger Brandl's avatar
Holger Brandl committed
241 242 243 244 245 246 247 248 249 250 251 252 253

    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
254
## replace R within pipe change just use ... %>% do(replaceNA(0)) %>% ...
Holger Brandl's avatar
Holger Brandl committed
255
replaceNA <- function(x, withValue) { x[is.na(x)] <- withValue; x }
Holger Brandl's avatar
Holger Brandl committed
256

Holger Brandl's avatar
Holger Brandl committed
257

Holger Brandl's avatar
Holger Brandl committed
258

Holger Brandl's avatar
Holger Brandl committed
259
## todo still needed since there's if_else in dplyr now
Holger Brandl's avatar
Holger Brandl committed
260
safe_ifelse <- function(cond, yes, no) {
Holger Brandl's avatar
Holger Brandl committed
261

Holger Brandl's avatar
Holger Brandl committed
262 263 264 265 266 267
#browser()
    isfacOrChar <- function(x) class(x)  %in% c("factor", "character")

    if (isfacOrChar(yes) | isfacOrChar(no)) {
        yes <- ac(yes)
        no <- ac(no)
268
    }
Holger Brandl's avatar
Holger Brandl committed
269 270

    ifelse(cond,yes,no)
Holger Brandl's avatar
Holger Brandl committed
271 272
}

Holger Brandl's avatar
Holger Brandl committed
273

Holger Brandl's avatar
Holger Brandl committed
274
## for na instead use mutate_each with:
Holger Brandl's avatar
Holger Brandl committed
275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302
#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
303 304 305
## see http://stackoverflow.com/questions/24172111/change-the-blank-cells-to-na/33952598#33952598


Holger Brandl's avatar
Holger Brandl committed
306 307 308 309 310 311 312
## apply dplyr::filter to df but use filter criterions for cross-tabulation beforehand
filter_count <- function(df, ...){
    print(count(df, ...))
    filter(df, ...)
}


313 314 315 316 317 318 319
reload_dplyr <- function(){
    unloadNamespace('tidyr')
    unloadNamespace('dplyr')
    require(tidyr);require(dplyr)
}


320
## from http://stackoverflow.com/questions/7505547/detach-all-packages-while-working-in-r
321
unload_packages <- function() {
322 323 324 325 326
    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
327 328


Holger Brandl's avatar
Holger Brandl committed
329 330 331 332
## workaround for biomart
## Deprecated: load dplyr after biomart to avoid this problem
#dselect <- function(...) dplyr::select(...)

Holger Brandl's avatar
Holger Brandl committed
333

Holger Brandl's avatar
Holger Brandl committed
334
########################################################################################################################
Holger Brandl's avatar
Holger Brandl committed
335
#### Result Caching for long running tasks
Holger Brandl's avatar
Holger Brandl committed
336

Holger Brandl's avatar
Holger Brandl committed
337 338
## related: http://cran.r-project.org/web/packages/R.cache/R.cache.pdf

339
cache_it <- function(expr, filePrefix="cache"){
340
    cacheFile <- paste0(filePrefix, "_", substr(digest::digest(deparse(expr)), 1,6)) %>% paste0(".", ., ".RData")
Holger Brandl's avatar
Holger Brandl committed
341

Holger Brandl's avatar
Holger Brandl committed
342 343 344 345 346 347 348 349 350 351 352
    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
353 354
#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
355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372

########################################################################################################################
#### 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
373
## 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
374 375 376 377 378 379 380 381 382 383
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
384 385
## Deprecated: use trim_ext instead
trimEnd <- function(fileNames, ...) trim_ext(fileNames, ...)
Holger Brandl's avatar
Holger Brandl committed
386

Holger Brandl's avatar
Holger Brandl committed
387
trim_ext <-function(fileNames, exts=c()){
Holger Brandl's avatar
Holger Brandl committed
388 389 390 391 392 393 394
    for(fileExt in exts){
        fileNames <- str_replace(fileNames, paste(fileExt, "$",sep=""), "")
    }

    fileNames
}

Holger Brandl's avatar
Holger Brandl committed
395
## DEPRECATED Use write_tsv instead
Holger Brandl's avatar
Holger Brandl committed
396 397
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
398 399 400 401 402 403 404 405 406 407
}

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

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


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

Holger Brandl's avatar
Holger Brandl committed
408

Holger Brandl's avatar
Holger Brandl committed
409
# improved list of objects
Holger Brandl's avatar
Holger Brandl committed
410 411
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
412
    names <- ls(pos = pos, pattern = pattern)
Holger Brandl's avatar
Holger Brandl committed
413

Holger Brandl's avatar
Holger Brandl committed
414 415 416 417 418
    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
419

420
    as.numeric(dim(x))[1:2]))
Holger Brandl's avatar
Holger Brandl committed
421 422
    vec <- is.na(obj.dim)[, 1] & (obj.type != "function")
    obj.dim[vec, 1] <- napply(names, length)[vec]
Holger Brandl's avatar
Holger Brandl committed
423

Holger Brandl's avatar
Holger Brandl committed
424 425
    out <- data.frame(obj.type, obj.size, obj.dim)
    names(out) <- c("Type", "Size", "Rows", "Columns")
Holger Brandl's avatar
Holger Brandl committed
426

Holger Brandl's avatar
Holger Brandl committed
427
    if (!missing(order.by))
428
    out <- out[order(out[[order.by]], decreasing=decreasing), ]
Holger Brandl's avatar
Holger Brandl committed
429 430 431

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

Holger Brandl's avatar
Holger Brandl committed
432 433 434 435 436
    out <- transform(out, var_name=rownames(out))
    rownames(out) <- NULL
    arrange(out, Size)
}

Holger Brandl's avatar
Holger Brandl committed
437
# shorthand that just shows top 1 results
Holger Brandl's avatar
Holger Brandl committed
438 439 440 441
lsosh <- function(..., n=10) {
    lsos(..., order.by="Size", decreasing=TRUE, head=TRUE, n=n)
}

Holger Brandl's avatar
Holger Brandl committed
442 443 444 445 446 447 448 449 450 451
########################################################################################################################
### 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
452
se <- function(x) sd(x, na.rm=TRUE) / sqrt(sum(!is.na(x)))
453 454 455 456 457 458 459 460 461


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

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