core_commons.R 12.4 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 24
## 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
25 26 27
# for sqldf to avoid the use of tckl
options(gsubfn.engine = "R")

Holger Brandl's avatar
Holger Brandl committed
28

Holger Brandl's avatar
Holger Brandl committed
29
########################################################################################################################
Holger Brandl's avatar
Holger Brandl committed
30
## automatic package installation
Holger Brandl's avatar
Holger Brandl committed
31

32
require_auto <-  function(x){
33 34
    warning("require_auto is deprecated, use loadpack() instead")

Holger Brandl's avatar
Holger Brandl committed
35 36 37 38 39
    x <- as.character(substitute(x))

    if(isTRUE(x %in% .packages(all.available=TRUE))) {
        eval(parse(text=paste("require(", x, ",  quietly=T)", sep="")))
    } else {
40
        #        update.packages(ask=F) # update dependencies, if any.
Holger Brandl's avatar
Holger Brandl committed
41 42 43 44 45 46 47
        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")
48
        #        biocLite(character(), ask=FALSE) # update dependencies, if any.
49
        eval(parse(text=paste("biocLite('", x, "', ask=FALSE)", sep="")))
Holger Brandl's avatar
Holger Brandl committed
50 51 52 53
        eval(parse(text=paste("require(", x, ",  quietly=T)", sep="")))
    }
}

54

55
loadpack <-  function(x, warn_conflicts=T){
56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73
    x <- as.character(substitute(x));

    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="")))
        }
    }

    ## load it using a library function so that loadpack errors if package is still not ins
Holger Brandl's avatar
Holger Brandl committed
74
    eval(parse(text=paste("library(", x, ",  quietly=T, warn.conflicts=", warn_conflicts, ")", sep="")))
75 76
}

77 78 79 80 81 82
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
83

Holger Brandl's avatar
Holger Brandl committed
84
########################################################################################################################
Holger Brandl's avatar
Holger Brandl committed
85
## load core packages
Holger Brandl's avatar
Holger Brandl committed
86

87 88 89
loadpack(plyr)
loadpack(stringr)
loadpack(reshape2)
90
#loadpack(reshape2, quietly=T, warn_conflicts=F)
Holger Brandl's avatar
Holger Brandl committed
91

Holger Brandl's avatar
Holger Brandl committed
92
## load on purpose after plyr
93 94 95
loadpack(dplyr, warn_conflicts=F)
loadpack(magrittr, warn_conflicts=F)
loadpack(tidyr, warn_conflicts=F)
96

Holger Brandl's avatar
Holger Brandl committed
97
## needed for caching
98 99 100 101
loadpack(digest)
loadpack(readr)
suppressWarnings(loadpack(readxl)) ## supress differring build number
#loadpack(readxl) ## supress differring build number
102 103 104


## common plotting requirements since they are omnipresent
105
loadpack(ggplot2)
106
loadpack(scales, warn_conflicts=F)
107
loadpack(grid)
108

Holger Brandl's avatar
Holger Brandl committed
109
## for table exploration without using Rstudio
110
loadpack(DT)
Holger Brandl's avatar
Holger Brandl committed
111

Holger Brandl's avatar
Holger Brandl committed
112 113

## moved into datatable_commons because replaced almost everywhere with dplyr
114
#loadpack(data.table)
Holger Brandl's avatar
Holger Brandl committed
115 116 117 118 119




########################################################################################################################
Holger Brandl's avatar
Holger Brandl committed
120
#### Convenience aliases
Holger Brandl's avatar
Holger Brandl committed
121 122


Holger Brandl's avatar
Holger Brandl committed
123
echo <- function(...) cat(paste(...), fill=T)
Holger Brandl's avatar
Holger Brandl committed
124 125 126 127

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

# string concatenation without space gaps (could/should use paste0 instead)
Holger Brandl's avatar
Holger Brandl committed
128 129
## Deprecated: use paste0 instead
#concat <- function(...) paste(..., sep="")
Holger Brandl's avatar
Holger Brandl committed
130 131 132 133 134

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

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

Holger Brandl's avatar
Holger Brandl committed
135 136
as.df <- function(dt) as.data.frame(dt)

Holger Brandl's avatar
Holger Brandl committed
137

Holger Brandl's avatar
Holger Brandl committed
138 139 140 141 142 143 144 145 146 147
########################################################################################################################
#### 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
148 149 150
#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
151 152

rownames2column <- function(df, colname){
153 154
    warning("DEPRECATED Use dplyr::add_rownames directly")
    add_rownames(df, var=colname)
Holger Brandl's avatar
Holger Brandl committed
155 156 157 158
}


column2rownames<- function(df, colname){
159
    #browser()
Holger Brandl's avatar
Holger Brandl committed
160 161 162
    ## force into df to avoid dplyr problems
    df <- as.df(df)

Holger Brandl's avatar
Holger Brandl committed
163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178
    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
179 180
#http://astrostatistics.psu.edu/datasets/R/html/base/html/formals.html
set_names <- function(df, ...){
181
    #browser()
Holger Brandl's avatar
Holger Brandl committed
182
    newnames <- as.character(unlist(list(...)))
183 184 185 186

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

187
    names(df) <- newnames;
Holger Brandl's avatar
Holger Brandl committed
188 189
    return(df)
}
Holger Brandl's avatar
Holger Brandl committed
190 191
#iris %>% set_names(c("setosa", "hallo")) %>% head
#iris %>% set_names("setosa", "hallo") %>% head
Holger Brandl's avatar
Holger Brandl committed
192 193


194 195 196 197 198 199 200
rify_names  <- function(df){
    warning("DEPRECATED: use pretty_columns instead");
    names(df) <- names(df) %>% str_replace_all(c(" "="_", "-"="_"));
    df
}

pretty_columns  <- function(df){ names(df) <- names(df) %>% str_replace_all(c(" "="_", "-"="_", "[.]"="_")) %>% tolower; df}
Holger Brandl's avatar
Holger Brandl committed
201 202


Holger Brandl's avatar
Holger Brandl committed
203 204
head_html <- function(df, n=5) head(df, n) %>% knitr::kable(format="html") %>% print()

Holger Brandl's avatar
Holger Brandl committed
205 206
print_head <- function(df, desc=NULL){
    print(head(df))
Holger Brandl's avatar
Holger Brandl committed
207
    print(nrow(df))
Holger Brandl's avatar
Holger Brandl committed
208 209 210 211 212
    return(df)
}


fac2char <- function(mydata, convert=names(mydata)[sapply(mydata, is.factor)]){
Holger Brandl's avatar
Holger Brandl committed
213 214 215
    if(length(convert)==0){
        return(mydata)
    }
Holger Brandl's avatar
Holger Brandl committed
216 217 218 219 220 221 222 223 224 225 226 227 228

    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
229
## replace R within pipe change just use ... %>% do(replaceNA(0)) %>% ...
Holger Brandl's avatar
Holger Brandl committed
230
replaceNA <- function(x, withValue) { x[is.na(x)] <- withValue; x }
Holger Brandl's avatar
Holger Brandl committed
231

Holger Brandl's avatar
Holger Brandl committed
232

Holger Brandl's avatar
Holger Brandl committed
233

Holger Brandl's avatar
Holger Brandl committed
234
## todo still needed
Holger Brandl's avatar
Holger Brandl committed
235
safe_ifelse <- function(cond, yes, no) {
Holger Brandl's avatar
Holger Brandl committed
236 237 238 239 240 241
#browser()
    isfacOrChar <- function(x) class(x)  %in% c("factor", "character")

    if (isfacOrChar(yes) | isfacOrChar(no)) {
        yes <- ac(yes)
        no <- ac(no)
242
    }
Holger Brandl's avatar
Holger Brandl committed
243 244

    ifelse(cond,yes,no)
Holger Brandl's avatar
Holger Brandl committed
245 246
}

Holger Brandl's avatar
Holger Brandl committed
247

Holger Brandl's avatar
Holger Brandl committed
248
## for na instead use mutate_each with:
Holger Brandl's avatar
Holger Brandl committed
249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276
#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
277 278 279
## see http://stackoverflow.com/questions/24172111/change-the-blank-cells-to-na/33952598#33952598


280 281 282 283 284 285 286
reload_dplyr <- function(){
    unloadNamespace('tidyr')
    unloadNamespace('dplyr')
    require(tidyr);require(dplyr)
}


Holger Brandl's avatar
Holger Brandl committed
287 288


Holger Brandl's avatar
Holger Brandl committed
289 290 291 292
## workaround for biomart
## Deprecated: load dplyr after biomart to avoid this problem
#dselect <- function(...) dplyr::select(...)

Holger Brandl's avatar
Holger Brandl committed
293

Holger Brandl's avatar
Holger Brandl committed
294
########################################################################################################################
Holger Brandl's avatar
Holger Brandl committed
295
#### Result Caching for long running tasks
Holger Brandl's avatar
Holger Brandl committed
296

Holger Brandl's avatar
Holger Brandl committed
297 298
## related: http://cran.r-project.org/web/packages/R.cache/R.cache.pdf

299
cache_it <- function(expr, filePrefix="cache"){
300
    cacheFile <- paste0(filePrefix, "_", substr(digest::digest(deparse(expr)), 1,6)) %>% paste0(".", ., ".RData")
Holger Brandl's avatar
Holger Brandl committed
301

Holger Brandl's avatar
Holger Brandl committed
302 303 304 305 306 307 308 309 310 311 312
    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
313 314
#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
315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332

########################################################################################################################
#### 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
333
## 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
334 335 336 337 338 339 340 341 342 343
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
344 345
## Deprecated: use trim_ext instead
trimEnd <- function(fileNames, ...) trim_ext(fileNames, ...)
Holger Brandl's avatar
Holger Brandl committed
346

Holger Brandl's avatar
Holger Brandl committed
347
trim_ext <-function(fileNames, exts=c()){
Holger Brandl's avatar
Holger Brandl committed
348 349 350 351 352 353 354
    for(fileExt in exts){
        fileNames <- str_replace(fileNames, paste(fileExt, "$",sep=""), "")
    }

    fileNames
}

Holger Brandl's avatar
Holger Brandl committed
355
## DEPRECATED Use write_tsv instead
Holger Brandl's avatar
Holger Brandl committed
356 357
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
358 359 360 361 362 363 364 365 366 367
}

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

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


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

Holger Brandl's avatar
Holger Brandl committed
368

Holger Brandl's avatar
Holger Brandl committed
369
# improved list of objects
Holger Brandl's avatar
Holger Brandl committed
370 371
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
372
    names <- ls(pos = pos, pattern = pattern)
Holger Brandl's avatar
Holger Brandl committed
373

Holger Brandl's avatar
Holger Brandl committed
374 375 376 377 378
    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
379

380
    as.numeric(dim(x))[1:2]))
Holger Brandl's avatar
Holger Brandl committed
381 382
    vec <- is.na(obj.dim)[, 1] & (obj.type != "function")
    obj.dim[vec, 1] <- napply(names, length)[vec]
Holger Brandl's avatar
Holger Brandl committed
383

Holger Brandl's avatar
Holger Brandl committed
384 385
    out <- data.frame(obj.type, obj.size, obj.dim)
    names(out) <- c("Type", "Size", "Rows", "Columns")
Holger Brandl's avatar
Holger Brandl committed
386

Holger Brandl's avatar
Holger Brandl committed
387
    if (!missing(order.by))
388
    out <- out[order(out[[order.by]], decreasing=decreasing), ]
Holger Brandl's avatar
Holger Brandl committed
389 390 391

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

Holger Brandl's avatar
Holger Brandl committed
392 393 394 395 396
    out <- transform(out, var_name=rownames(out))
    rownames(out) <- NULL
    arrange(out, Size)
}

Holger Brandl's avatar
Holger Brandl committed
397
# shorthand that just shows top 1 results
Holger Brandl's avatar
Holger Brandl committed
398 399 400 401
lsosh <- function(..., n=10) {
    lsos(..., order.by="Size", decreasing=TRUE, head=TRUE, n=n)
}

Holger Brandl's avatar
Holger Brandl committed
402 403 404 405 406 407 408 409 410 411
########################################################################################################################
### 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
412
se <- function(x) sd(x, na.rm=TRUE) / sqrt(sum(!is.na(x)))
413 414 415 416 417 418 419 420 421


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

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