core_commons.R 13.2 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) as.data.frame(dt)

Holger Brandl's avatar
Holger Brandl committed
126

Holger Brandl's avatar
Holger Brandl committed
127 128 129 130 131
install_package("tibble")

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

Holger Brandl's avatar
Holger Brandl committed
132 133 134 135 136 137 138 139 140 141
########################################################################################################################
#### 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
142 143 144
#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
145 146

rownames2column <- function(df, colname){
147 148
    warning("DEPRECATED Use dplyr::add_rownames directly")
    add_rownames(df, var=colname)
Holger Brandl's avatar
Holger Brandl committed
149 150 151 152
}


column2rownames<- function(df, colname){
153
    #browser()
Holger Brandl's avatar
Holger Brandl committed
154 155 156
    ## force into df to avoid dplyr problems
    df <- as.df(df)

Holger Brandl's avatar
Holger Brandl committed
157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172
    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
173 174
#http://astrostatistics.psu.edu/datasets/R/html/base/html/formals.html
set_names <- function(df, ...){
175
    #browser()
Holger Brandl's avatar
Holger Brandl committed
176
    newnames <- as.character(unlist(list(...)))
177 178 179 180

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

181
    names(df) <- newnames;
Holger Brandl's avatar
Holger Brandl committed
182 183
    return(df)
}
Holger Brandl's avatar
Holger Brandl committed
184 185
#iris %>% set_names(c("setosa", "hallo")) %>% head
#iris %>% set_names("setosa", "hallo") %>% head
Holger Brandl's avatar
Holger Brandl committed
186 187


188 189 190 191 192 193
rify_names  <- function(df){
    warning("DEPRECATED: use pretty_columns instead");
    names(df) <- names(df) %>% str_replace_all(c(" "="_", "-"="_"));
    df
}

194 195 196 197 198 199 200 201
pretty_columns  <- function(df){
    names(df) <- names(df) %>%
        str_replace_all("[#=.()-]+", "_") %>%
        str_replace("[_]+$", "") %>%
        str_replace("^[_]+", "") %>%
        tolower;
    df
}
Holger Brandl's avatar
Holger Brandl committed
202

203 204
# 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
205

Holger Brandl's avatar
Holger Brandl committed
206 207
head_html <- function(df, n=5) head(df, n) %>% knitr::kable(format="html") %>% print()

Holger Brandl's avatar
Holger Brandl committed
208 209
print_head <- function(df, desc=NULL){
    print(head(df))
Holger Brandl's avatar
Holger Brandl committed
210
    print(nrow(df))
Holger Brandl's avatar
Holger Brandl committed
211 212 213 214 215
    return(df)
}


fac2char <- function(mydata, convert=names(mydata)[sapply(mydata, is.factor)]){
Holger Brandl's avatar
Holger Brandl committed
216 217 218
    if(length(convert)==0){
        return(mydata)
    }
Holger Brandl's avatar
Holger Brandl committed
219 220 221 222 223 224 225 226 227 228 229 230 231

    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
232
## replace R within pipe change just use ... %>% do(replaceNA(0)) %>% ...
Holger Brandl's avatar
Holger Brandl committed
233
replaceNA <- function(x, withValue) { x[is.na(x)] <- withValue; x }
Holger Brandl's avatar
Holger Brandl committed
234

Holger Brandl's avatar
Holger Brandl committed
235

Holger Brandl's avatar
Holger Brandl committed
236

Holger Brandl's avatar
Holger Brandl committed
237
## todo still needed since there's if_else in dplyr now
Holger Brandl's avatar
Holger Brandl committed
238
safe_ifelse <- function(cond, yes, no) {
Holger Brandl's avatar
Holger Brandl committed
239

Holger Brandl's avatar
Holger Brandl committed
240 241 242 243 244 245
#browser()
    isfacOrChar <- function(x) class(x)  %in% c("factor", "character")

    if (isfacOrChar(yes) | isfacOrChar(no)) {
        yes <- ac(yes)
        no <- ac(no)
246
    }
Holger Brandl's avatar
Holger Brandl committed
247 248

    ifelse(cond,yes,no)
Holger Brandl's avatar
Holger Brandl committed
249 250
}

Holger Brandl's avatar
Holger Brandl committed
251

Holger Brandl's avatar
Holger Brandl committed
252
## for na instead use mutate_each with:
Holger Brandl's avatar
Holger Brandl committed
253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280
#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
281 282 283
## see http://stackoverflow.com/questions/24172111/change-the-blank-cells-to-na/33952598#33952598


Holger Brandl's avatar
Holger Brandl committed
284 285 286 287 288 289 290
## 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
291 292
distinct_all = function (x, ...) distinct(..., .keep_all=F)

293 294 295 296 297 298 299
reload_dplyr <- function(){
    unloadNamespace('tidyr')
    unloadNamespace('dplyr')
    require(tidyr);require(dplyr)
}


300
## from http://stackoverflow.com/questions/7505547/detach-all-packages-while-working-in-r
301
unload_packages <- function() {
302 303 304 305 306
    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
307 308


Holger Brandl's avatar
Holger Brandl committed
309 310 311 312
## workaround for biomart
## Deprecated: load dplyr after biomart to avoid this problem
#dselect <- function(...) dplyr::select(...)

Holger Brandl's avatar
Holger Brandl committed
313

Holger Brandl's avatar
Holger Brandl committed
314
########################################################################################################################
Holger Brandl's avatar
Holger Brandl committed
315
#### Result Caching for long running tasks
Holger Brandl's avatar
Holger Brandl committed
316

Holger Brandl's avatar
Holger Brandl committed
317 318
## related: http://cran.r-project.org/web/packages/R.cache/R.cache.pdf

319
cache_it <- function(expr, filePrefix="cache"){
320
    cacheFile <- paste0(filePrefix, "_", substr(digest::digest(deparse(expr)), 1,6)) %>% paste0(".", ., ".RData")
Holger Brandl's avatar
Holger Brandl committed
321

Holger Brandl's avatar
Holger Brandl committed
322 323 324 325 326 327 328 329 330 331 332
    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
333 334
#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
335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352

########################################################################################################################
#### 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
353
## 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
354 355 356 357 358 359 360 361 362 363
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
364 365
## Deprecated: use trim_ext instead
trimEnd <- function(fileNames, ...) trim_ext(fileNames, ...)
Holger Brandl's avatar
Holger Brandl committed
366

Holger Brandl's avatar
Holger Brandl committed
367
trim_ext <-function(fileNames, exts=c()){
Holger Brandl's avatar
Holger Brandl committed
368 369 370 371 372 373 374
    for(fileExt in exts){
        fileNames <- str_replace(fileNames, paste(fileExt, "$",sep=""), "")
    }

    fileNames
}

Holger Brandl's avatar
Holger Brandl committed
375
## DEPRECATED Use write_tsv instead
Holger Brandl's avatar
Holger Brandl committed
376 377
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
378 379 380 381 382 383 384 385 386 387
}

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

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


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

Holger Brandl's avatar
Holger Brandl committed
388

Holger Brandl's avatar
Holger Brandl committed
389
# improved list of objects
Holger Brandl's avatar
Holger Brandl committed
390 391
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
392
    names <- ls(pos = pos, pattern = pattern)
Holger Brandl's avatar
Holger Brandl committed
393

Holger Brandl's avatar
Holger Brandl committed
394 395 396 397 398
    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
399

400
    as.numeric(dim(x))[1:2]))
Holger Brandl's avatar
Holger Brandl committed
401 402
    vec <- is.na(obj.dim)[, 1] & (obj.type != "function")
    obj.dim[vec, 1] <- napply(names, length)[vec]
Holger Brandl's avatar
Holger Brandl committed
403

Holger Brandl's avatar
Holger Brandl committed
404 405
    out <- data.frame(obj.type, obj.size, obj.dim)
    names(out) <- c("Type", "Size", "Rows", "Columns")
Holger Brandl's avatar
Holger Brandl committed
406

Holger Brandl's avatar
Holger Brandl committed
407
    if (!missing(order.by))
408
    out <- out[order(out[[order.by]], decreasing=decreasing), ]
Holger Brandl's avatar
Holger Brandl committed
409 410 411

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

Holger Brandl's avatar
Holger Brandl committed
412 413 414 415 416
    out <- transform(out, var_name=rownames(out))
    rownames(out) <- NULL
    arrange(out, Size)
}

Holger Brandl's avatar
Holger Brandl committed
417
# shorthand that just shows top 1 results
Holger Brandl's avatar
Holger Brandl committed
418 419 420 421
lsosh <- function(..., n=10) {
    lsos(..., order.by="Size", decreasing=TRUE, head=TRUE, n=n)
}

Holger Brandl's avatar
Holger Brandl committed
422 423 424 425 426 427 428 429 430 431
########################################################################################################################
### 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
432
se <- function(x) sd(x, na.rm=TRUE) / sqrt(sum(!is.na(x)))
433 434 435 436 437 438 439 440


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

## inspired by http://stackoverflow.com/questions/8343509/better-error-message-for-stopifnot
assert <- function (expr, error) {
    if (! expr) stop(error, call. = FALSE)
441 442 443
}

### table rendering
Holger Brandl's avatar
Holger Brandl committed
444 445
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,...)
446
}