Package 'profr'

Visualise profiling data with ggplot2. Visualise profiling data stored in a profr data.frame.

Description

This will plot the call tree of the specified stop watch object. If you only want a small part, you will need to subset the object

Usage

ggplot.profr(data, ..., minlabel = 0.1, angle = 0)

Arguments

data	profile output to plot
...	other arguments passed on to ggplot
minlabel	minimum percent of time for function to get a label
angle	function label angle

See Also

plot.profr

Examples

if (require("ggplot2")) {
  ggplot(nesting_prof)
  ggplot(reshape_prof)
}

---

Parse Rprof output.

Description

Parses the output of Rprof into an alternative format described in profr. This produces a flat data frame, which is somewhat easier to summarise and visualise.

Usage

parse_rprof(path, interval = 0.02)

Arguments

path	path to Rprof output
interval	real-time interval between samples (in seconds)

Value

data.frame of class profr

See Also

profr for profiling and parsing

Examples

nesting_ex <- system.file("samples", "nesting.rprof", package="profr")
nesting <- parse_rprof(nesting_ex)

reshape_ex <- system.file("samples", "reshape.rprof", package="profr")
diamonds <- parse_rprof(reshape_ex)

---

Visualise profiling data with base graphics. Visualise profiling data stored in a profr data.frame.

Description

If you only want a small part of the total call tree, you will need to subset the object as demonstrated by the example.

Usage

## S3 method for class 'profr'
plot(x, ..., minlabel = 0.1, angle = 0)

Arguments

x	profile output to plot
...	other arguments passed on to plot.default
minlabel	minimum percent of time for function to get a label
angle	function label angle

See Also

ggplot.profr

Examples

plot(nesting_prof)
plot(reshape_prof)

---

Profile the performance of a function call.

Description

This is a wrapper around Rprof that provides results in an alternative data structure, a data.frame. The columns of the data.frame are:

Usage

profr(expr, interval = 0.02, quiet = TRUE)

Arguments

expr	expression to profile
interval	interval between samples (in seconds)
quiet	should output be discarded?

Details

f

name of function

level

level in call stack

time

total time (seconds) spent in function

start

time at which control entered function

end

time at which control exited function

leaf

TRUE if the function is a terminal node in the call tree, i.e. didn't call any other functions

source

guess at the package that the function came from

Value

data.frame of class profr

See Also

parse_rprof to parse standalone Rprof file, plot.profr and ggplot.profr to visualise the profiling data

Examples

## Not run: 
glm_ex <- profr({Sys.sleep(1); example(glm)}, 0.01)
head(glm_ex)
summary(glm_ex)
plot(glm_ex)

## End(Not run)

---

Sample profiling datasets

Description

These two datasets illustrate the results of running parse_rprof on the sample Rprof output stored in the samples directory. The output was generated by the code in samples/generate.r.

Usage

nesting_prof

reshape_prof

Format

a data frame

---