README.md

In petarmimica/estatusbar: Register Task Progress, Display Statusbar, Estimate Time of Completion

estatusbar

Petar Mimica

Register Task Progress, Display Statusbar, Estimate Time of Completion

Introduction

This is an R port of a Fortran module that was used in the astrophysical codes SPEV and MRGENESIS. You can register the task completion fraction as it progresses and display a progress/status bar. The package uses several algorithms to estimate the time of completion.

The best way to install this package directly from my GitHub page:

devtools::install_github("petarmimica/estatusbar")

Basic usage

The idea is that you start performing some task, and then periodically add entries about the fraction of that task that has been completed. estatusbar will automatically add time stamps.

First, an estatusbar object is created:

est <- estatusbar$new()

To add an entry after 15 percent of the task has been completed you can use:

est$add(0.15)

estatusbar also allows you to display the current status and also to estimate when the task will be complete.

est$display()

##

|=====>                                  | (15%) ETA: 2017-10-25 12:51:48

The display routine is configurable. The width can be adjusted:

est$display(width=20)

##

|==>                 | (15%) ETA: 2017-10-25 12:51:48

Also, a custom text string can be provided:

est$display(text = "Working...")

##

|=====>                                  | Working... (15%) ETA: 2017-10-25 12:51:48

The percentage can be hidden, as well as the ETA:

est$display(perc = FALSE)

##

|=====>                                  | ETA: 2017-10-25 12:51:48

est$display(eta = FALSE)

##

|=====>                                  | (15%)

Examples

Example 1

In the first example, the status bar is registered, and then 20 entries are added. Before each add the system sleeps for a random time interval (between 0 and 2 seconds). Afterward an entry is registered and the display is updated.

est <- estatusbar$new()
for (i in 1:20) {
   Sys.sleep(2 * runif(1))
   est$add(i / 20)
   est$display()
}

##

|=>                                      | (05%) ETA: 2017-10-25 12:51:49

|===>                                    | (10%) ETA: 2017-10-25 12:53:02

|=====>                                  | (15%) ETA: 2017-10-25 12:54:27

|=======>                                | (20%) ETA: 2017-10-25 12:52:24

|=========>                              | (25%) ETA: 2017-10-25 12:52:22

|===========>                            | (30%) ETA: 2017-10-25 12:52:15

|=============>                          | (35%) ETA: 2017-10-25 12:52:13

|===============>                        | (40%) ETA: 2017-10-25 12:52:14

|=================>                      | (45%) ETA: 2017-10-25 12:52:15

|===================>                    | (50%) ETA: 2017-10-25 12:52:14

|=====================>                  | (55%) ETA: 2017-10-25 12:52:14

|=======================>                | (60%) ETA: 2017-10-25 12:52:14

|=========================>              | (65%) ETA: 2017-10-25 12:52:13

|===========================>            | (70%) ETA: 2017-10-25 12:52:13

|=============================>          | (75%) ETA: 2017-10-25 12:52:12

|===============================>        | (80%) ETA: 2017-10-25 12:52:11

|=================================>      | (85%) ETA: 2017-10-25 12:52:10

|===================================>    | (90%) ETA: 2017-10-25 12:52:09

|=====================================>  | (95%) ETA: 2017-10-25 12:52:09

|=======================================>| (100%) ETA: 2017-10-25 12:52:09

Sys.time()

## [1] "2017-10-25 12:52:09 CEST"

If this were run in terminal, the statusbar would be animated. As can be seen, the time estimate varies considerable at the early times, but after 55% of the task has been completed it is quite close to the final value, shown by calling Sys.time().

Example 2

In this example the user supplied text is used instead of the percentage display to show progress.

est <- estatusbar$new()
for (i in 1:10) {
   Sys.sleep(5 * runif(1))
   est$add(i / 10)
   my.text <- paste0(formatC(i, width=2, flag="0"), "/", "10", collapse="")
   est$display(text=my.text, perc=FALSE)
}

##

|===>                                    | 01/10 ETA: 2017-10-25 12:52:09

|=======>                                | 02/10 ETA: 2017-10-25 12:54:16

|===========>                            | 03/10 ETA: 2017-10-25 12:52:46

|===============>                        | 04/10 ETA: 2017-10-25 12:52:37

|===================>                    | 05/10 ETA: 2017-10-25 12:52:34

|=======================>                | 06/10 ETA: 2017-10-25 12:52:34

|===========================>            | 07/10 ETA: 2017-10-25 12:52:32

|===============================>        | 08/10 ETA: 2017-10-25 12:52:31

|===================================>    | 09/10 ETA: 2017-10-25 12:52:31

|=======================================>| 10/10 ETA: 2017-10-25 12:52:33

Sys.time()

## [1] "2017-10-25 12:52:33 CEST"

Estimators

Currently only two algorithms are implemented.

First-last

The first algorithm takes the last known entry and uses its time T and its fraction f to estimate the total duration (T(1)), assuming the following model: T(f)=T(1)*fα It is easy to compute T(1) (or any T) from the preceding equation. In the current implementation α takes values 1 and 2/

Linear fit

The second algorithm performs a linear fit on the last n points and uses that model to predict the total duration or any other desired value: T(f)=A + B * f Currently n is either equal to 5 or to the total number of points so far.

This means that currently we are using 4 estimators for the total time.

Voting

When a new entry is added, the prediction of each estimator for that entry's time is recored. This information is then used when estimating the total time. Assuming that so far n entries have been recorded, for each estimator j we define its weight as: $$ w_j = 1 / \sum_{i=1}^n (T_i - P_j(f_i) )^2 $$ where Pj(fi) is the model prediction of the estimator j for the fracton fi. The final prediction is computed as: $$ T_{pred} = (\sum_{j=1}^{N_{est}} w_j P_j(1)) / \sum_{j=1}^{N_{est}} w_j $$ where Pj(1) is the current prediction for the total time of the estimator j. Nest is the total number of estimators.

petarmimica/estatusbar documentation built on May 29, 2019, 7:36 a.m.