status

#! /bin/sh
#
# Show the currently running KONECT-Analysis processes.
# This calls ps(1) and presents the results in a useful way.   
#
# All shown runtimes are CPU times, not wall clock times. 
#
# MECHANISM 
#
# To show the time left for a task, a program outputs, at regular
# intervals, strings of the form
#
#     [TIMESPEC left]
#
# where TIMESPEC is the time left in free form.  The string does not
# need to be alone on a line.
#
# Additionally, each program that should be inspected in this way should
# be started with a logfile name as (any) positional argument, which
# must start with the value of $TMPDIR.  This is done automatically by
# the wrappers {octave,matlab,julia}, but must be implemented by hand
# with other languages.  This is the reason why certain C or Perl
# programs are passed a logfile name, even if they do not (yet) make use
# of it.  In principle however, every program that has a long runtime
# should log such [... left] message to its logfile.
#
# Logfiles names have the general format
#
#    $TMPDIR/$LETTER.$PROGRAM[.$PARAMETERS].log
#
# where $LETTER is a single letter denoting the programming language
# (usually the programming language's name first letter), $PROGRAM is
# the general name of the program being run, and [.$PARAMETERS] are
# optional parameters (such as the network name, etc.)
#
# BUGS
#
#   * We don't manage the situation when the list is longer than the screen. 
#     Currently, we just make the window higher in that case.
#   * The shown CPU percentage is as returned by ps(1), which gives the
#     average CPU usage since the process was started.  This is
#     different (and less useful) to what top(1) is doing, which shows
#     the CPU usage since the last screen update. 
#

# Show all running Stu instances
# Note: ps -C is not POSIX
ps -C stu -o user,pid,pcpu,etime,cputime,command
printf '%s\n' -

# Format for the printf(1) command
format='%4s %13s %12s %13s %s\n'

# Header line
printf "$format" "%CPU" "MEM [KiB]" "TIME" "LEFT" "LOG"

TMPDIR=${TMPDIR-/tmp}
TMPDIR_esc=$(echo "$TMPDIR" | sed -E -e 's,/,\\/,g')

ps -A k-cputime -o pcpu,cputime,vsize,user,command | 
sed -E -e '1d;/'"$TMPDIR_esc"'/!d' |
while IFS= read -r line
do
	cpu=$(echo "$line" | sed -E -e 's,^\s*(\S+)\s.*$,\1,;s,\.[0-9]$,,')
	vsize=$(echo "$line" | sed -E -e 's,^\s*\S+\s+\S+\s+(\S+)\s.*$,\1,' | sed -E -e '
s/^([^.]*[0-9])([0-9]{21}(\.|$))/\1,\2/
s/^([^.]*[0-9])([0-9]{18}(\.|$))/\1,\2/
s/^([^.]*[0-9])([0-9]{15}(\.|$))/\1,\2/
s/^([^.]*[0-9])([0-9]{12}(\.|$))/\1,\2/
s/^([^.]*[0-9])([0-9]{9}(\.|$))/\1,\2/
s/^([^.]*[0-9])([0-9]{6}(\.|$))/\1,\2/
s/^([^.]*[0-9])([0-9]{3}(\.|$))/\1,\2/
')
	time=$(echo "$line" | sed -E -e 's,^\s*,,' | cut -f 2 -d ' ' | sed -E -e 's,^[0:]*(.*....)$,\1,')
	logfile=$(echo "$line" | sed -E -e 's,^.*\s('"$TMPDIR_esc"'\S+).*$,\1,')

	command=$(echo "$line" | sed -E -e 's,^\s*\S+\s+\S+\s+\S+\s+\S+\s+(\S+)\s.*$,\1,')
	command=$(basename "$command")

	# Don't show clearly non-computational processes.  For instance,
	# 'less' is often used to look at logfiles, and would thus show
	# up here because it has a logfile name as a positional argument 
	if [ "$command" = less ] || [ "$command" = more ] || [ "$command" = tail ] || [ "$command" = sed ] ; then
		continue
	fi
	if printf "%s\n" "$command" | grep -E -q ^main.stu ; then
		continue
	fi
	# This refers to stu-utils/output-finally 
	if printf "%s\n" "$line" | grep -E -q 'output-finally' ;
	then
		continue
	fi
	
	# 'tac' reports an error if the logfile was not (yet)
	# created.  Ignore this error.  
	left=$(tac "$logfile" 2>/dev/null | sed -E -e '/left]/!d;s,^.*\[(.*) left\].*$,\1,;q')
	if [ -z "$left" ] ; then
		# If there is a "time left" information available
		# for a sub-task, show that prefixed with '+'. 
		left=$(tac "$logfile" 2>/dev/null | sed -E -e '/left}/!d;s,^.*\{(.*) left\}.*$,+\1,;q')
	fi

	logfile_color=$(echo "$logfile" | sed -E -e 's,^(.*)/([^.]+)\.(.+).log$,\1/\2.\3.log,')
	
	printf "$format" "$cpu" "$vsize" "$time" "$left" "$logfile_color"
done				

printf '%s\n' -