#!/usr/bin/perl -w
#	summarize_epics.perl [filename]
#
# 	by: Jaideep Singh
#
#	date: 01/26/2006 
#
#	about:	reads a raw epics data text file and create a summary file
#		creates a file:
#				(0) first line is the prescalar string
#				(1) afterwards, each line is an epics variable
#				(2) if the variable is numeric, then the line will be:
#
#					variable name <tab> number of values <tab> minimum <tab> maximum <tab> mean <tab> standard deviation
#			        (3) the numeric variables come at the beginning of the file and are separated from the alphanumeric variables by a "-----" line
#				(4) if the variable is alphanumeric, then the line will be:
#
#					variable name <tab> number of values <tab> first value <tab> number of occurences <tab> second value <tab> number of occurences ...
#
#					Note that some EPICS-like stuff still leaks through...
#
#	input file is of the format: 
#		first line might be prescalar string
#		all subsequent lines:
#					timestamp<tab>seconds from 0000 Jan 1 2003<tab>variable name<tab>varicable value
#
#
#	EPICS info is injected into the data stream at certain times invertals.
#	More important stuff is logged more often.
#	Immediately preceding one log cycle, a timestamp is grabbed
#	Each variable is recorded in ASCII format on its own line: 
#
#	[variable name, upto 30 characters][at least one blank space][upto 20 characters for value] 
#	
#	for example:
#
#	IPM1H04B.XPOS                  -0.0408508
#	000000000111111111122222222223333333333444444444455
#	123456789012345678901234567890123456789012345678901
#
#	Variable names can have ".", ":","+","_" as well as alphanumeric characters

use strict "subs";

#getting name of file to search + ARGUMENT ERROR HANDLING
$filename = $ARGV[0];
die "Uh, what file do you want me to look at dude?: Too few arguments\n" unless ($filename);
die "Scared and confused: Too many arguments\n" unless (($#ARGV+1) < 2);
open(FILE,$filename) or die "d'oh!: - $filename - not found\n";
#print "\nSearching raw data file:\t$filename\n";

#initializing variables
$line_count = 0;

LINENUM: while(<FILE>)
{
 $line_count++;
 chomp;
 if (/(ps\d\=\d+\,ps\d\=\d+\,ps\d\=\d+\,ps\d\=\d+\,ps\d\=\d+\,ps\d\=\d+\,ps\d\=\d+\,ps\d\=\d+)/)
 {
	$prescalars = $1;
 }
 elsif ($_ ne "")
 {
	@line_parts = split(/\t/);
	if ($line_parts[3] =~ /([0123456789eE\.\+\-]+)/) { $poop = $1; if ($line_parts[3] eq $1) { $type = "num" } else { $type = "str" } } else {$poop = ""; $type = "str"}
#	$poop is used as just a diagnostic variable
#	print "\*$line_parts[3]\*\t\*$poop\*\t$type\n";
	if ($type eq "num")
	{
	   if (not $data_num{$line_parts[2]}{"count"})
	   {
		$data_num{$line_parts[2]}{"min"} = $line_parts[3];
		$data_num{$line_parts[2]}{"max"} = $line_parts[3];
		$data_num{$line_parts[2]}{"count"} = 1.0;
		$data_num{$line_parts[2]}{"sum"} = $line_parts[3];
		$data_num{$line_parts[2]}{"sum_sq"} = $line_parts[3]*$line_parts[3];
		$data_num{$line_parts[2]}{"mean"} = $data_num{$line_parts[2]}{"sum"}/$data_num{$line_parts[2]}{"count"};
		$data_num{$line_parts[2]}{"std"} = sqrt(abs($data_num{$line_parts[2]}{"sum_sq"}/$data_num{$line_parts[2]}{"count"}-$data_num{$line_parts[2]}{"mean"}*$data_num{$line_parts[2]}{"mean"}));
	   }
	   else
	   {
		$data_num{$line_parts[2]}{"min"} = $line_parts[3] if ($data_num{$line_parts[2]}{"min"} > $line_parts[3]);
		$data_num{$line_parts[2]}{"max"} = $line_parts[3] if ($data_num{$line_parts[2]}{"max"} < $line_parts[3]);
		$data_num{$line_parts[2]}{"count"}++;
		$data_num{$line_parts[2]}{"sum"} = $data_num{$line_parts[2]}{"sum"} + $line_parts[3];
		$data_num{$line_parts[2]}{"sum_sq"} = $data_num{$line_parts[2]}{"sum_sq"} + $line_parts[3]*$line_parts[3];
		$data_num{$line_parts[2]}{"mean"} = $data_num{$line_parts[2]}{"sum"}/$data_num{$line_parts[2]}{"count"};
		$data_num{$line_parts[2]}{"std"} = sqrt(abs($data_num{$line_parts[2]}{"sum_sq"}/$data_num{$line_parts[2]}{"count"}-$data_num{$line_parts[2]}{"mean"}*$data_num{$line_parts[2]}{"mean"}));
		$data_num{$line_parts[2]}{"std"} = 0.0 if ($data_num{$line_parts[2]}{"max"} == $data_num{$line_parts[2]}{"min"});
	   }
	}
	else { if (not $data_str{$line_parts[2]}{$line_parts[3]}) {$data_str{$line_parts[2]}{$line_parts[3]} = 1.0} else {$data_str{$line_parts[2]}{$line_parts[3]}++} }
 } 
};
close(FILE);
#setting up output file
@path_name_parts = split(/\//,$filename);
@file_name_parts = split(/\./,$path_name_parts[$#path_name_parts]);
$file_out = $file_name_parts[0]."_summary.txt";
#print "OUTPUT filename : $file_out\n\n";
open(OUTPUT,">$file_out");
print OUTPUT "$prescalars\n" if ($prescalars);
foreach $variable_name (sort keys %data_num) 
{ 
	$count = $data_num{$variable_name}{"count"};
	  $min = sprintf( "%.9E" , $data_num{$variable_name}{"min"}   );
	  $max = sprintf( "%.9E" , $data_num{$variable_name}{"max"}   );
	 $mean = sprintf( "%.9E" , $data_num{$variable_name}{"mean"}  );
	  $std = sprintf( "%.9E" , $data_num{$variable_name}{"std"}   );
	print OUTPUT "$variable_name\t$count\t$min\t$max\t$mean\t$std\n"
}
print OUTPUT "-----\n";
# because %data_str is a hash of hashes, $data_str{$variable_name} is a reference to a hash
# therefore $temp is a reference to a hash
#       and %$temp is the hash
#       and $$temp{$variable_value} is the value corresponding to the $variable_value key
foreach $variable_name (sort keys %data_str) 
{
	$temp = $data_str{$variable_name};
	$num_values = scalar keys %$temp;
	print OUTPUT "$variable_name\t$num_values";
	foreach $variable_value (sort keys %$temp) { print OUTPUT "\t$variable_value\t$$temp{$variable_value}"} 
	print OUTPUT "\n";
}
close(OUTPUT);