$opt{anchor} = qr/$_/;
} or die $@ =~ s/(?:\ at\ \N+)?\Z/ for option $_[0]/r;
},
+ 'count|c!',
'human-readable|H!',
'sexagesimal!',
'reformat!',
202 208 214 220 226 227 228 229 230 231 159
)],
whites => [qw( 1;30 0;37 1;37 )],
- greys => [map {"38;5;$_"} 0, 232..255, 15],
+ grays => [map {"38;5;$_"} 0, 232..255, 15],
random => [map {"38;5;$_"} List::Util::shuffle(17..231)],
rainbow=> [map {"38;5;$_"}
196, # r
$opt{trim} *= $opt{width} / 100 if $opt{trimpct};
$opt{units} = [split //, ' kMGTPEZYRQqryzafpn'.($opt{ascii} ? 'u' : 'μ').'m']
if $opt{'human-readable'};
-$opt{anchor} //= qr/\A/;
$opt{'value-length'} = 4 if $opt{units};
$opt{'value-length'} = 1 if $opt{unmodified};
$opt{'signal-stat'} //= exists $SIG{INFO} ? 'INFO' : 'QUIT';
$opt{markers} //= '=avg >31.73v <68.27v +50v |0';
-$opt{report} //= join(', ',
- '${min; color(31)} min',
- '${avg; $opt{reformat} or $_ = sprintf "%0.2f", $_; color(36)} avg',
- '${max; color(32)} max',
+$opt{report} //= join('',
+ '${partsum+; $_ .= " of "}',
+ '${sum+; color(1); $_ .= " total in "}',
+ '${count#} values',
+ '${lines#; $_ = $_ != @order && " over $_ lines"}',
+ sprintf('${count: (%s)}', join ', ',
+ '${min; color(31)} min',
+ '${avg; $opt{reformat} or $_ = sprintf "%0.2f", $_; color(36)} avg',
+ '${max; color(32)} max',
+ ),
);
$opt{palette} //= $opt{color} && [31, 90, 32];
$opt{indicators} = [split //, $opt{indicators} ||
$opt{'value-format'} ||= sub { sprintf '%.8g', $_[0] };
-my (@lines, @values, @order);
+my (@lines, @values, @order, %uniq);
$SIG{$_} = \&show_stat for $opt{'signal-stat'} || ();
$SIG{ALRM} = sub {
eval {
require Tie::Array::Sorted;
tie @order, 'Tie::Array::Sorted', sub { $_[1] <=> $_[0] };
- } or warn $@, "Expect slowdown with large datasets!\n";
+ } or warn $@, "Expect slowdown with large datasets!\n"
+ unless $opt{count};
}
my $float = qr<[0-9]* [.]? [0-9]+ (?: e[+-]?[0-9]+ )?>; # positive numberish
-my $valmatch = qr< $opt{anchor} ( \h* -? $float |) >x;
+my $valmatch = $opt{anchor} // qr/\A/;
+$valmatch .= !$opt{count} ? qr/( \h* -? $float |)/ :
+ $opt{anchor} ? qr/(\S*)/ : qr/(.*)/;
+
while (defined ($_ = $opt{input} ? shift @{ $opt{input} } : readline)) {
s/\r?\n\z//;
- s/\A\h*// unless $opt{unmodified};
- my $valnum = s/$valmatch/\n/ && $1;
- push @values, $valnum;
- push @order, $valnum if length $valnum;
+ my $valnum;
+ if ($opt{count}) {
+ $valnum = m/$valmatch/ && $1;
+ $uniq{$valnum}++ and next;
+ push @values, $valnum;
+ s/\A/\n /;
+ }
+ else {
+ s/\A\h*// unless $opt{unmodified};
+ $valnum = s/$valmatch/\n/ && $1;
+ push @values, $valnum;
+ push @order, $valnum if length $valnum;
+ }
+
if (defined $opt{trim} and defined $valnum) {
my $trimpos = abs $opt{trim};
$trimpos -= length $valnum if $opt{unmodified};
}
}
push @lines, $_;
+}
+continue {
show_lines() if defined $opt{interval} and $opt{interval} < 0
and $. % $opt{interval} == 0;
}
my $limit = $opt{hidemax} ? $opt{hidemax}->($#lines, $nr) : $#lines;
+if ($opt{count}) {
+ $_ = $uniq{$_} for @values[$nr .. $limit];
+ @order = @values;
+}
+
@order = sort { $b <=> $a } @order unless tied @order;
my $maxval = $opt{maxval} // (
$opt{hidemax} ? max grep { length } @values[$nr .. $limit] :
if ($opt{markers} and $size > 0) {
for my $markspec (split /\h/, $opt{markers}) {
my ($char, $func) = split //, $markspec, 2;
+ my $increment = $func =~ s/[+]\z//;
my @pos = eval {
if ($func eq 'avg') {
return sum(@order) / @order;
$pos -= $minval;
$pos &&= log $pos if $opt{log};
$pos >= 0 or next;
- color(36) for $barmark[$pos / $range * $size] = $char;
+ $increment ||= $minval && !$pos;
+ color(36) for $barmark[$pos / $range * $size + $increment + .5] = $char;
}
}
state $lastmax = $maxval;
if ($maxval > $lastmax) {
print ' ' x ($lenval + $len);
- printf color(90);
+ print color(90);
printf '%-*s',
($lastmax - $minval) * $size / $range + .5,
'-' x (($values[$nr - 1] - $minval) * $size / $range);
$nr++;
}
say $opt{palette} ? color(0) : '' if $opt{spark};
+%uniq = () if $opt{interval} and $opt{count};
return $nr;
}
sub show_stat {
- if ($opt{hidemin} or $opt{hidemax}) {
- my $linemin = $opt{hidemin} ? $opt{hidemin}->($#lines) : 0;
- my $linemax = $opt{hidemax} ? $opt{hidemax}->($#lines, $linemin) : $#lines;
- print varfmt('${sum+} of ', {
- lines => $linemax - $linemin + 1,
- sum => sum(0, grep {length} @values[$linemin .. $linemax]),
- }) if $linemin <= $linemax;
- }
+ my %vars = (
+ count => int @order,
+ lines => int @lines,
+ );
+ my $linemin = !$opt{hidemin} ? 0 :
+ ($vars{start} = $opt{hidemin}->($#lines));
+ my $linemax = !$opt{hidemax} ? $#lines :
+ ($vars{end} = $opt{hidemax}->($#lines, $vars{start}));
if (@order) {
- my $total = sum @order;
- my $fmt = '${sum+;color(1)} total in ${count#} values';
- $fmt .= ' over ${lines#} lines' if @order != @lines;
- $fmt .= " ($_)" for $opt{report} || ();
- print varfmt($fmt, {
- sum => $total,
- count => int @order,
- lines => int @lines,
+ $vars{partsum} = sum(0, grep {length} @values[$linemin .. $linemax])
+ if $linemin <= $linemax and ($opt{hidemin} or $opt{hidemax});
+ %vars = (%vars,
+ sum => sum(@order),
min => $order[-1],
max => $order[0],
- avg => $total / @order,
- });
+ );
+ $vars{avg} = $vars{sum} / @order;
}
- say '';
+ say varfmt($opt{report}, \%vars);
return 1;
}
sub varfmt {
my ($fmt, $vars) = @_;
$fmt =~ s[\$\{ \h*+ ((?: [^{}]++ | \{(?1)\} )+) \}]{
- my ($name, $cmd) = split /\s*;/, $1, 2;
+ my ($name, $op, $cmd) = split /\s*([;:])/, $1, 2;
my $format = $name =~ s/\+// || $name !~ s/\#// && $opt{reformat};
local $_ = $vars->{$name};
- if (defined) {
+ defined && do {
$_ = $opt{'value-format'}->($_) if $format;
- if ($cmd) {
+ if ($cmd and $op eq ':') {
+ $_ = varfmt($cmd, $vars);
+ }
+ elsif ($cmd) {
eval $cmd;
warn "Error in \$$name report: $@" if $@;
}
$_;
}
- else {
- warn "Unknown variable \$$name in report\n";
- "\$$name";
- }
}eg;
return $fmt;
}
Options:
-a, --[no-]ascii Restrict user interface to ASCII characters
-C, --[no-]color Force colored output of values and bar markers
+ -c, --count Omit repetitions and count the number of
+ occurrences
-f, --field=([+]N|REGEXP)
Compare values after a given number of whitespace
separators
=head1 SYNOPSIS
-B<barcat> [<options>] [<file>... | <numbers>]
+B<barcat> [I<options>] [I<file>... | I<numbers>]
=head1 DESCRIPTION
Contents are concatenated similar to I<cat>,
but numbers are reformatted and a bar graph is appended to each line.
-Don't worry, barcat does not drink and divide.
It can has various options for input and output (re)formatting,
but remains limited to one-dimensional charts.
For more complex graphing needs
=over
-=item -a, --[no-]ascii
+=item B<-a>, B<-->[B<no->]B<ascii>
Restrict user interface to ASCII characters,
replacing default UTF-8 by their closest approximation.
Input is always interpreted as UTF-8 and shown as is.
-=item -C, --[no-]color
+=item B<-C>, B<-->[B<no->]B<color>
Force colored output of values and bar markers.
Defaults on if output is a tty,
disabled otherwise such as when piped or redirected.
-Can also be disabled by setting I<-M>
+Can also be disabled by setting B<-M>
or the I<NO_COLOR> environment variable.
-=item -f, --field=([+]<number> | <regexp>)
+=item B<-c>, B<--count>
+
+Omit repetitions and count the number of occurrences.
+Similar to piping input through C<sort | uniq -c>
+but keeping the order of first appearances.
+
+Lines are omitted if they (or a specified field) are identical,
+and the amount of matches is prepended and used as values
+for bars and subsequent statistics.
+
+=item B<-f>, B<--field>=([B<+>]I<number> | I<regexp>)
Compare values after a given number of whitespace separators,
or matching a regular expression.
-Unspecified or I<-f0> means values are at the start of each line.
-With I<-f1> the second word is taken instead.
+Unspecified or B<-f0> means values are at the start of each line.
+With B<-f1> the second word is taken instead.
A string can indicate the starting position of a value
-(such as I<-f:> if preceded by colons),
+(such as B<-f:> if preceded by colons),
or capture the numbers itself,
-for example I<-f'(\d+)'> for the first digits anywhere.
-A shorthand for this is I<+0>, or I<+N> to find the Nth number.
+for example B<-f'(\d+)'> for the first digits anywhere.
+A shorthand for this is C<+0>, or C<+N> to find the Nth number.
-=item --header
+=item B<--header>
Prepend a chart axis with minimum and maximum values labeled.
-=item -H, --human-readable
+=item B<-H>, B<--human-readable>
Format values using SI unit prefixes,
-turning long numbers like I<12356789> into I<12.4M>.
-Also changes an exponent I<1.602176634e-19> to I<160.2z>.
+turning long numbers like C<12356789> into C<12.4M>.
+Also changes an exponent C<1.602176634e-19> to C<160.2z>.
Short integers are aligned but kept without decimal point.
-=item --sexagesimal
+=item B<--sexagesimal>
Convert seconds to HH:MM:SS time format.
-=item -t, --interval[=(<seconds> | -<lines>)]
+=item B<-t>, B<--interval>[=(I<seconds> | B<->I<lines>)]
Output partial progress every given number of seconds or input lines.
An update can also be forced by sending a I<SIGALRM> alarm signal.
-=item -l, --length=[-]<size>[%]
+=item B<-l>, B<--length>=[B<->]I<size>[B<%>]
Trim line contents (between number and bars)
to a maximum number of characters.
The exceeding part is replaced by an abbreviation sign,
-unless C<--length=0>.
+unless B<--length=0>.
Prepend a dash (i.e. make negative) to enforce padding
regardless of encountered contents.
-=item -L, --limit=[<count> | [-]<start>(-[<end>] | +<count>)]
+=item B<-L>, B<--limit>=[I<count> | [B<->]I<start>(B<->[I<end>] | B<+>I<count>)]
Select a range of lines to display.
-A single integer indicates the last line number (like C<head>),
-or first line counting from the bottom if negative (like C<tail>).
+A single integer indicates the last line number (like I<head>),
+or first line counting from the bottom if negative (like I<tail>).
A range consists of a starting line number followed by either
a dash C<-> to an optional end, or plus sign C<+> with count.
All hidden input is still counted and analyzed for statistics,
but disregarded for padding and bar size.
-=item -e, --log
+=item B<-e>, B<--log>
-Logarithmic (I<e>xponential) scale instead of linear
+Logarithmic (B<e>xponential) scale instead of linear
to compare orders of magnitude.
-=item --graph-format=<character>
+=item B<--graph-format>=I<character>
Glyph to repeat for the graph line.
Defaults to a dash C<->.
-=item -m, --markers=<format>
+=item B<-m>, B<--markers>=I<format>
Statistical positions to indicate on bars.
A single indicator glyph precedes each position:
=over 2
-=item <number>
+=item I<number>
Exact value to match on the axis.
-A vertical bar at the zero crossing is displayed by I<|0>
+A vertical bar at the zero crossing is displayed by C<|0>
for negative values.
-For example I<π3.14> would locate pi.
+For example C<π3.14> would locate pi.
-=item I</><interval>
+=item B</>I<interval>
Repeated at every multiple of a number.
-For example I<:/1> for a grid at every integer.
+For example C<:/1> for a grid at every integer.
-=item <percentage>I<v>
+=item I<percentage>B<v>
Ranked value at the given percentile.
-The default shows I<+> at I<50v> for the mean or median;
+The default shows C<+> at C<50v> for the mean or median;
the middle value or average between middle values.
-One standard deviation right of the mean is at about I<68.3v>.
-The default includes I<< >31.73v <68.27v >>
-to encompass all I<normal> results, or 68% of all entries, by B<< <--> >>.
+One standard deviation right of the mean is at about C<68.3v>.
+The default includes C<< >31.73v <68.27v >>
+to encompass all I<normal> results, or 68% of all entries, by I<< <--> >>.
-=item I<avg>
+=item B<avg>
Matches the average;
the sum of all values divided by the number of counted lines.
-Indicated by default as I<=>.
+Indicated by default as C<=>.
=back
-=item --min=<number>, --max=<number>
+=item B<--min>=I<number>, B<--max>=I<number>
Bars extend from 0 or the minimum value if lower,
to the largest value encountered.
These options can be set to customize this range.
-=item --palette=(<preset> | <color>...)
+=item B<--palette>=(I<preset> | I<color>...)
Override colors of parsed numbers.
-Can be any CSI escape, such as I<90> for default dark grey,
-or alternatively I<1;30> for bright black.
+Can be any CSI escape, such as C<90> for default dark gray,
+or alternatively C<1;30> for bright black.
In case of additional colors,
the last is used for values equal to the maximum, the first for minima.
-If unspecified, these are green and red respectively (I<31 90 32>).
+If unspecified, these are green and red respectively (C<31 90 32>).
Multiple intermediate colors will be distributed
relative to the size of values.
-Predefined color schemes are named I<whites> and I<fire>,
-or I<greys> and I<fire256> for 256-color variants.
+A non-numeric name can refer to a predefined color scheme:
+
+=over 8
+
+=item B<whites>
-=item -_, --spark
+Minimal set of monochrome brightnesses.
+
+=item B<grays>
+
+Utilize the 24 grayscale ramp in 256-color terminals.
+
+=item B<fire>
+
+Gradient red to white in 7 out of 16 colors.
+
+=item B<fire256>
+
+Extended to 17 colors out of 256.
+
+=item B<rainbow>
+
+Saturated red to green to blue to red.
+
+=item B<random>
+
+All 215 extended colors in unrelated orders.
+
+=back
+
+=item B<-_>, B<--spark>
Replace lines by I<sparklines>,
-single characters (configured by C<--indicators>)
+single characters (configured by B<--indicators>)
corresponding to input values.
-=item --indicators[=<characters>]
+=item B<--indicators>[=I<characters>]
Prefix a unicode character corresponding to each value.
The first specified character will be used for non-values,
the remaining sequence will be distributed over the range of values.
Unspecified, block fill glyphs U+2581-2588 will be used.
-=item -s, --stat
+=item B<-s>, B<--stat>
Total statistics after all data.
-=item -u, --unmodified
+While processing (possibly a neverending pipe),
+intermediate results are also shown on signal I<SIGINFO> if available (control+t on BSDs)
+or I<SIGQUIT> otherwise (ctrl+\ on linux).
+
+=item B<-u>, B<--unmodified>
Do not reformat values, keeping leading whitespace.
Keep original value alignment, which may be significant in some programs.
-=item --value-length=<size>
+=item B<--value-length>=I<size>
Reserved space for numbers.
-=item -w, --width=<columns>
+=item B<-w>, B<--width>=I<columns>
Override the maximum number of columns to use.
Appended graphics will extend to fill up the entire screen,
otherwise determined by the environment variable I<COLUMNS>
-or by running the C<tput> command.
+or by running the I<tput> command.
-=item -h, --usage
+=item B<-h>, B<--usage>
Overview of available options.
-=item --help
+=item B<--help>
Full pod documentation
as rendered by perldoc.
-=item -V, --version
+=item B<-V>, B<--version>
Version information.
ping google.com | barcat -f'time=\K' -t
-Commonly used after counting, for example users on the current server:
-
- users | tr ' ' '\n' | sort | uniq -c | barcat
-
-Letter frequencies in text files:
+Commonly used after counting, eg letter frequencies in text files:
cat /usr/share/games/fortunes/*.u8 |
perl -CS -nE 'say for grep length, split /\PL*/, uc' |
sort | uniq -c | barcat
+Users on the current server while preserving order:
+
+ users | tr ' ' '\n' | barcat -c
+
Number of HTTP requests per day:
- cat httpd/access.log | cut -d\ -f4 | cut -d: -f1 | uniq -c | barcat
+ barcat -cf'\[([^:]+)' httpd/access.log
Any kind of database query results, preserving returned alignment: