use open qw( :std :utf8 );
use experimental qw( lexical_subs );
-our $VERSION = '1.02';
+our $VERSION = '1.04';
use Getopt::Long '2.33', qw( :config gnu_getopt );
sub podexit {
GetOptions(\%opt,
'color|c!',
'C' => sub { $opt{color} = 0 },
- 'field|f=s',
+ 'field|f=s' => sub {
+ eval {
+ local $_ = $_[1];
+ $opt{anchor} = /^[0-9]+$/ ? qr/(?:\S*\h+){$_}\K/ : qr/$_/;
+ } or die $@ =~ s/(?: at .+)?$/ for option $_[0]/r;
+ },
+ 'human-readable|H!',
'interval|t:i',
'trim|length|l=s' => sub {
my ($optname, $optval) = @_;
$opt{trim} = $optval;
},
'value-length=i',
+ 'hidemin=i',
+ 'hidemax=i',
+ 'limit|L=s' => sub {
+ my ($optname, $optval) = @_;
+ $optval ||= 0;
+ ($opt{hidemin}, $opt{hidemax}) =
+ $optval =~ m/\A (?: ([0-9]+)? - )? ([0-9]+)? \z/x or die(
+ "Value \"$optval\" invalid for option limit",
+ " (range expected)\n"
+ );
+ },
'markers|m=s',
+ 'stat|s!',
'unmodified|u!',
'width|w=i',
'usage|h' => sub { podexit() },
$opt{width} ||= $ENV{COLUMNS} || 80;
$opt{color} //= -t *STDOUT; # enable on tty
$opt{trim} *= $opt{width} / 100 if $opt{trimpct};
+$opt{units} = [split //, ' kMGTPEZYyzafpnμm'] if $opt{'human-readable'};
+$opt{anchor} //= qr/\A/;
+$opt{'value-length'} = 6 if $opt{units};
+
+my (@lines, @values, @order);
if (defined $opt{interval}) {
$opt{interval} ||= 1;
alarm $opt{interval};
};
alarm $opt{interval};
+
+ eval {
+ require Tie::Array::Sorted;
+ tie @order, 'Tie::Array::Sorted', sub { $_[1] <=> $_[0] };
+ } or warn $@, "Expect slowdown with large datasets!\n";
}
$SIG{INT} = 'IGNORE'; # continue after assumed eof
-my (@lines, @values);
-my $anchor = !defined $opt{field} ? qr/\A/ :
- $opt{field} =~ /^[0-9]+$/ ? qr/(?:\S*\h+){$opt{field}}\K/ :
- $opt{field};
+my $valmatch = qr/$opt{anchor} ( \h* -? [0-9]* \.? [0-9]+ (?: e[+-]?[0-9]+ )? |)/x;
while (readline) {
s/\r?\n\z//;
s/^\h*// unless $opt{unmodified};
- push @values, s/$anchor ( \h* -? [0-9]* \.? [0-9]+ |)/\n/x && $1;
- if (defined $opt{trim}) {
+ push @values, s/$valmatch/\n/ && $1;
+ push @order, $1 if length $1;
+ if (defined $opt{trim} and defined $1) {
my $trimpos = abs $opt{trim};
if ($trimpos <= 1) {
$_ = substr $_, 0, 1;
sub show_lines {
-state $nr = 0;
+state $nr = $opt{hidemin} ? $opt{hidemin} - 1 : 0;
@lines and @lines > $nr or return;
+@lines or return;
+@lines > $nr or return unless $opt{hidemin};
-my @order = sort { $b <=> $a } grep { length } @values;
-my $maxval = $order[0];
-my $minval = min $order[-1], 0;
+@order = sort { $b <=> $a } @order unless tied @order;
+my $maxval = ($opt{hidemax} ? max grep { length } @values[0 .. $opt{hidemax} - 1] : $order[0]) // 0;
+my $minval = min $order[-1] // (), 0;
my $lenval = $opt{'value-length'} // max map { length } @order;
my $len = defined $opt{trim} && $opt{trim} <= 0 ? -$opt{trim} + 1 :
- max map { length $values[$_] && length $lines[$_] } 0 .. $#lines; # left padding
+ max map { length $values[$_] && length $lines[$_] }
+ 0 .. min $#lines, $opt{hidemax} || (); # left padding
my $size = ($maxval - $minval) &&
($opt{width} - $lenval - $len) / ($maxval - $minval); # bar multiplication
}
}
+@lines > $nr or return if $opt{hidemin};
+
+sub sival {
+ my $unit = int(log(abs $_[0] || 1) / log(10) - 3*($_[0] < 1) + 1e-15);
+ my $float = $_[0] !~ /^0*[-0-9]{1,3}$/;
+ sprintf('%3.*f%1s',
+ $float && ($unit % 3) == ($unit < 0), # tenths
+ $_[0] / 1000 ** int($unit/3), # number
+ $#{$opt{units}} * 1.5 < abs $unit ? "e$unit" : $opt{units}->[$unit/3]
+ );
+}
+
while ($nr <= $#lines) {
+ $nr >= $opt{hidemax} and last if defined $opt{hidemax};
my $val = $values[$nr];
if (length $val) {
my $color = !$opt{color} ? 0 :
$val == $order[0] ? 32 : # max
$val == $order[-1] ? 31 : # min
90;
- $val = sprintf "%*s", $lenval, $val;
+ $val = $opt{units} ? sival($val) : sprintf "%*s", $lenval, $val;
$val = "\e[${color}m$val\e[0m" if $color;
}
my $line = $lines[$nr] =~ s/\n/$val/r;
printf '%-*s', $len + length($val), $line;
print $barmark[$_] // '-' for 1 .. $size && (($values[$nr] || 0) - $minval) * $size + .5;
say '';
+
$nr++;
}
}
show_lines();
+if ($opt{stat}) {
+ if ($opt{hidemin} or $opt{hidemax}) {
+ $opt{hidemin} ||= 1;
+ $opt{hidemax} ||= @lines;
+ printf '%s of ', sum(@values[$opt{hidemin} - 1 .. $opt{hidemax} - 1]) // 0;
+ }
+ if (@order) {
+ my $total = sum @order;
+ printf '%s total', $total;
+ printf ' in %d values', scalar @values;
+ printf ' (%s min, %*.*f avg, %s max)',
+ $order[-1], 0, 2, $total / @order, $order[0];
+ }
+ say '';
+}
+
__END__
+=encoding utf8
=head1 NAME
-graph - append bar chart to input numbers
+barcat - graph to visualize input values
=head1 SYNOPSIS
-B<graph> [<options>] [<input>]
+B<barcat> [<options>] [<input>]
=head1 DESCRIPTION
or capture the numbers itself,
for example I<-f'(\d+)'> for the first digits anywhere.
+=item -H, --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>.
+Short integers are aligned but kept without decimal point.
+
=item -t, --interval[=<seconds>]
Interval time to output partial progress.
Prepend a dash (i.e. make negative) to enforce padding
regardless of encountered contents.
+=item -L, --limit=(<count>|<start>-[<end>])
+
+Stop output after a number of lines.
+All input is still counted and analyzed for statistics,
+but disregarded for padding and bar size.
+
=item -m, --markers=
Statistical positions to indicate on bars.
=back
+=item -s, --stat
+
+Total statistics after all data.
+
=item -u, --unmodified
Do not strip leading whitespace.
=head1 EXAMPLES
-Commonly used after counting, such as users on the current server:
+Draw a sine wave:
- users | sed 's/ /\n/g' | sort | uniq -c | graph
+ seq 30 | awk '{print sin($1/10)}' | barcat
-Letter frequencies in text files:
+Compare file sizes (with human-readable numbers):
- cat /usr/share/games/fortunes/*.u8 |
- perl -CO -nE 'say for grep length, split /\PL*/, uc' |
- sort | uniq -c | graph
+ du -d0 -b * | barcat -H
-Memory usage of user processes:
+Memory usage of user processes with long names truncated:
- ps xo %mem,pid,cmd | graph -l40
+ ps xo %mem,pid,cmd | barcat -l40
-Sizes (in megabytes) of all root files and directories:
+Monitor network latency from prefixed results:
- du -d0 -m * | graph
+ ping google.com | barcat -f'time=\K' -t
-Number of HTTP requests per day:
+Commonly used after counting, for example users on the current server:
- cat log/access.log | cut -d\ -f4 | cut -d: -f1 | uniq -c | graph
+ users | sed 's/ /\n/g' | sort | uniq -c | barcat
-Any kind of database query with leading counts:
+Letter frequencies in text files:
- echo 'SELECT count(*),schemaname FROM pg_tables GROUP BY 2' |
- psql -t | graph -u
+ cat /usr/share/games/fortunes/*.u8 |
+ perl -CS -nE 'say for grep length, split /\PL*/, uc' |
+ sort | uniq -c | barcat
-Exchange rate USD/EUR history from CSV download provided by ECB:
+Number of HTTP requests per day:
- curl https://sdw.ecb.europa.eu/export.do \
- -Gd 'node=SEARCHRESULTS&q=EXR.D.USD.EUR.SP00.A&exportType=csv' |
- grep '^[12]' | graph -f',\K' --value-length=7
+ cat log/access.log | cut -d\ -f4 | cut -d: -f1 | uniq -c | barcat
-Total population history from the World Bank dataset (XML):
+Any kind of database query with counts, preserving returned alignment:
- curl http://api.worldbank.org/v2/country/1W/indicator/SP.POP.TOTL |
- xmllint --xpath '//*[local-name()="date" or local-name()="value"]' - |
- sed -r 's,</wb:value>,\n,g; s,(<[^>]+>)+, ,g' | graph -f1
+ echo 'SELECT count(*),schemaname FROM pg_tables GROUP BY 2' |
+ psql -t | barcat -u
-Movies per year from prepared JSON data:
+External datasets, like movies per year:
curl https://github.com/prust/wikipedia-movie-data/raw/master/movies.json |
- jq '.[].year' | uniq -c | graph
+ perl -054 -nlE 'say if s/^"year"://' | uniq -c | barcat
+
+But please get I<jq> to process JSON
+and replace the manual selection by C<< jq '.[].year' >>.
Pokémon height comparison:
- curl https://github.com/Biuni/PokemonGO-Pokedex/raw/master/pokedex.json |
- jq -r '.pokemon[] | [.height,.num,.name] | join(" ")' | graph
+ curl https://github.com/Biuni/PokemonGO-Pokedex/raw/master/pokedex.json |
+ jq -r '.pokemon[] | [.height,.num,.name] | join(" ")' | barcat
+
+USD/EUR exchange rate from CSV provided by the ECB:
+
+ curl https://sdw.ecb.europa.eu/export.do \
+ -Gd 'node=SEARCHRESULTS&q=EXR.D.USD.EUR.SP00.A&exportType=csv' |
+ grep '^[12]' | barcat -f',\K' --value-length=7
-Git statistics, such commit count by year:
+Total population history from the World Bank dataset (XML):
+External datasets, like total population in XML from the World Bank:
- git log --pretty=%ci | cut -b-4 | uniq -c | graph
+ curl http://api.worldbank.org/v2/country/1W/indicator/SP.POP.TOTL |
+ xmllint --xpath '//*[local-name()="date" or local-name()="value"]' - |
+ sed -r 's,</wb:value>,\n,g; s,(<[^>]+>)+, ,g' | barcat -f1 -H
-Or the most frequent authors:
+And of course various Git statistics, such commit count by year:
- git shortlog -sn | graph | head -3
+ git log --pretty=%ci | cut -b-4 | uniq -c | barcat
-Latency history:
+Or the top 3 most frequent authors with statistics over all:
- ping google.com | graph -f'time=\K' -t
+ git shortlog -sn | barcat -L3 -s
=head1 AUTHOR