X-Git-Url: http://git.shiar.nl/barcat.git/blobdiff_plain/047d20c16603fb80b5abfe2fe1d1bb073e117eab..ca6b20aa1d3f9ca870d0a26a802b9a263218ea79:/barcat diff --git a/barcat b/barcat index a664d0d..15a2269 100755 --- a/barcat +++ b/barcat @@ -1,4 +1,4 @@ -#!/usr/bin/env perl +#!/usr/bin/perl -CA use 5.018; use warnings; use utf8; @@ -6,13 +6,9 @@ use List::Util qw( min max sum ); use open qw( :std :utf8 ); use experimental qw( lexical_subs ); -our $VERSION = '1.04'; +our $VERSION = '1.06'; use Getopt::Long '2.33', qw( :config gnu_getopt ); -sub podexit { - require Pod::Usage; - Pod::Usage::pod2usage(-exitval => 0, -perldocopt => '-oman', @_); -} my %opt; GetOptions(\%opt, 'color|c!', @@ -37,6 +33,8 @@ GetOptions(\%opt, 'value-length=i', 'hidemin=i', 'hidemax=i', + 'minval=f', + 'maxval=f', 'limit|L=s' => sub { my ($optname, $optval) = @_; $optval ||= 0; @@ -46,30 +44,73 @@ GetOptions(\%opt, " (range expected)\n" ); }, + 'header!', 'markers|m=s', + 'graph-format=s' => sub { + $opt{'graph-format'} = substr $_[1], 0, 1; + }, + 'spark:s' => sub { + $opt{spark} = [split //, $_[1] || '▁▂▃▄▅▆▇█']; + }, + 'palette:s' => sub { + $opt{palette} = [ split /\s/, $_[1] ]; + }, 'stat|s!', + 'signal-stat=s', 'unmodified|u!', 'width|w=i', - 'usage|h' => sub { podexit() }, - 'help' => sub { podexit(-verbose => 2) }, + 'version' => sub { + say "barcat version $VERSION"; + exit; + }, + 'usage|h' => sub { + local $/; + my $pod = readline *DATA; + $pod =~ s/^=over\K/ 22/m; # indent options list + $pod =~ s/^=item \N*\n\n\N*\n\K(?:(?:^=over.*?^=back\n)?(?!=)\N*\n)*/\n/msg; + + require Pod::Usage; + my $parser = Pod::Usage->new; + $parser->select('SYNOPSIS', 'OPTIONS'); + $parser->output_string(\my $contents); + $parser->parse_string_document($pod); + + $contents =~ s/\n(?=\n\h)//msg; # strip space between items + print $contents; + exit; + }, + 'help|?' => sub { + require Pod::Usage; + Pod::Usage::pod2usage( + -exitval => 0, -perldocopt => '-oman', -verbose => 2, + ); + }, ) or exit 64; # EX_USAGE $opt{width} ||= $ENV{COLUMNS} || 80; $opt{color} //= -t *STDOUT; # enable on tty +$opt{'graph-format'} //= '-'; $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 $opt{units}; +$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{palette} //= $opt{color} && [31, 90, 32]; my (@lines, @values, @order); +$SIG{$_} = \&show_stat for $opt{'signal-stat'} || (); +$SIG{ALRM} = sub { + show_lines(); + alarm $opt{interval} if defined $opt{interval} and $opt{interval} > 0; +}; +$SIG{INT} = \&show_exit; + if (defined $opt{interval}) { $opt{interval} ||= 1; - $SIG{ALRM} = sub { - show_lines(); - alarm $opt{interval}; - }; - alarm $opt{interval}; + alarm $opt{interval} if $opt{interval} > 0; eval { require Tie::Array::Sorted; @@ -77,8 +118,6 @@ if (defined $opt{interval}) { } or warn $@, "Expect slowdown with large datasets!\n"; } -$SIG{INT} = 'IGNORE'; # continue after assumed eof - my $valmatch = qr/$opt{anchor} ( \h* -? [0-9]* \.? [0-9]+ (?: e[+-]?[0-9]+ )? |)/x; while (readline) { s/\r?\n\z//; @@ -87,18 +126,27 @@ while (readline) { push @order, $1 if length $1; if (defined $opt{trim} and defined $1) { my $trimpos = abs $opt{trim}; + $trimpos -= length $1 if $opt{unmodified}; if ($trimpos <= 1) { - $_ = substr $_, 0, 1; + $_ = substr $_, 0, 2; } elsif (length > $trimpos) { substr($_, $trimpos - 1) = '…'; } } push @lines, $_; + show_lines() if defined $opt{interval} and $opt{interval} < 0 + and $. % $opt{interval} == 0; } $SIG{INT} = 'DEFAULT'; +sub color { + $opt{color} and defined $_[0] or return ''; + return "\e[$_[0]m" if defined wantarray; + $_ = color(@_) . $_ . color(0) if defined; +} + sub show_lines { state $nr = $opt{hidemin} ? $opt{hidemin} - 1 : 0; @@ -107,8 +155,8 @@ state $nr = $opt{hidemin} ? $opt{hidemin} - 1 : 0; @lines > $nr or return unless $opt{hidemin}; @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 $maxval = $opt{maxval} // ($opt{hidemax} ? max grep { length } @values[0 .. $opt{hidemax} - 1] : $order[0]) // 0; +my $minval = $opt{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 $lines[$_] } @@ -117,25 +165,35 @@ my $size = ($maxval - $minval) && ($opt{width} - $lenval - $len) / ($maxval - $minval); # bar multiplication my @barmark; -if ($opt{markers} // 1 and $size > 0) { - my sub orderpos { (($order[$_[0]] + $order[$_[0] + .5]) / 2 - $minval) * $size } - $barmark[ (sum(@order) / @order - $minval) * $size ] = '='; # average - $barmark[ orderpos($#order * .31731) ] = '>'; - $barmark[ orderpos($#order * .68269) ] = '<'; - $barmark[ orderpos($#order / 2) ] = '+'; # mean - $barmark[ -$minval * $size ] = '|' if $minval < 0; # zero - defined and $opt{color} and $_ = "\e[36m$_\e[0m" for @barmark; +if ($opt{markers} and $size > 0) { + for my $markspec (split /\h/, $opt{markers}) { + my ($char, $func) = split //, $markspec, 2; + my $pos = eval { + if ($func eq 'avg') { + return sum(@order) / @order; + } + elsif ($func =~ /\A([0-9.]+)v\z/) { + my $index = $#order * $1 / 100; + return ($order[$index] + $order[$index + .5]) / 2; + } + else { + return $func; + } + } - $minval; + $pos >= 0 or next; + color(36) for $barmark[$pos * $size] = $char; + } state $lastmax = $maxval; if ($maxval > $lastmax) { print ' ' x ($lenval + $len); - printf "\e[90m" if $opt{color}; + printf color(90); printf '%-*s', ($lastmax - $minval) * $size + .5, '-' x (($values[$nr - 1] - $minval) * $size); - print "\e[92m" if $opt{color}; + print color(92); say '+' x (($maxval - $lastmax - $minval) * $size + .5); - print "\e[0m" if $opt{color}; + print color(0); $lastmax = $maxval; } } @@ -143,40 +201,55 @@ if ($opt{markers} // 1 and $size > 0) { @lines > $nr or return if $opt{hidemin}; sub sival { - my $unit = int(log($_[0]) / log(1000) - ($_[0] < 1) + 1e-15); - my $float = sprintf '%e', $_[0] / 1000 ** $unit; #TODO: or $_[0] =~ /\./ - $float -= int($float); - sprintf('%*.*f%*s', - $float ? (5,1) : (3,0), # length and tenths - $_[0] / 1000 ** $unit, # number - $float ? 0 : 3, # unit size - $#{$opt{units}} >> 1 < abs $unit ? "e$unit" : $opt{units}->[$unit] + 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] ); } +say( + color(31), sprintf('%*s', $lenval, $minval), + color(90), '-', color(36), '+', + color(32), sprintf('%*s', $size * ($maxval - $minval) - 3, $maxval), + color(90), '-', color(36), '+', + color(0), +) if $opt{header}; + while ($nr <= $#lines) { $nr >= $opt{hidemax} and last if defined $opt{hidemax}; my $val = $values[$nr]; + my $rel = length $val && ($val - $minval) / ($maxval - $minval); + + if ($opt{spark}) { + print color($opt{palette}->[ $rel * $#{$opt{palette}} ]) if $opt{palette}; + print $opt{spark}->[ $rel * $#{$opt{spark}} ]; + next; + } + if (length $val) { - my $color = !$opt{color} ? 0 : - $val == $order[0] ? 32 : # max - $val == $order[-1] ? 31 : # min - 90; + my $color = !$opt{palette} ? undef : + $val == $order[0] ? $opt{palette}->[-1] : # max + $val == $order[-1] ? $opt{palette}->[0] : # min + $opt{palette}->[ $rel * ($#{$opt{palette}} - 1) + 1 ]; $val = $opt{units} ? sival($val) : sprintf "%*s", $lenval, $val; - $val = "\e[${color}m$val\e[0m" if $color; + color($color) for $val; } 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; + print $barmark[$_] // $opt{'graph-format'} for 1 .. $size && (($values[$nr] || 0) - $minval) * $size + .5; say ''; - +} +continue { $nr++; } +say $opt{palette} ? color(0) : '' if $opt{spark}; } -show_lines(); -if ($opt{stat}) { +sub show_stat { if ($opt{hidemin} or $opt{hidemax}) { $opt{hidemin} ||= 1; $opt{hidemax} ||= @lines; @@ -184,14 +257,26 @@ if ($opt{stat}) { } if (@order) { my $total = sum @order; - printf '%s total', $total; + printf '%s total', color(1) . $total . color(0); printf ' in %d values', scalar @values; - printf ' (%s min, %*.*f avg, %s max)', - $order[-1], 0, 2, $total / @order, $order[0]; + printf(' (%s min, %s avg, %s max)', + color(31) . $order[-1] . color(0), + color(36) . (sprintf '%*.*f', 0, 2, $total / @order) . color(0), + color(32) . $order[0] . color(0), + ); } say ''; } +sub show_exit { + show_lines(); + show_stat() if $opt{stat}; + exit 130 if @_; # 0x80+signo + exit; +} + +show_exit(); + __END__ =encoding utf8 @@ -209,6 +294,12 @@ Visualizes relative sizes of values read from input (file(s) or STDIN). Contents are concatenated similar to I, 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 +you'll need a larger animal like I. + =head1 OPTIONS =over @@ -231,6 +322,10 @@ A string can indicate the starting position of a value or capture the numbers itself, for example I<-f'(\d+)'> for the first digits anywhere. +=item --header + +Prepend a chart axis with minimum and maximum values labeled. + =item -H, --human-readable Format values using SI unit prefixes, @@ -238,9 +333,10 @@ 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[=] +=item -t, --interval[=(|-)] -Interval time to output partial progress. +Output partial progress every given number of seconds or input lines. +An update can also be forced by sending a I alarm signal. =item -l, --length=[-][%] @@ -258,38 +354,69 @@ 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= +=item --graph-format= -Statistical positions to indicate on bars. -Cannot be customized yet, -only disabled by providing an empty argument. +Glyph to repeat for the graph line. +Defaults to a dash C<->. -Any value enables all marker characters: +=item -m, --markers= + +Statistical positions to indicate on bars. +A single indicator glyph precedes each position: =over 2 -=item B<=> +=item -Average: -the sum of all values divided by the number of counted lines. +Exact value to match on the axis. +A vertical bar at the zero crossing is displayed by I<|0> +for negative values. +For example I<:3.14> would show a colon at pi. -=item B<+> +=item I -Mean, median: +Ranked value at the given percentile. +The default shows I<+> at I<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 results, or 68% of all entries, by B<< <--> >>. -=item B<<> +=item I -Standard deviation left of the mean. -Only 16% of all values are lower. +Matches the average; +the sum of all values divided by the number of counted lines. +Indicated by default as I<=>. -=item B<< > >> +=back -Standard deviation right of the mean. -The part between B<< <--> >> encompass all I results, -or 68% of all entries. +=item --min=, --max= -=back +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=... + +Override colors of parsed numbers. +Can be any CSI escape, such as I<90> for default dark grey, +or alternatively I<1;30> for bold 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>). + +=item --spark[=] + +Replace lines by I, +single characters corresponding to input values. +A specified sequence of unicode characters will be used for +Of a specified sequence of unicode characters, +the first one will be used for non-values, +the last one for the maximum, +the second (if any) for the minimum, +and any remaining will be distributed over the range of values. +Unspecified, block fill glyphs U+2581-2588 will be used. =item -s, --stat @@ -297,7 +424,7 @@ Total statistics after all data. =item -u, --unmodified -Do not strip leading whitespace. +Do not reformat values, keeping leading whitespace. Keep original value alignment, which may be significant in some programs. =item --value-length= @@ -309,6 +436,19 @@ Reserved space for numbers. Override the maximum number of columns to use. Appended graphics will extend to fill up the entire screen. +=item -h, --usage + +Overview of available options. + +=item --help + +Full documentation +rendered by perldoc. + +=item --version + +Version information. + =back =head1 EXAMPLES @@ -348,6 +488,11 @@ Any kind of database query with counts, preserving returned alignment: echo 'SELECT count(*),schemaname FROM pg_tables GROUP BY 2' | psql -t | barcat -u +Earthquakes worldwide magnitude 1+ in the last 24 hours: + + https://earthquake.usgs.gov/earthquakes/feed/v1.0/summary/1.0_day.csv | + column -tns, | graph -f4 -u -l80% + External datasets, like movies per year: curl https://github.com/prust/wikipedia-movie-data/raw/master/movies.json | @@ -382,6 +527,12 @@ Or the top 3 most frequent authors with statistics over all: git shortlog -sn | barcat -L3 -s +Activity of the last days (substitute date C<-v-{}d> on BSD): + + ( git log --pretty=%ci --since=30day | cut -b-10 + seq 0 30 | xargs -i date +%F -d-{}day ) | + sort | uniq -c | awk '$1--' | graph --spark + =head1 AUTHOR Mischa POSLAWSKY