X-Git-Url: http://git.shiar.nl/barcat.git/blobdiff_plain/13a52d166f90627fdffd7ed362b753bc4bdb4110..a0ff197f92a8630d87509509ad635c8f194fa29b:/barcat diff --git a/barcat b/barcat index 5ac55ba..650e406 100755 --- a/barcat +++ b/barcat @@ -1,4 +1,4 @@ -#!/usr/bin/env perl +#!/usr/bin/perl -CA use 5.018; use warnings; use utf8; @@ -6,18 +6,19 @@ use List::Util qw( min max sum ); use open qw( :std :utf8 ); use experimental qw( lexical_subs ); -our $VERSION = '1.03'; +our $VERSION = '1.05'; 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!', '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 { @@ -42,58 +43,111 @@ GetOptions(\%opt, ); }, 'markers|m=s', + 'graph-format=s' => sub { + $opt{'graph-format'} = substr $_[1], 0, 1; + }, + 'spark:s' => sub { + $opt{spark} = [split //, $_[1] || '⎽▁▂▃▄▅▆▇█']; + }, + 'stat|s!', 'unmodified|u!', 'width|w=i', - 'usage|h' => sub { podexit() }, - 'help' => sub { podexit(-verbose => 2) }, + '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} = $opt{'human-readable'} && ['', qw( k M G T <> n μ m )]; +$opt{units} = [split //, ' kMGTPEZYyzafpnμm'] if $opt{'human-readable'}; +$opt{anchor} //= qr/\A/; +$opt{'value-length'} = 6 if $opt{units}; +$opt{'value-length'} = 1 if $opt{unmodified}; + +my (@lines, @values, @order); + +$SIG{QUIT} = \&show_stat; +$SIG{ALRM} = sub { + show_lines(); + alarm $opt{interval} if defined $opt{interval} and $opt{interval} > 0; +}; 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; + tie @order, 'Tie::Array::Sorted', sub { $_[1] <=> $_[0] }; + } or warn $@, "Expect slowdown with large datasets!\n"; } -$SIG{INT} = 'IGNORE'; # continue after assumed eof +$SIG{INT} = sub { + $SIG{INT} = 'DEFAULT'; # reset for subsequent attempts + exit if !$.; + '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}; + $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; @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 = $opt{hidemax} ? max @values[0 .. $opt{hidemax} - 1] : $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 $lines[$_] } @@ -109,47 +163,86 @@ if ($opt{markers} // 1 and $size > 0) { $barmark[ orderpos($#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; + color(36) for @barmark; 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; } } +@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 $opt{hidemax}; + $nr >= $opt{hidemax} and last if defined $opt{hidemax}; my $val = $values[$nr]; + + if ($opt{spark}) { + print $opt{spark}->[ ($val - $minval) / $maxval * $#{$opt{spark}} ]; + next; + } + if (length $val) { - my $color = !$opt{color} ? 0 : + my $color = !$opt{color} ? undef : $val == $order[0] ? 32 : # max $val == $order[-1] ? 31 : # min 90; - $val = sprintf "%3.1f%1s", $val / 1000**$_, $opt{units}->[$_] - for $opt{units} ? int(log($val) / log(1000) - ($val < 1)) : (); - $val = sprintf "%*s", $lenval, $val; - $val = "\e[${color}m$val\e[0m" if $color; + $val = $opt{units} ? sival($val) : sprintf "%*s", $lenval, $val; + color($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 '' if $opt{spark}; } show_lines(); +sub show_stat { + if ($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', color(1) . $total . color(0); + printf ' in %d values', scalar @values; + 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 ''; +} +show_stat() if $opt{stat}; + __END__ +=encoding utf8 =head1 NAME @@ -165,6 +258,12 @@ 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 @@ -191,10 +290,13 @@ for example I<-f'(\d+)'> for the first digits anywhere. 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[=] +=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=[-][%] @@ -212,6 +314,11 @@ Stop output after a number of lines. All input is still counted and analyzed for statistics, but disregarded for padding and bar size. +=item --graph-format= + +Glyph to repeat for the graph line. +Defaults to a dash C<->. + =item -m, --markers= Statistical positions to indicate on bars. @@ -245,9 +352,13 @@ or 68% of all entries. =back +=item -s, --stat + +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= @@ -259,70 +370,102 @@ 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 -Commonly used after counting, such as users on the current server: +Draw a sine wave: - users | sed 's/ /\n/g' | sort | uniq -c | barcat + 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 | barcat + du -d0 -b * | barcat -H -Memory usage of user processes: +Memory usage of user processes with long names truncated: 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 * | barcat + ping google.com | barcat -f'time=\K' -t + +Commonly used after counting, for example users on the current server: + + users | sed 's/ /\n/g' | sort | uniq -c | barcat + +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 Number of HTTP requests per day: cat log/access.log | cut -d\ -f4 | cut -d: -f1 | uniq -c | barcat -Any kind of database query with leading counts: +Any kind of database query with counts, preserving returned alignment: echo 'SELECT count(*),schemaname FROM pg_tables GROUP BY 2' | psql -t | barcat -u -Exchange rate USD/EUR history from CSV download provided by ECB: +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 | + perl -054 -nlE 'say if s/^"year"://' | uniq -c | barcat + +But please get I 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(" ")' | 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 Total population history from the World Bank dataset (XML): +External datasets, like total population in XML from the World Bank: curl http://api.worldbank.org/v2/country/1W/indicator/SP.POP.TOTL | xmllint --xpath '//*[local-name()="date" or local-name()="value"]' - | sed -r 's,,\n,g; s,(<[^>]+>)+, ,g' | barcat -f1 -H -Movies per year from prepared JSON data: - - curl https://github.com/prust/wikipedia-movie-data/raw/master/movies.json | - jq '.[].year' | uniq -c | barcat - -Pokémon height comparison: - - curl https://github.com/Biuni/PokemonGO-Pokedex/raw/master/pokedex.json | - jq -r '.pokemon[] | [.height,.num,.name] | join(" ")' | barcat - -Git statistics, such commit count by year: +And of course various Git statistics, such commit count by year: git log --pretty=%ci | cut -b-4 | uniq -c | barcat -Or the most frequent authors: +Or the top 3 most frequent authors with statistics over all: - git shortlog -sn | barcat -L3 + git shortlog -sn | barcat -L3 -s -Latency history: +Activity of the last days (substitute date C<-v-{}d> on BSD): - ping google.com | barcat -f'time=\K' -t + ( 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