X-Git-Url: http://git.shiar.nl/barcat.git/blobdiff_plain/13a52d166f90627fdffd7ed362b753bc4bdb4110..9a0f5056e03159f5a58f23385e2995a54b83cbd8:/barcat diff --git a/barcat b/barcat index 5ac55ba..d4a8274 100755 --- a/barcat +++ b/barcat @@ -1,29 +1,30 @@ -#!/usr/bin/env perl -use 5.018; +#!/usr/bin/perl -CA +use 5.014; use warnings; use utf8; use List::Util qw( min max sum ); use open qw( :std :utf8 ); -use experimental qw( lexical_subs ); +use re '/msx'; -our $VERSION = '1.03'; +our $VERSION = '1.07'; 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]; + $opt{anchor} = /\A[0-9]+\z/ ? qr/(?:\S*\h+){$_}\K/ : qr/$_/; + } or die $@ =~ s/(?:\ at\ \N+)?\Z/ for option $_[0]/r; + }, 'human-readable|H!', 'interval|t:i', 'trim|length|l=s' => sub { my ($optname, $optval) = @_; $optval =~ s/%$// and $opt{trimpct}++; - $optval =~ m/^-?[0-9]+$/ or die( + $optval =~ m/\A-?[0-9]+\z/ or die( "Value \"$optval\" invalid for option $optname", " (number or percentage expected)\n" ); @@ -32,124 +33,308 @@ GetOptions(\%opt, 'value-length=i', 'hidemin=i', 'hidemax=i', - 'limit|L=s' => sub { + 'minval=f', + 'maxval=f', + 'limit|L:s' => sub { my ($optname, $optval) = @_; $optval ||= 0; ($opt{hidemin}, $opt{hidemax}) = - $optval =~ m/\A (?: ([0-9]+)? - )? ([0-9]+)? \z/x or die( + $optval =~ m/\A (?: ([0-9]+)? - )? ([0-9]+)? \z/ or die( "Value \"$optval\" invalid for option limit", " (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 { + $opt{palette} = { + fire => [qw( 90 31 91 33 93 97 96 )], + fire88 => [map {"38;5;$_"} qw( + 80 32 48 64 68 72 76 77 78 79 47 + )], + fire256=> [map {"38;5;$_"} qw( + 235 52 88 124 160 196 + 202 208 214 220 226 227 228 229 230 231 159 + )], + ramp88 => [map {"38;5;$_"} qw( + 64 65 66 67 51 35 39 23 22 26 25 28 + )], + whites => [qw( 1;30 0;37 1;37 )], + greys => [map {"38;5;$_"} 52, 235..255, 47], + }->{$_[1]} // [ split /[^0-9;]/, $_[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 $/ = undef; # slurp + my $pod = readline *DATA; + $pod =~ s/^=over\K/ 25/; # indent options list + $pod =~ s/^=item\ \N*\n\n\N*\n\K (?:(?:^=over.*?^=back\n)?(?!=)\N*\n)*/\n/g; + $pod =~ s/[.,](?=\n)//g; # trailing punctuation + $pod =~ s/^=item\ \K(?=--)/____/g; # align long options + # abbreviate indicators + $pod =~ s/\Q>.../s>/g; + $pod =~ s/<(?:number|count|seconds)>/N/g; + $pod =~ s//\Uchar$1/g; + $pod =~ s/\Q | /|/g; + $pod =~ s/(?/\U$1/g; # uppercase + + require Pod::Usage; + my $parser = Pod::Usage->new(USAGE_OPTIONS => { + -indent => 2, -width => 78, + }); + $parser->select('SYNOPSIS', 'OPTIONS'); + $parser->output_string(\my $contents); + $parser->parse_string_document($pod); + + $contents =~ s/\n(?=\n\h)//msg; # strip space between items + $contents =~ s/^\ \ \K____/ /g; # nbsp substitute + 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{width} ||= $ENV{COLUMNS} || qx(tput cols) || 80 unless $opt{spark}; $opt{color} //= -t *STDOUT; # enable on tty +$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 $opt{unmodified}; +$opt{'signal-stat'} //= exists $SIG{INFO} ? 'INFO' : 'QUIT'; +$opt{markers} //= '=avg >31.73v <68.27v +50v |0'; +$opt{palette} //= $opt{color} && [31, 90, 32]; +$opt{hidemin} = ($opt{hidemin} || 1) - 1; +$opt{input} = (@ARGV && $ARGV[0] =~ m/\A[-0-9]/) ? \@ARGV : undef + and undef $opt{interval}; + +my (@lines, @values, @order); + +$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 $opt{interval} > 0; -$SIG{INT} = 'IGNORE'; # continue after assumed eof + eval { + require Tie::Array::Sorted; + tie @order, 'Tie::Array::Sorted', sub { $_[1] <=> $_[0] }; + } or warn $@, "Expect slowdown with large datasets!\n"; +} -my (@lines, @values); -my $anchor = !defined $opt{field} ? qr/\A/ : - $opt{field} =~ /^[0-9]+$/ ? qr/(?:\S*\h+){$opt{field}}\K/ : - $opt{field}; -while (readline) { +my $valmatch = qr< + $opt{anchor} ( \h* -? [0-9]* [.]? [0-9]+ (?: e[+-]?[0-9]+ )? |) +>x; +while (defined ($_ = $opt{input} ? shift @{ $opt{input} } : 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}) { + s/\A\h*// unless $opt{unmodified}; + my $valnum = s/$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 $opt{unmodified}; if ($trimpos <= 1) { - $_ = substr $_, 0, 1; + $_ = substr $_, 0, 2; } elsif (length > $trimpos) { + # cut and replace (intentional lvalue for speed, contrary to PBP) substr($_, $trimpos - 1) = '…'; } } push @lines, $_; + show_lines() if defined $opt{interval} and $opt{interval} < 0 + and $. % $opt{interval} == 0; +} + +if ($opt{'zero-missing'}) { + push @values, (0) x 10; } $SIG{INT} = 'DEFAULT'; -sub show_lines { +sub color { + $opt{color} and defined $_[0] or return ''; + return "\e[$_[0]m" if defined wantarray; + $_ = color(@_) . $_ . color(0) if defined; +} + +sub sival { + my $unit = int(log(abs $_[0] || 1) / log(10) - 3*($_[0] < 1) + 1e-15); + my $float = $_[0] !~ /\A0*[-0-9]{1,3}\z/; + return 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] + ); +} -state $nr = $opt{hidemin} ? $opt{hidemin} - 1 : 0; -@lines and @lines > $nr or return; +sub show_lines { -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; +state $nr = $opt{hidemin}; +@lines or return; +@lines > $nr or return; + +@order = sort { $b <=> $a } @order unless tied @order; +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 $range = $maxval - $minval; 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 .. min $#lines, $opt{hidemax} || (); # left padding -my $size = ($maxval - $minval) && - ($opt{width} - $lenval - $len) / ($maxval - $minval); # bar multiplication +my $size = defined $opt{width} && $range && + ($opt{width} - $lenval - $len) / $range; # bar multiplication my @barmark; -if ($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/) { + die "Invalid marker $char: percentile $1 out of bounds\n" if $1 > 100; + my $index = $#order * $1 / 100; + return ($order[$index] + $order[$index + .5]) / 2; + } + elsif ($func =~ /\A-?[0-9.]+\z/) { + return $func; + } + else { + die "Unknown marker $char: $func\n"; + } + }; + defined $pos or do { + warn $@ if $@; + next; + }; + $pos -= $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 $opt{color}; - say '+' x (($maxval - $lastmax - $minval) * $size + .5); - print "\e[0m" if $opt{color}; + print color(92); + say '+' x (($range - $lastmax) * $size + .5); + print color(0); $lastmax = $maxval; } } +say( + color(31), sprintf('%*s', $lenval, $minval), + color(90), '-', color(36), '+', + color(32), sprintf('%*s', $size * $range - 3, $maxval), + color(90), '-', color(36), '+', + color(0), +) if $opt{header}; + while ($nr <= $#lines) { - $nr >= $opt{hidemax} and last if $opt{hidemax}; + $nr >= $opt{hidemax} and last if defined $opt{hidemax}; my $val = $values[$nr]; + my $rel = length $val && $range && ($val - $minval) / $range; + my $color = !length $val || !$opt{palette} ? undef : + $val == $order[0] ? $opt{palette}->[-1] : # max + $val == $order[-1] ? $opt{palette}->[0] : # min + $opt{palette}->[ $rel * ($#{$opt{palette}} - 1) + 1 ]; + + if ($opt{spark}) { + say '' if $opt{width} and $nr and $nr % $opt{width} == 0; + print color($color), $opt{spark}->[ + !$val || !$#{$opt{spark}} ? 0 : # blank + $val == $order[0] ? -1 : # max + $val == $order[-1] ? 1 : # min + $#{$opt{spark}} < 3 ? 1 : + $rel * ($#{$opt{spark}} - 3) + 2.5 + ]; + next; + } + if (length $val) { - my $color = !$opt{color} ? 0 : - $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 $opt{palette} ? color(0) : '' if $opt{spark}; + + return $nr; +} + +sub show_stat { + if ($opt{hidemin} or $opt{hidemax}) { + printf '%s of ', sum(grep { length } + @values[$opt{hidemin} .. ($opt{hidemax} || @lines) - 1] + ) // 0; + } + if (@order) { + my $total = sum @order; + printf '%s total', color(1) . sprintf('%.8g', $total) . color(0); + printf ' in %d values', scalar @order; + printf ' over %d lines', scalar @lines if @order != @lines; + 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 ''; + return 1; +} +sub show_exit { + show_lines(); + show_stat() if $opt{stat}; + exit 130 if @_; # 0x80+signo + exit; } -show_lines(); + +show_exit(); __END__ +=encoding utf8 =head1 NAME @@ -157,14 +342,21 @@ barcat - graph to visualize input values =head1 SYNOPSIS -B [] [] +B [] [... | ] =head1 DESCRIPTION -Visualizes relative sizes of values read from input (file(s) or STDIN). +Visualizes relative sizes of values read from input +(parameters, 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 @@ -175,7 +367,7 @@ Force colored output of values and bar markers. Defaults on if output is a tty, disabled otherwise such as when piped or redirected. -=item -f, --field=(|) +=item -f, --field=( | ) Compare values after a given number of whitespace separators, or matching a regular expression. @@ -187,14 +379,21 @@ 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, 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=[-][%] @@ -206,48 +405,88 @@ unless C<--length=0>. Prepend a dash (i.e. make negative) to enforce padding regardless of encountered contents. -=item -L, --limit=(|-[]) +=item -L, --limit[=( | -[])] 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 I + +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 left of the mean. -Only 16% of all values are lower. +=item --min=, --max= -=item B<< > >> +Bars extend from 0 or the minimum value if lower, +to the largest value encountered. +These options can be set to customize this range. -Standard deviation right of the mean. -The part between B<< <--> >> encompass all I results, -or 68% of all entries. +=item --palette=( | ...) -=back +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. + +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>). +Multiple intermediate colors will be distributed +relative to the size of values. + +Predefined color schemes are named I and I, +or I and I for 256-color variants. + +=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 + +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 +498,109 @@ 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: + + ping google.com | barcat -f'time=\K' -t + +Commonly used after counting, for example users on the current server: - du -d0 -m * | barcat + users | tr ' ' '\n' | 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: +In PostgreSQL from within the client: + + postgres=> SELECT sin(generate_series(0, 3, .1)) \g |barcat + +Earthquakes worldwide magnitude 1+ in the last 24 hours: + + curl https://earthquake.usgs.gov/earthquakes/feed/v1.0/summary/1.0_day.csv | + column -tns, | barcat -f4 -u -l80% + +External datasets, like movies per year: + + curl https://github.com/prust/wikipedia-movie-data/raw/master/movies.json -L | + 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 -L | + 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): +Total population history in XML from the World Bank: - curl http://api.worldbank.org/v2/country/1W/indicator/SP.POP.TOTL | + curl http://api.worldbank.org/v2/country/1W/indicator/SP.POP.TOTL -L | 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 +And of course various Git statistics, such commit count by year: -Pokémon height comparison: + git log --pretty=%ci | cut -b-4 | uniq -c | barcat - curl https://github.com/Biuni/PokemonGO-Pokedex/raw/master/pokedex.json | - jq -r '.pokemon[] | [.height,.num,.name] | join(" ")' | barcat +Or the top 3 most frequent authors with statistics over all: -Git statistics, such commit count by year: + git shortlog -sn | barcat -L3 -s - git log --pretty=%ci | cut -b-4 | uniq -c | barcat +Sparkline graphics of simple input given as inline parameters: -Or the most frequent authors: + barcat --spark= 3 1 4 1 5 0 9 2 4 - git shortlog -sn | barcat -L3 +Activity graph of the last days (substitute date C<-v-{}d> on BSD): -Latency history: - - 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--' | barcat --spark =head1 AUTHOR