# perl -w # ShowTable.pm # # Copyright (C) 1996,1997 Alan K. Stebbens # # This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # You should have received a copy of the GNU General Public License # along with this program; if not, write to the Free Software # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. # # $Id: ShowTable.pm,v 1.20 1997/03/02 22:04:50 aks Exp $ # package Data::ShowTable; =head1 NAME B - routines to display tabular data in several formats. =head1 USAGE C B { I => I, ... }; B I<\@titles>, I<\@types>, I<\@widths>, I<\&row_sub> [, I<\&fmt_sub> ]; B I<\@dbnames>; B { I => I, ... }; B I<\@tblnames>; B { I => I, ... }; B I<\@columns>, I<\@col_types>, I<\@col_lengths>, I<\@col_attrs>; B { I => I, ... }; B I<\@titles>, I<\@types>, I<\@widths>, I<\&row_sub> [, I<\&fmt_sub> ]; B { I => I, ... }; B I<\@titles>, I<\@types>, I<\@widths>, I<\&row_sub> [, I<\&fmt_sub>]; B { I => I, ... }; B I<\@titles>, I<\@types>, I<\@widths>, I<\&row_sub> [, I<\&fmt_sub>]; B { I => I, ... }; B I<\@titles>, I<\@types>, I<\@widths>, I<\&row_sub> [, I<\&fmt_sub>]; B { I => I, ... }; C; B<$Show_Mode> = 'I'; B<$Max_Table_Width> = I; B<$Max_List_Width> = I; B<$No_Escape> = I; B<%URL_Keys> = { "I<$colname>" => "I<$col_URL>", ... }; B<@Title_Formats> = ( I, , ... ); B<@Data_Formats> = ( I, , ... ); B I<$rewindflag>, I<\$index>, I<$col_array_1> [, I<$col_array_2>, ...;] I<$fmt> = B I<$value>, I<$type>, I<$max_width>, I<$width>, I<$precision>, I<$showmode>; [I<$plaintext> = ] B [I<$htmltext>]; =head1 DESCRIPTION The B<ShowTable> module provides subroutines to display tabular data, typially from a database, in nicely formatted columns, in several formats. Its arguments can either be given in a fixed order, or, as a single, anonymous hash-array. The output format for any one invocation can be one of four possible styles: =over 10 =item Box A tabular format, with the column titles and the entire table surrounded by a "box" of "C<+>", "C<->", and "C<|>" characters. See L<"ShowBoxTable"> for details. =item Table A simple tabular format, with columns automatically aligned, with column titles. See L<"ShowSimpleTable">. =item List A I<list> style, where columns of data are listed as a I<name>:I<value> pair, one pair per line, with rows being one or more column values, separated by an empty line. See L<"ShowListTable">. =item HTML The data is output as an HTML I<TABLE>, suitable for display through a I<Web>-client. See L<"ShowHTMLTable">. Input can either be plain ASCII text, or text with embedded HTML elements, depending upon an argument or global parameter. =back The subroutines which perform these displays are listed below. =head1 EXPORTED NAMES This module exports the following subroutines: ShowDatabases - show list of databases ShowTables - show list of tables ShowColumns - show table of column info ShowTable - show a table of data ShowRow - show a row from one or more columns ShowTableValue - show a single column's value ShowBoxTable - show a table of data in a box ShowListTable - show a table of data in a list ShowSimpleTable - show a table of data in a simple table ShowHTMLTable - show a table of data using HTML PlainText - convert HTML text into plain text All of these subroutines, and others, are described in detail in the following sections. =cut use Exporter; @ISA = qw( Exporter ); @EXPORT = qw( ShowDatabases ShowTables ShowColumns ShowTable ShowRow ShowBoxTable ShowHTMLTable ShowListTable ShowSimpleTable ShowTableValue Show_Mode PlainText URL_Keys ); @EXPORT_OK = qw( Show_Mode URL_Keys Title_Formats Data_Formats ); # Some control variables -- the user may set these $Show_Mode = 'Box'; # one of: List, Table, Box, or HTML $List_Wrap_Margin = 10; # break words up to this long $Max_Table_Width = ''; # if defined, scale tables $Max_List_Width = $ENV{'COLUMNS'} || 80; $No_Escape = ''; # escape by default %URL_Keys = (); @Title_Formats = (); # formats for HTML formatting @Data_Formats = (); use Carp; unshift(@INC, '.'); sub ShowDatabases; sub ShowTables; sub ShowColumns; sub ShowTable; sub ShowRow; sub PlainText; sub htmltext; sub get_params; sub html_formats; sub center; sub max_length; sub max; sub out; sub put; =head1 MODULES =head1 ShowTable Format and display the contents of one or more rows of data. S< >B<ShowTable> { I<parameter> => I<value>, ... }; S< >B<ShowTable> I<\@titles>, I<\@types>, I<\@widths>, I<\&row_sub> [, I<\&fmt_sub> [, I<$max_width> ] [, I<$show_mode> ] ]; The B<ShowTable> subroutine displays tabular data aligned in columns, with headers. B<ShowTable> supports four I<modes> of display: B<Box>, B<Table>, B<List>, and B<HTML>. Each mode is described separately below. The arguments to B<ShowTable> may be given in one of two ways: as a hashed-array, or by a combination of fixed order arguments, and some package-global variable settings. The hash-array parameters correspond to the fixed arguments and the global-parameter settings. In the list below, both the hash-array parameter name and the fixed-order argument name is given as the value. In the case where there is no fixed-order argument for a given parameter-value pair, then the corresponding global variable name is given. =over 10 =item C<titles> => I<\@titles> A reference to an array of column names, or titles. If a particular column name is null, then the string C<Field_I<num>> is used by default. To have a column have no title, use the empty string. =item C<types> => I<\@types> A reference to an array of types, one for each column. These types are passed to the I<fmt_sub> for appropriate formatting. Also, if a column type matches the regexp "C</text|char|string/i>", then the column alignment will be left-justified, otherwise it will be right-justified. =item C<widths> => I<\@widths> A reference to an array of column widths, which may be given as an integer, or as a string of the form: "I<width>.I<precision>". =item C<row_sub> => I<\&row_sub> A reference to a subroutine which successively returns rows of values in an array. It is called for two purposes, each described separately: * To fetch successive rows of data: @row = &$row_sub(0); When given a null, zero, or empty argument, the next row is returned. * To initialize or rewind the data traversal. $rewindable = &$row_sub(1); When invoked with a non-null argument, the subroutine should rewind its row pointer to start at the first row of data. If the data which I<row_sub> is traversing is not rewindable, it must return zero or null. If the data is rewindable, a non-null, non-zero value should be returned. The I<row_sub> must expect to be invoked once with a non-null argument, in order to discover whether or not the data is rewindable. If the data cannot be rewound, I<row_sub> will thereafter only be called with a zero argument. Specifically, I<row_sub> subroutine is used in this manner: $rewindable = &$row_sub(1); if ($rewindable) { while ((@row = &$row_sub(0)), $#row >= 0) { # examine lengths for optimal formatting } &$row_sub(1); # rewind } while ((@row = &$row_sub(0)), $#row >= 0) { # format the data } The consequence of data that is not rewindable, a reasonably nice table will still be formatted, but it may contain fairly large amounts of whitespace for wide columns. =item C<fmtsub> => I<\&fmt_sub> A reference to a subroutine which formats a value, according to its type, width, precision, and the current column width. It is invoked either with a fixed list of arguments, or with a hash-array of parameter and value pairs. $string = &fmt_sub { I<parameter> => I<value>, ... }; $string = &fmt_sub($value, $type, $max_width, $width, $precision) If I<\&fmt_sub> is omitted, then a default subroutine, B<ShowTableValue>, will be used, which will use Perl's standard string formatting rules. The arguments to I<\&fmt_sub>, either as values passed in a fixed order, or as part of the parameter value pair, are described in the section on L<"ShowTableValue> below. =item C<max_width> => I<number>, The maximum table width, including the table formatting characters. If not given, defaults to the global variable B<$Max_Table_Width>; =item C<show_mode> => 'I<mode>', The display mode of the output. One of five strings: C<'Box'>, C<'Table'>, C<'Simple'>, C<'List'>, and C<'HTML'>. =back =cut sub ShowTable { my @argv = @_; local ($_,$titles,$types,$widths,$row_sub,$fmt_sub, $max_width, $show_mode, $wrap_margin, $url_keys, $no_escape, $title_formats, $data_formats); my $args = get_params \@argv, { titles => \$titles, types => \$types, widths => \$widths, row_sub => \$row_sub, fmt_sub => \$fmt_sub, max_width => \$max_width, show_mode => \$show_mode, }, [qw(titles types widths row_sub fmt_sub max_width show_mode)]; # Default mode is from $Show_Mode global $show_mode = $args->{'show_mode'} = $Show_Mode unless $show_mode ne ''; $_ = $show_mode; if (/List/i) { &ShowListTable($args); } elsif (/HTML/i) { &ShowHTMLTable($args); } elsif (/Table/i) { &ShowSimpleTable($args); } else { &ShowBoxTable($args); } } =head1 ShowDatabases Show a list of database names. S< >B<ShowDatabases> I<\@dbnames>; S< >B<ShowDatabases> { 'data' => I<\@dbnames>, I<parameter> => I<value>, ...}; B<ShowDatabases> is intended to be used to display a list of database names, under the column heading of "Databases". It is a special case usage of B<ShowTable> (and can thus be passed any parameter suitable for B<ShowTable>. The argument, I<\@dbnames>, is a reference to an array of strings, used as the values of the single column display. =cut sub ShowDatabases { my @argv = @_; local $databases; my $args = get_params \@argv, {data => \$databases}, ['data']; $databases ne '' or croak "Missing array of databases.\n"; $args->{'titles'} = 'Databases' unless exists $args->{'titles'}; $args->{'types'} = [ 'char' ]; $args->{'width'} = max_length $databases; $args->{'lengths'} = $args->{'width'}; local( $current_row ) = 0; $args->{'row_sub'} = sub { &ShowRow( $_[0], \$current_row, $databases ); }; ShowTable $args; } =head1 ShowTables Show an array of table names. S< >B<ShowTables> I<\@tblnames>; S< >B<ShowTables> { 'data' => I<\@tblnames>, I<parameter> => I<value>, ...}; B<ShowTables> is used to display a list of table names, under the column heading of "Tables". It is a special case usage of B<ShowTable>, and can be passed any L<"ShowTable"> argument parameter. =cut sub ShowTables { my @argv = @_; local $tables; my $args = get_params \@argv, {data => \$tables}, ['data']; $tables ne '' or croak "Missing array of tables.\n"; $args->{'titles'} = 'Tables' unless exists $args->{'titles'}; $args->{'types'} = 'char'; $args->{'width'} = max_length $tables; $args->{'lengths'} = $args->{'width'}; local( $current_row ) = 0; $args->{'row_sub'} = sub { &ShowRow( $_[0], \$current_row, $tables ); }; ShowTable $args; } =head1 ShowColumns Display a table of column names, types, and attributes. S< >B<ShowColumns> { I<parameter> => I<values>, ... }; S< >B<ShowColumns> I<\@columns>, I<\@col_types>, I<\@col_lengths>, I<\@col_attrs>; The B<ShowColumns> subroutine displays a table of column names, types, lengths, and other attributes in a nicely formatted table. It is a special case usage of B<ShowTable>, and can be passed any argument suitable for L<"ShowTable">; The arguments are: =over 10 =item C<columns> = I<\@columns> An array of column names. This provides the value for the first column of the output. =item C<col_types> = I<\@col_types> An array of column types names. This provides the value for the second column. =item C<col_lengths> = I<\@col_lengths> An array of maximum lengths for corresponding columns. This provides the value for the third column of the output. =item C<col_attrs> = I<\@col_attrs> An array of column attributes array references (ie: an array of arrays). The attributes array for the first column are at "I<$col_attrs>-\>[0]". The first attribute of the second column is "I<$col_attrs>-\>[1][0]". =back The columns, types, lengths, and attributes are displayed in a table with the column headings: "Column", "Type", "Length", and "Attributes". This is a special case usage of B<ShowTable>, and can be passed additional arguments suitable for L<"ShowTable">. =cut sub ShowColumns { my @argv = @_; local ($col_names, $col_types, $col_lengths, $col_attributes); my $args = get_params \@argv, { col_names => \$col_names, col_types => \$col_types, col_lengths => \$col_lengths, col_attributes => \$col_attributes, },[qw(col_names col_types col_lengths col_attributes)]; $col_names ne '' or croak "Missing array of column names.\n"; $col_types ne '' or croak "Missing array of column types.\n"; $col_lengths ne '' or croak "Missing array of column lengths.\n"; $col_attributes ne '' or croak "Missing array of column attributes.\n"; # setup the descriptor arrays $args->{'titles'} = [ qw(Column Type Length Attributes) ]; $args->{'types'} = [ qw(varchar varchar int varchar) ]; # Do some data conversions before displaying # Convert attribute array to a string of attributes local @col_attrs = (); my $i; for ($i = 0; $i <= $#{$col_attributes}; $i++) { $col_attrs[$i] = join(', ',@{$col_attributes->[$i]}); } # count the widths, to setup the Column name column width $args->{'lengths'} = [ (max_length $col_names), (max_length $col_types), (max_length $col_lengths), (max_length \@col_attrs) ]; local($current_row) = 0; $args->{'row_sub'} = sub { &ShowRow($_[0], \$current_row, $col_names, $col_types, $col_lengths, \@col_attrs); }; # Finally, show the darn thing ShowTable $args; } =head1 ShowBoxTable Show tabular data in a box. S< >B<ShowBoxTable> { I<parameter> = I<value>, ... }; S< >B<ShowBoxTable> I<\@titles>, I<\@types>, I<\@widths>, I<\&row_sub> S< >[, [ I<\&fmt_sub> ] [, I<$max_width> ] ]; The B<ShowBoxTable> displays tabular data in titled columns using a "box" of ASCII graphics, looking something like this: +------------+----------+-----+----------+ | Column1 | Column2 | ... | ColumnN | +------------+----------+-----+----------+ | Value11 | Value12 | ... | Value 1M | | Value21 | Value22 | ... | Value 2M | | Value31 | Value32 | ... | Value 3M | | ... | ... | ... | ... | | ValueN1 | ValueN2 | ... | Value NM | +------------+----------+-----+----------+ The arguments are the same as with L<"ShowTable">. If the I<@titles> array is empty, the header row is omitted. =cut sub ShowBoxTable { my @argv = @_; local ($titles, $types, $col_widths, $row_sub, $fmt_sub, $max_width); my $args = get_params \@argv, { titles => \$titles, types => \$types, widths => \$col_widths, row_sub => \$row_sub, fmtsub => \$fmt_sub, max_width => \$max_width, }, [qw(titles types widths row_sub fmtsub max_width)]; $titles ne '' or croak "Missing column names array.\n"; $types ne '' or croak "Missing column types array.\n"; $col_widths ne '' or croak "Missing column width array.\n"; $row_sub ne '' or croak "Missing row subroutine.\n"; $fmt_sub = \&ShowTableValue if !defined($fmt_sub) || $fmt_sub eq ''; $max_width = $Max_Table_Width if !defined($max_width) || $max_width eq ''; my $rewindable = &$row_sub(1); # see if data is rewindable my ($num_cols, $widths, $precision, $max_widths) = &calc_widths($col_widths, $titles, $rewindable, $row_sub, $fmt_sub, $types, 'box', $max_width); my $width = 1; my $dashes = ' +'; my $title_line = ' |'; my $title; my $fmt = ' |'; # initial format string my $c; # Compose the box header for ($c = 0; $c < $num_cols; $c++) { $width = $max_widths->[$c]; # get previously calculated max col width $width += 2; # account for a blank on either # side of each value $dashes .= ('-' x $width); $dashes .= '+'; $title = $#$titles >= 0 && defined($titles->[$c]) ? $titles->[$c] : sprintf("Field_%d", $c+1); $title_line .= center $title, $width; $title_line .= '|'; } out $dashes; if ($#$titles >= 0) { out $title_line; out $dashes; } my @values; my @prefix = (" ", "<"); my @suffix = (" |", ">|"); my @cell; # loop over the data, formatting it into cells, one row at a time. while ((@values = &$row_sub(0)), $#values >= $[) { # first pass -- format each value into a string @cell = (); for ($c = 0; $c <= $#values; $c++) { $cell[$c] = &$fmt_sub($values[$c], $types->[$c], $max_widths->[$c], $widths->[$c], $precision->[$c], 'box'); } # second pass -- output each cell, wrapping if necessary my $will_wrap; my $wrapped = 0; do { $will_wrap = 0; put " |"; # start a line for ($c = 0; $c <= $#cell; $c++) { $will_wrap |= &putcell(\@cell, $c, $max_widths->[$c], \@prefix, \@suffix, $wrapped); } out ""; $wrapped++; } while ($will_wrap); } out $dashes; out ""; } =head1 ShowSimpleTable Display a table of data using a simple table format. S< >B<ShowSimpleTable> I<\@titles>, I<\@types>, I<\@widths>, I<\&row_sub> [, I<\&fmt_sub>]; S< >B<ShowSimpleTable> { I<parameter> => I<values>, ... }; The B<ShowSimpleTable> subroutine formats data into a simple table of aligned columns, in the following example: Column1 Column2 Column3 ------- ------- ------- Value1 Value2 Value3 Value12 Value22 Value32 Columns are auto-sized by the data's widths, plus two spaces between columns. Values which are too long for the maximum colulmn width are wrapped within the column. =cut sub ShowSimpleTable { my @argv = @_; local ($titles, $types, $col_widths, $row_sub, $fmt_sub, $max_width); my $args = get_params \@argv, { titles => \$titles, types => \$types, widths => \$col_widths, row_sub => \$row_sub, fmtsub => \$fmt_sub, max_width => \$max_width, }, [qw(titles types widths row_sub fmtsub max_width)]; $titles ne '' or croak "Missing column names array.\n"; $types ne '' or croak "Missing column types array.\n"; $col_widths ne '' or croak "Missing column width array.\n"; $row_sub ne '' or croak "Missing row sub array.\n"; $fmt_sub = \&ShowTableValue if !defined($fmt_sub) || $fmt_sub eq ''; $max_width = $Max_Table_Width if !defined($max_width) || $max_width eq ''; my $rewindable = &$row_sub(1); # see if data is rewindable my ($num_cols, $widths, $precision, $max_widths) = &calc_widths($col_widths, $titles, $rewindable, $row_sub, $fmt_sub, $types, 'table', $max_width); my $width = 1; my $dashes = ' '; my $title_line = ' '; my $title ; my $postfix = shift; my $c ; # Calculate the maximum widths for ($c = 0; $c < $num_cols; $c++) { $width = $max_widths->[$c]; $dashes .= ('-' x $width); $dashes .= ' '; next if $#$titles < 0; $title = center $titles->[$c], $width; $title_line .= $title; $title_line .= ' '; } out $title_line if $#$titles >= 0; out $dashes; my @values; my @prefix = (" ", "<"); my @suffix = (" ", ">"); while ((@values = &$row_sub(0)), $#values >= $[) { # first pass -- format each value into a string my @cell; for ($c = 0; $c <= $#values; $c++) { $cell[$c] = &$fmt_sub($values[$c], $types->[$c], $max_widths->[$c], $widths->[$c], $precision->[$c], 'table'); } # second pass -- output each cell, wrapping if necessary my $will_wrap; my $wrapped = 0; do { $will_wrap = 0; for ($c = 0; $c <= $#cell; $c++) { $will_wrap |= &putcell(\@cell, $c, $max_widths->[$c], \@prefix, \@suffix, $wrapped); } out ""; $wrapped++; } while ($will_wrap); } out ""; } =head1 ShowHTMLTable Display a table of data nicely using HTML tables. S< >B<ShowHTMLTable> { I<parameter> => I<value>, ... }; S< >B<ShowHTMLTable> I<\@titles>, I<\@types>, I<\@widths>, I<\&row_sub> [, I<\&fmt_sub> [, I<$max_width> [, I<\%URL_Keys> [, I<$no_escape> [, I<\@title_formats [, I<\@data_formats [, I<$table_attrs> ] ] ] ] ] ] ]; The B<ShowHTMLTable> displays one or more rows of columns of data using the HTML C<\<TABLE\>> feature. In addition to the usual parameter arguments of L<"ShowTable">, the following parameter arguments are defined: =over 10 =item C<url_keys> => I<\%URL_Keys>, This is a hash array of column names (titles) and corresponding base URLs. The values of any column names or indexes occuring as keys in the hash array will be generated as hypertext anchors using the associated I<printf>-like string as the base URL. Either the column name or the column index (beginning with 1) may be used as the hash key. In the string value, these macros can be substituted: "C<%K>" is replaced with the column name. "C<%V>" is replaced with the column value; "C<%I>" is replaced with the column index. For example, if we define the array: $base_url = "http://www.$domain/cgi/lookup?col=%K?val=%V"; %url_cols = ('Author' => $base_url, 'Name' => $base_url); Then, the values in the C<Author> column will be generated with the following HTML text: <A HREF="http://www.$domain/cgi/lookup?col=Author?val=somevalue>somevalue</A> and the values in the C<Name> column will be generated with the URL: <A HREF="http://www.$domain/cgi/lookup?col=Name?val=othervalue>othervalue</A> If this variable is not given, it will default to the global variable C<\%URL_Keys>. =item C<no_escape> => I<boolean>, Unless B<$no_escape> is set, HTML-escaping is performed on the data values in order to properly display the special HTML formatting characters : '\<', '\>', and '&'. If you wish to display data with embedded HTML text, you must set B<$no_escape>. Enabling embedded HTML, turns on certain heuristics which enable the user to more completely define appearance of the table. For instance, any C<\<TR\>> tokens found embedded *anywhere* within a row of data will be placed at the front of the row, within the generated C<\<TR\>>. Similarly, a row of data containing the C<\<THEAD\>> or C<\<TFOOT\>> tokens, and their closing counterparts, will begin and end, respectively a table header or footer data. =item C<title_formats> => I<\@title_formats>, =item C<tformats> => I<\@title_formats>, An array of HTML formatting elements for the column titles, one for each column. Each array element is a list of one or more HTML elements, given as C<\<ELEMENT\>> or plainly, C<ELEMENT>, and separated by a comma C<','>, semi-colon C<';'>, or vertical bar C<'|'>. Each given HTML element is prepended to the corresponding column title, in the order given. The corresponding HTML closing elements are appended in the opposite order. For example, if I<\@title_formats> contains the two elements: [ 'FONT SIZE=+2,BOLD', 'FONT COLOR=red,EM' ] then the text output for the title of the first column would be: <FONT SIZE=+2><BOLD>I<column_1_title></BOLD></FONT> If C<title_formats> is omitted, the global variable B<@Title_Formats> is used by default. =item C<data_formats> => I<\@data_formats>, =item C<dformats> => I<\@data_formats>, Similar to C<title_formats>, this array provides HTML formatting for the columns of each row of data. If C<data_formats> is omitted or null, then the global variable B<\@Data_Formats> is used by default. =item C<table_attrs> => I<$table_attrs>, This variable defines a string of attributes to be inserted within the C<\<TABLE\>> token. For example, if the user wishes to have no table border: ShowHTMLTable { ... table_attrs => 'BORDER=0', ... }; =back =cut sub ShowHTMLTable { my @argv = @_; local ($titles, $types, $col_widths, $row_sub, $fmt_sub, $max_width, $url_keys, $no_escape, $title_formats, $data_formats, $show_mode, $table_attrs); my $args = get_params \@argv, { titles => \$titles, types => \$types, widths => \$col_widths, row_sub => \$row_sub, fmtsub => \$fmt_sub, max_width => \$max_width, url_keys => \$url_keys, no_escape => \$no_escape, tformats => \$title_formats, dformats => \$data_formats, table_attrs => \$table_attrs, data_formats => 'tformats', title_formats => 'tformats', }, [qw(titles types widths row_sub fmtsub max_width url_keys no_escape title_formats data_formats table_attrs)]; $titles ne '' or croak "Missing column names array.\n"; $types ne '' or croak "Missing column types array.\n"; $col_widths ne '' or croak "Missing column width array.\n"; $row_sub ne '' or croak "Missing row sub array.\n"; # Defaults $fmt_sub = \&ShowTableValue if !defined($fmt_sub) || $fmt_sub eq ''; $max_width = $Max_Table_Width if !defined($max_width) || $max_width eq ''; $url_keys = \%URL_Keys if !defined($url_keys) || $url_keys eq ''; $title_formats = \@Title_Formats if !defined($title_formats) || $title_formats eq ''; $data_formats = \@Data_Formats if !defined($data_formats) || $data_formats eq ''; $no_escape = $No_Escape if !defined($no_escape); my $rewindable = &$row_sub(1); # see if rewindable my ($num_cols, $widths, $precision, $max_widths) = &calc_widths($col_widths, $titles, $rewindable, $row_sub, $fmt_sub, $types, 'html', $max_width); my $width = 1; my $total_width = 0; my $title_line = ''; my $title; my ($c,$x); my ($tprefixes,$tsuffixes,$dprefixes,$dsuffixes); # prepare the HTML prefixes and suffixes, if any ($tprefixes,$tsuffixes) = html_formats $title_formats if defined($title_formats) && $title_formats ne ''; ($dprefixes,$dsuffixes) = html_formats $data_formats if defined($data_formats) && $data_formats ne ''; if ($table_attrs) { # any table attributes? local($_) = $table_attrs; $table_attrs .= ' BORDER=1' unless /\bBORDER=/i; $table_attrs .= ' CELLPADDING=1' unless /\bCELLPADDING=/i; $table_attrs .= ' CELLSPACING=1' unless /\bCELLSPACING=/i; } else { $table_attrs = 'BORDER=2 CELLPADDING=1 CELLSPACING=1'; } out "<TABLE $table_attrs>\n<TR>" ; map { $total_width += defined($_) ? $_ : 0; } @$max_widths; for ($c = 0; $c < $num_cols; $c++) { # If the user specified a width, then use it. $width = defined($widths->[$c]) ? $widths->[$c] : $max_widths->[$c]; my $pct_width = int(100 * $width/$total_width); $title_line .= " <TH ALIGN=CENTER WIDTH=$pct_width%%>"; if ($#$titles >= 0) { if (($x = $#$tprefixes) >= 0) { $title_line .= $tprefixes->[$c > $x ? $x : $c]; } $title_line .= $no_escape ? $titles->[$c] : &htmltext($titles->[$c]); if (($x = $#$tsuffixes) >= 0) { $title_line .= $tsuffixes->[$c > $x ? $x : $c]; } } $title_line .= "</TH>\n"; } out $title_line; out "</TR>"; my ($href, $key, $val, $out); while ((@values = &$row_sub(0)), $#values >= $[) { out "<TR> "; # Walk through the values for ($c = 0; $c <= $#values; $c++) { $out = "<TD"; if (defined($val = $values[$c])) { # only worry about defined values # In HTML mode, all CHAR, TEXT, SYMBOL, or STRING data should # be escaped to protect HTML syntax "<", ">", "\", and "&". if ($types->[$c] =~ /char|text|symbol|string/i) { $val = &htmltext($val) unless $no_escape; $out .= " ALIGN=LEFT"; } else { $out .= " ALIGN=RIGHT"; } $out .= ">"; # Discover if either the column name or column index # have been mapped to a URL. $href = ''; foreach $key ( $#$titles >= 0 && &PlainText($titles->[$c]), sprintf("%d", $c+1)) { next unless $key ne '' && defined($url_keys->{$key}); $href = $url_keys->{$key}; last; } if ($href ne '') { if ($href =~ /%K/) { my $s = &htmltext(&PlainText($titles->[$c]), 1); $href =~ s/%K/$s/g; } if ($href =~ /%V/) { my $s = &htmltext($val, 1); $href =~ s/%V/$s/g; } if ($href =~ /%I/) { my $s = sprintf("%d", $c+1); $href =~ s/%I/$s/g; } $out .= sprintf("<A HREF=\"%s\">",$href); } $val = &$fmt_sub($val, $types->[$c], 0, $widths->[$c], $precision->[$c], 'html'); $val =~ s/^\s+//; # don't try to align $val =~ s/\s+$//; if (($x = $#$dprefixes) >= 0) { $out .= $dprefixes->[$c > $x ? $x : $c]; } $out .= $val; if (($x = $#$dsuffixes) >= 0) { $out .= $dsuffixes->[$c > $x ? $x : $c]; } $out .= "</A>" if $href; } else { $out .= ">"; } $out .= "</TD>"; out $out; } out "</TR>"; } out "</TABLE>"; } =head1 ShowListTable Display a table of data using a list format. S< >B<ShowListTable> { I<parameter> => I<value>, ... }; S< >B<ShowListTable> I<\@titles>, I<\@types>, I<\@widths>, I<\&row_sub> [, I<\&fmt_sub> [, I<$max_width> [, I<$wrap_margin> ] ] ]; The arguments for B<ShowListTable> are the same as for L<"ShowTable">, except for those described next. =over 10 =item C<max_width> = I<number>, =item C<wrap_margin> = I<number>, Lines are truncated, and wrapped when their length exceeds I<$max_width>. Wrapping is done on a word-basis, unless the resulting right margin exceeds I<$wrap_margin>, in which case the line is simply truncated at the I<$max_width> limit. The I<$max_width> variable defaults to B<$Max_List_Width>. The I<$wrap_margin> defaults to B<$List_Wrap_Margin>. =back In I<List> mode, columns (called "fields" in List mode) are displayed wth a field name and value pair per line, with records being one or more fields . In other words, the output of a table would look something like this: Field1_1: Value1_1 Field1_2: Value1_2 Field1_3: Value1_3 ... Field1-N: Value1_M <empty line> Field2_1: Value2_1 Field2_2: Value2_2 Field2_3: Value2_3 ... Field2_N: Value2_N ... FieldM_1: ValueM_1 FieldM_2: ValueM_2 ... FieldM_N: ValueM_N <empty line> <empty line> Characteristics of I<List> mode: =over 10 =item * two empty lines indicate the end of data. =item * An empty field (column) may be omitted, or may have a label, but no data. =item * A long line can be continue by a null field (column): Field2: blah blah blah : blah blah blah =item * On a continuation, the null field is an arbitrary number of leading white space, a colon ':', a single blank or tab, followed by the continued text. =item * Embedded newlines are indicated by the escape mechanism "\n". Similarly, embedded tabs are indicated with "\t", returns with "\r". =item * If the I<@titles> array is empty, the field names "C<Field_>I<NN>" are used instead. =back =cut sub ShowListTable { my @argv = @_; local ($titles, $types, $col_widths, $row_sub, $fmt_sub, $max_width, $wrap_margin); my $args = get_params \@argv, { titles => \$titles, types => \$types, widths => \$col_widths, row_sub => \$row_sub, fmtsub => \$fmt_sub, max_width => \$max_width, wrap_margin => \$wrap_margin, }, [qw(titles types widths row_sub fmt_sub max_width wrap_margin)]; defined($titles) && $titles ne '' or croak "Missing column names array.\n"; defined($types) && $types ne '' or croak "Missing column types array.\n"; defined($col_widths) && $col_widths ne '' or croak "Missing column width array.\n"; defined($row_sub) && $row_sub ne '' or croak "Missing row sub array.\n"; $fmt_sub = \&ShowTableValue if !defined($fmt_sub) || $fmt_sub eq ''; $max_width = $Max_List_Width if !defined($max_width) || $max_width eq ''; $wrap_margin = $List_Wrap_Margin if !defined($wrap_margin) || $wrap_margin eq ''; my $rewindable = &$row_sub(1); # init the row pointer my ($num_cols, $widths, $precision, $max_widths) = &calc_widths($col_widths, $titles, $rewindable, $row_sub, $fmt_sub, $types, 'list', ''); my $fmt = sprintf("%%-%ds : %%s\n", ($#$titles >= 0 ? &max_length($titles) : 8)); my @values; my ($value, $c, $cut, $line); my $col_limit = $max_width - 2; while ((@values = &$row_sub(0)), $#values >= $[) { for ($c = 0; $c <= $#values; $c++) { # get this column's title $title = $#$titles >= 0 ? $titles->[$c] : sprintf("Field_%d", $c+1); my $type = $types->[$c]; my $width = 0; my $prec = $precision->[$c]; $value = &$fmt_sub($values[$c], $type, 0, $width, $prec, 'list'); while (length($value)) { if (length($value) > ($cut = $col_limit)) { $line = substr($value, 0, $cut); if ($line =~ m/([-,;? \t])([^-,;? \t]*)$/ && length($2) <= $wrap_margin) { $cut = $col_limit - length($2); $line = substr($value, 0, $cut); } ($value = substr($value, $cut)) =~ s/^\s+//; } else { $line = $value; $value = ''; } out $fmt, $title, $line; $title = ''; } } out ""; } } =head1 ShowRow Fetch rows successively from one or more columns of data. S< >B<ShowRow> I<$rewindflag>, I<\$index>, I<$col_array_1> [, I<$col_array_2>, ...;] The B<ShowRow> subroutine returns a row of data from one or more columns of data. It is designed to be used as a I<callback> routine, within the B<ShowTable> routine. It can be used to select elements from one or more array reference arguments. If passed two or more array references as arguments, elements of the arrays selected by I<$index> are returned as the "row" of data. If a single array argument is passed, and each element of the array is itself an array, the subarray is returned as the "row" of data. If the I<$rewindflag> flag is set, then the I<$index> pointer is reset to zero, and "true" is returned (a scalar 1). This indicates that the data is rewindable to the B<ShowTable> routines. When the I<$rewindflag> is not set, then the current row of data, as determined by I<$index> is returned, and I<$index> will have been incremented. An actual invocation (from B<ShowColumns>) is: ShowTable \@titles, \@types, \@lengths, sub { &ShowRow( $_[0], \$current_row, $col_names, $col_types, $col_lengths, \@col_attrs); }; In the example above, after each invocation, the I<$current_row> argument will have been incremented. =cut sub ShowRow { my $rewind_flag = shift; my $index_ref = shift; # an indirect index my @columns = @_; # get rest of columns my @row; # we're selecting a row if ($rewind_flag) { $$index_ref = 0; # reset the pointer return 1; } return () if $#{$columns[0]} < $$index_ref; if ($#columns == 0) { # exactly one array ref argument my $data = $columns[0]->[$$index_ref]; # get the current data if (ref($data) eq 'ARRAY') { # if an array.. @row = @$data; # ..return the array of data } elsif (ref($data) eq 'HASH') {# if a hash.. @row = values %$data; # ..return the values } else { # otherwise.. @row = ($data); # ..return the data element } } else { # with two or more array refs.. my $col; # select elements from each for ($col = 0; $col <= $#columns; $col++) { push(@row, ${$columns[$col]}[$$index_ref]); } } ${$index_ref}++; # increment the index for the next call @row; # return this row of data } =head1 ShowTableValue Prepare and return a formatted representation of a value. A value argument, using its corresponding type, effective width, and precision is formatted into a field of a given maximum width. S< >I<$fmt> = B<ShowTableValue> I<$value>, I<$type>, I<$max_width>, I<$width>, I<$precision>, I<$showmode>; =over 10 =item C<width> => I<$width> =item I<$width> The width of the current value. If omittied, I<$max_width> is assumed. =item C<precision> => I<$precision> =item I<$precision> The number of decimal digits; zero is assumed if omittied. =item C<value> => I<$value> =item I<$value> The value to be formatted. =item I<$type> The type name of the value; eg: C<char>, C<varchar>, C<int>, etc. =item C<maxwidth> => I<$max_width> =item I<$max_width> The maximum width of any value in the current value's column. If I<$width> is zero or null, I<$max_width> is used by default. I<$max_width> is also used as a I<minimum> width, in case I<$width> is a smaller value. =item I<$width> The default width of the value, obtained from the width specification of the column in which this value occurs. =item I<$precision> The precision specification, if any, from the column width specification. =item I<$showmode> The mode of the output: one of "table", "list", "box", or "html". Currently, only the "html" mode is significant: it is used to avoid using HTML tokens as part of the formatted text and length calculations. =back =cut sub ShowTableValue { my $value = shift; my $type = shift; my $max_width = shift; my $width = shift; my $prec = shift || 2; my $showmode = shift; my $fmt = ($Type2Format{lc($type)} || $Type2Format{'char'}); my $str; $max_width = 0 if !defined($max_width) || $max_width eq ''; $width = $max_width if !defined($width) || $width eq ''; $width = min($width, $max_width) if $max_width > 0; if ($type =~ /money/i) { # money formatting is special if (($str = $value) !~ /[\$,]/) { # not already formatted? my ($d,$c) = split(/\./,$value,2); # reverse the digits $d = join('',reverse(split(//,abs($d)))); # do the grouping from the rightmost to the left $d =~ s/(...)(?=.)/$1,/g; # reverse the digits and grouping char $d = '$'.join('',reverse(split(//,$d))); # If there is any precision, add on pennies $d .= sprintf(".%02d",$c) if $prec > 0; # Mark as negative with '(xxx)' $d = '-'.$d if $value < 0; $str = $d; } } else { $fmt = sprintf ($fmt,$width,$prec); # If we are in HTML mode, and the value has any HTML tokens, # then format it always as a string (even if it might # be a decimal--this is a kluge but seems to work). if ($showmode =~ /html/i && $value =~ /<\/?($HTML_Elements)/) { $fmt =~ s/[df]/s/; # convert to string sub } $str = sprintf($fmt,$value); } if ($width > length(&PlainText($str))) { # right align the value if any kind of number $str = sprintf("%${width}s", $str) if $type =~ /int|float|real|numeric|money/i; } $str; } %Type2Format = ( 'char' => '%%-%ds', 'varchar' => '%%-%ds', 'symbol' => '%%-%ds', 'tinyint' => '%%%dd', 'shortint' => '%%%dd', 'int' => '%%%dd', 'real' => '%%%d.%df', 'float' => '%%%d.%df', 'numeric' => '%%%d.%df', 'text' => '%%-%ds', # The money types do not actually need to be in this table, since # ShowTableValue handle money formatting explicitly. However, some # one else might use this table, so we treat them like right-aligned # strings. 'money' => '%%%ds', 'smallmoney' => '%%%ds', ); =head1 PlainText S< >I<$plaintext> = B<&PlainText>(I<$htmltext>); S< >B<&PlainText> This function removes any HTML formatting sequences from the input argument, or from C<$_> if no argument is given. The resulting plain text is returned as the result. =cut # $plaintext = &PlainText($htmltext); # or: # &PlainText; # # Convert the argument and return as a string, or convert $_. sub PlainText { local($_) = shift if $#_ >= 0; # set local $_ if there's an argument # skip unless there's a sequence return $_ unless m=</?($HTML_Elements)=i; # HTML text? s{</?(?:$HTML_Elements)# # match and remove any HTML token.. (?:\ \w+# # ..then PARAM or PARAM=VALUE (?:\=(?:"(?:[^"]|\\")*"|# # ...."STRING" or.. [^"> ]+# # ....VALUE )# )?# # ..=VALUE is optional )*# # zero or more PARAM or PARAM=VALUE >}{}igx; # up to the closing '>' $_; # return the result } BEGIN { @HTML_Elements = qw( A ABBREV ACRONYM ADDRESS APP APPLET AREA AU B BANNER BASE BASEFONT BDO BGSOUND BIG BLINK BLOCKQUOTE BODY BQ BR CAPTION CENTER CITE CODE COL COLGROUP CREDIT DD DEL DFN DIR DIV DL DT EM EMBED FN FIG FONT FORM FRAME FRAMESET H1 H2 H3 H4 H5 H6 HEAD HP HR HTML I IMG INPUT INS ISINDEX KBD LANG LH LI LINK LISTING MAP MARQUEE MENU META NEXTID NOBR NOEMBED NOFRAMES NOTE OL OPTION OVERLAY P PARAM PERSON PLAINTEXT PRE Q S SAMP SELECT SMALL SPAN STRIKE STRONG SUB SUP TAB TABLE TBODY TD TEXTAREA TFOOT TH THEAD TITLE TR TT U UL VAR WBR XMP ); $HTML_Elements = join("|",@HTML_Elements); } =head1 VARIABLES The following variables may be set by the user to affect the display (with the defaults enclosed in square brackets [..]): =over 10 =item B<$Show_Mode> [Box] This is the default display mode when using B<ShowTable>. The environment variable, C<$ENV{'SHOW_MODE'}>, is used when this variable is null or the empty string. The possible values for this variable are: C<"Box">, C<"List">, C<"Table">, and C<"HTML">. Case is insignificant. =item B<$List_Wrap_Margin> [2] This variable's value determines how large a margin to keep before wrarpping a long value's display in a column. This value is only used in "List" mode. =item B<$Max_List_Width> [80] This variable, used in "List" mode, is used to determine how long an output line may be before wrapping it. The environment variable, C<$ENV{'COLUMNS'}>, is used to define this value when it is null. =item B<$Max_Table_Width> [''] This variable, when set, causes all tables to have their columns scaled such that their total combined width does not exceed this value. When this variable is not set, which is the default case, there is no maximum table width, and no scaling will be done. =item B<$No_Escape> [''] If set, allows embedded HTML text to be included in the data displayed in an HTML-formatted table. By default, the HTML formatting characters ("<", ">", and "&") occuring in values are escaped. =item B<%URL_Keys> In HTML mode, this variable is used to recognize which columns are to be displayed with a corresponding hypertext anchor. See L<"ShowHTMLTable"> for more details. =item B<@HTML_Elements> An array of HTML elements (as of HTML 3.0) used to recognize and strip for width calculations. =item B<$HTML_Elements> A regular expression string formed from the elements of B<@HTML_Elements>. =back =cut ############################## =head1 INTERNAL SUBROUTINES =head1 get_params S< >my I<$args> = B<&get_params> I<\@argv>, I<\%params>, I<\@arglist>; Given the I<@argv> originally passed to the calling sub, and the hash of named parameters as I<%params>, and the array of parameter names in the order expected for a pass-by-value invocation, set the values of each of the variables named in I<@vars>. If the only element of the I<@argv> is a hash array, then set the variables to the values of their corresponding parameters used as keys to the hash array. If the parameter is not a key of the I<%params> hash, and is not a key in the global hash B<%ShowTableParams>, then an error is noted. When I<@argv> has multiple elements, or is not a hash array, set each variable, in the order given within I<@arglist>, to the values from the I<@argv>, setting the variables named by each value in I<%params>. Variables may given either by name or by reference. The result is a HASH array reference, either corresponding directly to the HASH array passed as the single argument, or one created by associating the resulting variable values to the parameter names associated with the variable names. =cut sub get_params { my $argvref = shift or croak "Missing required argument.\n"; my $params = shift or croak "Missing required parameters hash.\n"; my $arglist = shift or croak "Missing required arglist array.\n"; my %args; my ($param, $var); if ($#$argvref == 0 && ref($argvref->[0]) eq 'HASH') { my $href = $argvref->[0]; %args = %$href; # initialize result with input hash foreach $param (keys %$href) { # for each named argument... # Is this a known parameter? if (exists($params->{$param})) { $var = $params->{$param}; while ($var ne '' && ref($var) eq '') { # indirect refs? $var = $params->{$param = $var}; } if ($var ne '') { $$var = $href->{$param}; # assign the param's variable $args{$param} = $$var; # make sure canonical param gets defined next; # go to the next parameter } } if (!exists($show_table_params{$param})) { croak "Unknown parameter: \"$param\"\n"; } } } else { # use args in the order given for variables my $i; for ($i = 0; $i <= $#$arglist; $i++) { $param = $arglist->[$i]; # get the next argument $var = $params->{$param}; # get it's variable next unless defined($var); while ($var ne '' && ref($var) eq '') { $var = $params->{$param = $var}; } if ($var ne '') { $$var = $i <= $#$argvref ? $argvref->[$i] : ''; $args{$param} = $$var; # assign to the hash } elsif (!exists($show_table_params{$param})) { croak "Unknown parameter: \"$param\" for argument $i.\n"; } } } # Now, make sure all variables get initialized foreach $param (keys %$params) { $var = $params->{$param}; while ($var ne '' && ref($var) eq '') { $var = $params->{$param = $var}; } if ($var ne '' && !exists($args{$param})) { $$var = $args{$param} = undef; } } \%args; # return the HASH ref } BEGIN { # A table of parameters used by all the external subroutines For # example, in order for parameters applicable to ShowHTMLTable to be # passed through ShowTable, they need to be defined in this table. @show_table_params = qw( caption col_attributes col_lengths col_names col_types data data_formats dformats fmt_sub fmtsub max_width no_escape row_sub show_mode table_attrs tformats title_formats titles types url_keys widths wrap_margin ); @show_table_params{@show_table_params} = () x (1 + $#show_table_params); undef @show_table_params; } =head1 html_formats S< >(I<$prefixes>,I<$suffixes>) = B<html_formats> I<\@html_formats>; The B<html_format> function takes an array reference of HTML formatting elements I<\@html_formats>, and builds two arrays of strings: the first: I<$prefixes>, is an array of prefixes containing the corresponding HTML formatting elements from I<\@html_formats>, and the second, I<$suffixes>, containing the appropriate HTML closing elements, in the opposite order. The result is designed to be used as prefixes and suffixes for the corresponding titles and column values. The array I<\@html_formats> contains lists of HTML formatting elements, one for each column (either title or data). Each array element is a list of one or more HTML elements, either given in HTML syntax, or as a "plain" name (ie: given as C<\<ELEMENT\>> or plainly, C<ELEMENT>). Multiple elements are separated by a comma C<','>. The resulting array of I<$prefixes> contains the corresponding opening elements, in the order given, with the proper HTML element syntax. The resulting array of I<$suffixes> contains the closing elements, in the opposite order given, with the proper HTML element syntax. For example, if I<\@html_formats> contains the two elements: [ 'FONT SIZE=+2,BOLD', 'FONT COLOR=red,EM' ] then the resulting two arrays will be returned as: [ [ '<FONT SIZE=+2><BOLD>', '<FONT COLOR=red><EM>' ], [ '</FONT></BOLD>', '</FONT></EM>' ] ] =cut sub html_formats { my $html_formats = shift; # array ref my $i; my (@prefixes, @suffixes); my ($html, $elt, $html_list, @html_list); my ($prefixes, $suffixes); local($_); foreach $html_list (@$html_formats) { @html_list = split(/,/,$html_list); $prefixes = $suffixes = ''; # initialize the list my %formats; # keep track of formats foreach (@html_list) { ($html, $elt) = (); if (($html, $elt) = /^(<)?\s*(\w+)/) {# <KEYWORD or KEYWORD next if $formats{$elt}++ > 0; # only do an element once $html = '<' unless $html; $prefixes .= $html.$elt.$'; $prefixes .= '>' unless $prefixes =~ />$/; $suffixes = $html.'/'.$elt.'>'.$suffixes; } } push(@prefixes, $prefixes); # even push empty items push(@suffixes, $suffixes); } ( \@prefixes, \@suffixes ); } =head1 calc_widths S< >(I<$num_cols>, I<$widths>, I<$precision>, I<$max_widths>) = S< >B<&calc_widths>( I<$widthspec>, I<$titles>, I<$rewindable>, S< >I<$row_sub>, I<$fmt_sub>, I<$types>, I<$showmode>, S< >I<$max_width>); =head2 B<DESCRIPTION> B<calc_widths> is a generalized subroutine used by all the B<ShowTable> variant subroutines to setup internal variables prior to formatting for display. B<Calc_widths> handles the column width and precision analysis, including scanning the data (if rewindable) for appropriate default values. The number of columns in the data is returned, as well as three arrays: the declared column widths, the column precision values, and the maximum column widths. =head2 B<RETURN VALUES> =over 10 =item I<$num_cols> is the number of columns in the data. If the data is not rewindable, this is computed as the maximum of the number of elements in the I<$widthspec> array and the number of elements in the I<$titles> array. When the data is rewindable, this is the maximum of the number of columns of each row of data. =item I<$widths> is the column widths array ref, without the precision specs (if any). Each column's width value is determined by the original I<$widthspec> value and/or the maximum length of the formatted data for the column. =item I<$precision> is the precision component (if any) of the original I<$widthspec> array ref. If there was no original precision component from the I<$widthspec>, and the data is rewindable, then the data is examined to determine the maximum default precision. =item I<$max_widths> is the ref to the array of maximum widths for the given columns. =head2 B<ARGUMENTS> =item I<$widthspec> A reference to an array of column width (or length) values, each given as an integer, real number, or a string value of "I<width>.I<precision>". If a value is zero or null, the length of the corresponding formatted data (if rewindable) and column title length are used to determine a reasonable default. If a column's I<width> portion is a positive, non-zero number, then the column will be this wide, regardless of the values lengths of the data in the column. If the column's I<width> portion is given as a negative number, then the positive value is used as a minimum column width, with no limit on the maximum column width. In other words, the column will be at least I<width> characters wide. If the data is not rewindable, and a column's width value is null or zero, then the length of the column title is used. This may cause severe wrapping of data in the column, if the column data lengths are much greater than the column title widths. =item I<$titles> The array ref to the column titles; used to determine the minimum acceptable width, as well as the default number of columns. If the C<$titles> array is empty, then the C<$widthspec> array is used to determine the default number of columns. =item I<$rewindable> A flag indicating whether or not the data being formatted is rewindable. If this is true, a pass over the data will be done in order to calculate the maximum lengths of the actual formatted data, using I<$fmt_sub> (below), rather than just rely on the declared column lengths. This allows for optimal column width adjustments (ie: the actual column widths may be less than the declared column widths). If it is not desired to have the column widths dynamically adjusted, then set the I<$rewindable> argument to 0, even if the data is rewindable. =item I<$row_sub> The code reference to the subroutine which returns the data; invoked only if I<$rewindable> is non-null. =item I<$fmt_sub> The subroutine used to determine the length of the data when formatted; if this is omitted or null, the length of the data is used by default. The I<$fmt_sub> is used only when the data is rewindable. =item I<$types> An array reference to the types of each of the value columns; used only when I<$fmt_sub> is invoked. =item I<$showmode> A string indicating the mode of the eventual display; one of four strings: "C<box>", "C<table>", "C<list>", and "C<html>". Used to adjust widths for formatting requirements. =item I<$max_width> The maximum width of the table being formatted. If set, and the total sum of the individual columns exceeds this value, the column widths are scaled down uniformly. If not set (null), no column width scaling is done. =back =cut sub calc_widths { my $widthspec = shift; my $titles = shift; my $rewindable = shift; my $row_sub = shift; my $fmt_sub = shift; my $types = shift; my $showmode = shift; my $max_width = shift; my @precision; # array of precision values my @setprec; # array of flags to set default precision my @widths; # array of widths my @max_widths; # array of max widths my @expandable; # flag if widths expandable my $num_cols; my $c; if ($#$widthspec >= 0) { @precision = @$widthspec; foreach (@precision) { s/^.*\.(\d+)/$1/ || ($_ = ''); } # The setprec array indicates which columns need a default precision @setprec = map { !length } @precision; # Get the integer portions @widths = map { length($_) ? int : 0 } @$widthspec; # Set @expandable if negative widths @expandable = map { $_ < 0 } @widths; # Convert widths to all positive values @widths = map abs, @widths; @max_widths = (0) x (1 + $#widths); # no maximums yet $num_cols = 1 + $#widths; } else { # No widths given @expandable = (1) x (1 + $#$titles); @precision = ('') x (1 + $#$titles); @setprec = @expandable; @max_widths = map length, @$titles; # initialize maximums to title widths $num_cols = 1 + $#$titles; } # If the data is rewindable, scan and accumulate *actual* widths for # each column, using the title lengths as a minimum. if ($rewindable) { my @values; my @prectype; if (ref($types) eq 'ARRAY') { @prectype = map {/float|num(eric|ber)|money|dec|real|precision|double/i } @$types; } # Scan the values while ((@values = &$row_sub(0)), $#values >= $[) { # If the new row is larger than the number of titles, adjust # the info arrays.. if ($num_cols < 1 + $#values) { # new column? $num_cols = 1 + $#values; # new # of columns for ($c = $#expandable + 1; $c <= $#values; $c++) { $expandable[$c] = 1; $precision[$c] = ''; $setprec[$c] = 1; $max_widths[$c] = 0; } } my $len; my $value; for ($c = 0; $c < $num_cols; $c++) { # Does this column's precision need setting? if ($setprec[$c]) { # Yes, is it a type of value which can use the precision? if ($prectype[$c]) { # yes, how much is the current value's default precision? if ($values[$c] =~ /\.(.*)$/) { $precision[$c] = length($1) if length($1) > $precision[$c]; } } else { # No, this column can't use the precision value -- don't # do this check on this column again $precision[$c] = $setprec[$c] = 0; } } # Now, let's get the formatted value so we can guess the best # default widths $value = # If a fmt_sub is available, use it to format the value $fmt_sub ? &$fmt_sub($values[$c], $types->[$c], 0, 0, $precision[$c], $showmode) # If no fmt sub, then use Perl stringify : length($showmode eq 'html' ? # in HTML mode? &PlainText($values[$c]) # use plain text : $values[$c]); # else, use raw text $len = length($value); $max_widths[$c] = $len if $c > $#max_widths || $len > $max_widths[$c]; } } # okay -- maximums scanned. # If the maximum table width set, scale the max_widths $max_width = 0 unless defined($max_width) && $max_width ne ''; if ($max_width > 0) { # Start with the given maximum, but adjust it to account for # the formatting and space characters. my $max_width = $max_width; $max_width -= $num_cols * 3 + 2 if $showmode eq 'box'; $max_width -= $num_cols * 2 - 1 if $showmode eq 'table'; my $total = 0; # Calculate the total table width for ($c = 0; $c <= $#max_widths; $c++) { $total += $max_widths[$c]; } if ($max_width < $total) { # Now scale it to the adjusted maximum table width for ($c = 0; $c <= $#max_widths; $c++) { $max_widths[$c] = int($max_widths[$c] * $max_width / $total); } } } # If the column is expandable, allow the width to grow to the max_width. # If the column is not expandable, allow the width to shrink to # the max_width if it is smaller. if ($#widths < 0) { # were there any widths? @widths = @max_widths; # nope, set them to the scanned values } else { $num_cols = max($num_cols, 1 + $#widths) if $#widths >= 0; my $len; for ($c = 0; $c < $num_cols; $c++) { # provide defaults first $max_widths[$c] = 0 if !defined($max_widths[$c]); $widths[$c] = $max_widths[$c] if $c > $#widths || !defined($widths[$c]); # if the column can shrink, let it if ($max_widths[$c] < $widths[$c]) { $widths[$c] = $max_widths[$c]; } elsif ($expandable[$c] || !$widths[$c]) { # allow the width to grow to the maximum width $widths[$c] = $max_widths[$c] if $widths[$c] < $max_widths[$c]; } elsif ($max_widths[$c] > $widths[$c] && $widths[$c] > 0) { # not expandable -- set the max width to the width value $max_widths[$c] = $widths[$c]; } # In either case, however, ensure that the widths are at # least as long as the title length if ($c <= $#$titles) { if (defined($titles->[$c])) { # If we're in HTML mode, get the length of the plaintext $len = length($showmode eq 'html' ? &PlainText($titles->[$c]) # else, use raw text. : $titles->[$c]); } else { $len = length("Field_$c"); } $widths[$c] = $len if $widths[$c] < $len; $max_widths[$c] = $len if $max_widths[$c] < $len; } } } &$row_sub(1); # reset the pointer for the next scan } else { # Use title width as default if original width is null or zero my $len; for ($c = 0; $c <= $#widths; $c++) { next unless $c <= $#$titles; # Get the length of the title (sans HTML text if in that mode) $len = length($showmode eq 'html' ? &PlainText($titles->[$c]) : $titles->[$c]); $widths[$c] = $len if $widths[$c] < $len; } # Can't scan the data, so the maximums can only be set by using the # explicit widths. @max_widths = @widths; } ($num_cols, \@widths, \@precision, \@max_widths); } ############################## =head1 putcell S< >I<$wrapped> = B<&putcell>( I<\@cells>, I<$c>, I<$cell_width>, I<\@prefix>, I<\@suffix>, I<$wrap_flag> ); Output the contents of an array cell at I<$cell>[I<$c>], causing text longer than I<$cell_width> to be saved for output on subsequent calls. Prefixing the output of each cell's value is a string from the two-element array I<@prefix>. Suffixing each cell's value is a string from the two-element array I<@suffix>. The first element of either array is selected when I<$wrap_flag> is zero or null, or when there is no more text in the current to be output. The second element is selected when I<$wrap_flag> is non-zero, and when there is more text in the current cell to be output. In the case of text longer than I<$cell_width>, a non-zero value is returned. Cells with undefined data are not output, nor are the prefix or suffix strings. =cut sub putcell { my $cells = shift; # ref to cell array my $c = shift; # index my $cell_width = shift; # maximum width of the cell my $prefix = shift; # 2-elt array of prefix strings my $suffix = shift; # 2-elt array of suffix strings my $wrap_flag = shift; # non-zero for wrapped lines my $fmt = sprintf("%%s%%-%ds%%s",$cell_width); my $more; my $v = $cells->[$c]; # get the data my $px = 0; # prefix index my $sx = 0; # suffix index if (defined $v) { # not undef data? my $text = $v; # save the text $cell_width = 1 unless $cell_width > 0; # be sane if ($cell_width <= length($text)) { $more = substr($text,$cell_width); $v = substr($text,0,$cell_width); } else { $v = $text; $more = ''; } # wrapping? if ($more ne '' && # See if we can wrap on a word boundary, instead of # arbitrarily splitting one; note that we try to not # split grouped numbers (1,345) or reals (1.234). $v =~ /([-,;? \t])([^-,;? \t0-9]*)$/ && # but also make sure that it is not too long length($2) <= $List_Wrap_Margin ) { # Okay, cut on the word boundary, leaving the break char # on the tail end of the current output value my $cut = $cell_width - length($2); $v = substr($text,0,$cut); # get new value $more = substr($text, $cut); # new remainder } $cells->[$c] = $more; # leave the rest for later $px = $wrap_flag != 0 && length($v) > 0; $sx = length($more) > 0; } put $fmt,$prefix->[$px],$v,$suffix->[$sx]; # output something (could be blanks) $sx; # leave wrapped flag } ############################## =head1 center Center a string within a given width. S< >I<$field> = B<center> I<$string>, I<$width>; =cut sub center { my($string,$width) = @_; $width = 0 if !defined($width); return $string if length($string) >= $width; my($pad) = int(($width - length($string))/2); # pad left half my($center) = (' ' x $pad) . $string; $pad = $width - length($center); $center .= ' ' x $pad; # pad right half $center; # return with the centered string } ############################## =head1 max Compute the maximum value from a list of values. S< >I<$max> = B<&max>( I<@values> ); =cut sub max { my ($max) = shift; foreach (@_) { $max = $_ if $max < $_; } $max; } ############################## =head1 min Compute the minum value from a list of values. S< >I<$min> = B<&min>( I<@values> ); =cut sub min { my ($min) = shift; foreach (@_) { $min = $_ if $min > $_; } $min; } ############################## =head1 max_length Compute the maximum length of a set of strings in an array reference. S< >I<$maxlength> = B<&max_length>( I<\@array_ref> ); =cut sub max_length { my($aref) = shift; my(@lens) = map { length } @$aref; my($maxlen) = max( @lens ); $maxlen; } ############################## =head1 htmltext Translate regular text for output into an HTML document. This means certain characters, such as "&", ">", and "<" must be escaped. S< >I<$output> = B<&htmltext>( I<$input> [, I<$allflag> ] ); If I<$allflag> is non-zero, then all characters are escaped. Normally, only the four HTML syntactic break characters are escaped. =cut # htmltext -- translate special text into HTML esacpes sub htmltext { local($_) = shift; my $all = shift; return undef unless defined($_); s/&(?!(?:amp|quot|gt|lt|#\d+);)/&amp;/g; s/\"/&quot;/g; s/>/&gt;/g; s/</\&lt;/g; if ($all) { s/ /\&#32;/g; s/\t/\&#09;/g; } $_; } ############################## =head1 out Print text followed by a newline. S< >B<out> I<$fmt> [, I<@text> ]; =cut sub out { my $fmt = shift; $fmt .= "\n" unless $fmt =~ /\n$/; printf STDOUT $fmt, @_; } ############################## =head1 put Print text (without a trailing newline). S< >B<out> I<$fmt> [, I<@text> ]; =cut sub put { printf STDOUT @_; } ############################## =head1 AUTHOR Alan K. Stebbens <aks@sgi.com> =cut =head1 BUGS =over 10 =item * Embedded HTML is how the user can insert formatting overrides. However, the HTML formatting techniques have not been given much consideration -- feel free to provide constructive feedback. =cut # 1;