diff options
| author | William Waites <wwaites@inf.ed.ac.uk> | 2020-02-26 17:08:26 +0300 |
|---|---|---|
| committer | William Waites <wwaites@inf.ed.ac.uk> | 2020-02-26 17:08:26 +0300 |
| commit | 696a5d98339b1b1e41480bed904e80355c49b947 (patch) | |
| tree | 9a520199abee44cba7553f0ba4d83729879d340b /scripts/ems/support | |
| parent | 22923ddcf049c694ac75158ead1613ea77ed18b1 (diff) | |
flag to turn off sentence splitter from emitting <P>
Diffstat (limited to 'scripts/ems/support')
| -rwxr-xr-x | scripts/ems/support/split-sentences.perl | 5 |
1 files changed, 4 insertions, 1 deletions
diff --git a/scripts/ems/support/split-sentences.perl b/scripts/ems/support/split-sentences.perl index 240195c7c..b6e9e2456 100755 --- a/scripts/ems/support/split-sentences.perl +++ b/scripts/ems/support/split-sentences.perl @@ -23,6 +23,7 @@ my $is_cjk = 0; my $QUIET = 0; my $HELP = 0; my $LIST_ITEM = 0; +my $NOP = 0; while (@ARGV) { $_ = shift; @@ -31,6 +32,7 @@ while (@ARGV) { /^-q$/ && ($QUIET = 1, next); /^-h$/ && ($HELP = 1, next); /^-i$/ && ($LIST_ITEM = 1, next); + /^-n$/ && ($NOP = 1, next); /^-b$/ && ($|++, next); # no output buffering } @@ -40,6 +42,7 @@ if ($HELP) { print "-b: no output buffering (for use in bidirectional pipes)\n"; print "-p: use a custom prefix file, overriding the installed one\n"; print "-i: avoid splitting on list items (e.g. 1. This is the first)\n"; + print "-n: do not emit <P> after paragraphs\n"; exit; } if (!$QUIET) { @@ -89,7 +92,7 @@ while (<STDIN>) { if (/^<.+>$/ || /^\s*$/) { # Time to process this block; we've hit a blank or <p> &do_it_for($text, $_); - print "<P>\n" if (/^\s*$/ && $text); ## If we have text followed by <P> + print "<P>\n" if $NOP == 0 && (/^\s*$/ && $text); ## If we have text followed by <P> $text = ""; } else { |