From: Tara L Andrews <tla\@mit.edu>
Date: Fri, 3 Feb 2012 19:54:45 +0000 (+0100)
Subject: post-processing script to get rid of punctuation
X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=commitdiff_plain;h=6f0ec5df2f4fb5ee38786f98a81439fcf8eab3ba;p=scpubgit%2Fstemmatology.git

post-processing script to get rid of punctuation
---

diff --git a/script/strip_punctuation.pl b/script/strip_punctuation.pl
new file mode 100755
index 0000000..2b4ec62
--- /dev/null
+++ b/script/strip_punctuation.pl
@@ -0,0 +1,38 @@
+#!/usr/bin/env perl
+
+use lib 'lib';
+use strict;
+use warnings;
+use Text::Tradition::Directory;
+
+my( $dsn, $user, $pass ) = @ARGV;
+
+my $connect_args = { dsn => $dsn };
+$connect_args->{'extra_args'} = { user => $user, password => $pass }
+	if $user && $pass;
+my $dir = Text::Tradition::Directory->new( $connect_args );
+
+foreach my $id ( $dir->tradition_ids ) {
+	my $scope = $dir->new_scope;
+	my $tradition = $dir->lookup( $id );
+	print STDERR "Processing tradition " . $tradition->name . "\n";
+	foreach my $reading ( $tradition->collation->readings ) {
+		$reading->alter_text( strip_punct( $reading->text ) );
+	}
+	$tradition->collation->flatten_ranks;
+	$dir->save( $tradition );
+}
+
+print STDERR "Done\n";
+
+sub strip_punct {
+	my( $rtext ) = @_;
+	my $orig_r = $rtext;
+	return $rtext unless $rtext =~ /\w/;
+	$rtext =~ s/^\W+//;
+	$rtext =~ s/\W+$//;
+	print STDERR "Altering $orig_r to $rtext\n"
+		unless $orig_r eq $rtext;
+	return $rtext;
+}
+	
\ No newline at end of file