From ebb1b1653f6f6ec67c9d7dfc521ec0981c60761e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Maty=C3=A1=C5=A1=20Kopp?= <kopp@ufal.mff.cuni.cz>
Date: Tue, 7 May 2024 15:48:51 +0200
Subject: [PATCH] inline media, fix CZ url #868

---
 Scripts/parlamint2teitok.pl | 19 ++++++++++++++++---
 1 file changed, 16 insertions(+), 3 deletions(-)

diff --git a/Scripts/parlamint2teitok.pl b/Scripts/parlamint2teitok.pl
index 7c584dbb7..47b5246de 100644
--- a/Scripts/parlamint2teitok.pl
+++ b/Scripts/parlamint2teitok.pl
@@ -117,7 +117,7 @@
 
 foreach $media ( $xml->findnodes("//media") ) {
   if ( $sname =~ /-CZ/ ) {
-    $mbase = "https://lindat.mff.cuni.cz/services/teitok/data/parczech/www.psp.cz/eknih/";
+    $mbase = "https://lindat.mff.cuni.cz/services/teitok/data/parczech/";
   };
   if ( $mbase ) {
     $media->setAttribute("url", $mbase.$media->getAttribute("url"));
@@ -162,6 +162,21 @@
   # Already pb'd
   if ( $verbose ) { print "Already pb'd - adding atts later"; };
   $dopb = 1;
+  # inline media if pb contain corresp
+  foreach $pb ( $xml->findnodes("//text//pb[\@corresp]") ) {
+  	print STDERR $pb->toString;
+    foreach $corresp (map {s/^#//;$_} split(/\s+/, $pb->getAttribute("corresp"))){
+    	print STDERR $corresp;
+      my ($audio) = $xml->findnodes("//media[\@xml:id = '$corresp']");
+      print STDERR $audio->toString;
+      if($audio){
+      	$audio->unbindNode();
+      	$pb->parentNode->insertAfter($audio,$pb);
+      }
+    };
+  };
+
+
 
 } else {
   $dcnt = 0; $pbcnt = 0; $scnt = 1;
@@ -398,8 +413,6 @@ ( $$$ )
           $pb->setAttribute("ana", $utt->getAttribute("ana"));
     if ( $debug ) { print $pb->toString; } ;
   };
-
-  # inline media if pb contain corresp
 };
 
 `mkdir -p $outf`;