From 7d2fd2bf643671377e990b1c944aa3650397e3da Mon Sep 17 00:00:00 2001
From: Patrick Simianer <patrick@lilt.com>
Date: Mon, 9 Mar 2020 17:48:55 +0000
Subject: de-sgm: match more stuff

---
 de-sgm | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

(limited to 'de-sgm')

diff --git a/de-sgm b/de-sgm
index 452edfe..fd4546e 100755
--- a/de-sgm
+++ b/de-sgm
@@ -1,10 +1,13 @@
 #!/bin/sh
 
-egrep -v "^[[:space:]]*(<\?xml.*\?>|</?(mteval|doc|srcset|refset|translator|reviewer)[^>]*>)[[:space:]]*$" \
-  | egrep -v "^[[:space:]]*<(url|description|keywords|talkid|title|translator|reviewer)[^>]*>.*</(url|description|keywords|talkid|title|translator|reviewer)>[[:space:]]*$" \
+egrep -v -i "^[[:space:]]*(<\?xml.*\?>|</?(mteval|doc|srcset|refset|translator|reviewer)[^>]*>)[[:space:]]*$" \
+  | egrep -v -i "^[[:space:]]*<(url|description|keywords|talkid|title|translator|reviewer)[^>]*>.*</(url|description|keywords|talkid|title|translator|reviewer)>[[:space:]]*$" \
   | sed "s|<seg[^>]*>\s*||" \
   | sed "s|\s*</seg>\s*$||" \
-  | egrep -v "^[[:space:]]*<p>[[:space:]]*$|^[[:space:]]*</p>[[:space:]]*$" \
+  | egrep -v -i "^[[:space:]]*<p>[[:space:]]*$|^[[:space:]]*</p>[[:space:]]*$" \
   | sed "s|<speaker>\s*||" \
-  | sed "s|\s*</speaker>\s*$||"
+  | sed "s|\s*</speaker>\s*$||" \
+  | sed "s|\s*<hl>\s*$||" \
+  | sed "s|\s*</hl>\s*$||" \
+  | grep -v -P "^\s*$"
 
-- 
cgit v1.2.3