From ead2c41bb34f240d4ae0796e27ef61602f9f1c1b Mon Sep 17 00:00:00 2001
From: lars <lars@systemausfall.org>
Date: Tue, 18 Aug 2009 09:30:58 +0000
Subject: [PATCH] OSM QA tools: * fixed parsing of number of inhabitants with
 html entities * fixed wikipedia name of Grabow

---
 osm/qa/sparsely_mapped/generate_unmapped_places.sh | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/osm/qa/sparsely_mapped/generate_unmapped_places.sh b/osm/qa/sparsely_mapped/generate_unmapped_places.sh
index 7d3b284..1f91f1c 100755
--- a/osm/qa/sparsely_mapped/generate_unmapped_places.sh
+++ b/osm/qa/sparsely_mapped/generate_unmapped_places.sh
@@ -36,6 +36,7 @@ Dmitz	D%C3%B6mitz
 Friedland	Friedland_(Mecklenburg)
 FrstenbergHavel	F%C3%BCrstenberg/Havel
 Goldberg	Goldberg_(Mecklenburg)
+Grabow	Grabow_(Elde)
 Malchow	Malchow_(Mecklenburg)
 OstseebadRerik	Rerik
 RbelMritz	R%C3%B6bel
@@ -56,11 +57,12 @@ get_place_data() {
 
 get_place_size() {
 	# call "zcat -f" to uncompress a page if necessary
+	# ignore stuff like "&#160;" (space) within the number
 	wget --quiet --output-document - "$WIKIPEDIA_URL/$1" \
 			| zcat -f \
 			| grep -A 1 "^<td>Einwohner:</td>$" \
 			| tail -1 | cut -d ">" -f 2 | cut -d "<" -f 1 \
-			| sed s/[^0-9]//g
+			| sed -r 's/&#[0-9]{3};//g' | sed s/[^0-9]//g
 }