[Header-description] Improved efficiency and capability of paragraph searching

1 files changed, 29 insertions, 14 deletions

diff --git a/awk-edits/Header-description.awk b/awk-edits/Header-description.awk
index bb245c6..645a1ac 100644
--- a/awk-edits/Header-description.awk
+++ b/awk-edits/Header-description.awk
@@ -7,30 +7,45 @@
 # either all of the text on the page, or only the first 200 characters.
 
 function appToDesc(l) {
-	gsub(/<[^>]*>/, "", l) ; gsub("&#8217;", "'", l)
-	gsub(/[^а-яА-Яa-zA-Z0-9 '.]/, "", l)
-	gsub(/^[ \t]*/, "", l) ; gsub(/[ \t]*$/, "", l)
+	l = substr(l, parstar, length(l) - parstar + 1 - parend);
+	gsub("<[^>]*>", "", l); gsub("&#8217;", "'", l); gsub(/&[^;]*;/, "", l);
+
+	if (desc == "") desc = l;
+	else            desc = desc " " l;
 
-	desc = desc l
 	if (length(desc) >= 200) {
-		desc = substr(desc, 1, 197) "..."
-		return 1
+		desc = substr(desc, 1, 197) "...";
+		return 1;
 	}
-	return 0
+	return 0;
 }
 
 BEGIN {
-	stop = 0
+	parcont = 0; # continue, we have found a <p> but not a </p>
+	parstar = 1; # we found a <p>
+	parends = 0; # we found a </p>
 
 	# Loop over the whole file, extract all text in first couple of paragraphs,
 	# store it in desc (but no more than 200 characters will be saved)
-	while(stop == 0 && getline < ARGV[1]) {
-		if (index($0, "<p") > 0)
-			stop = appToDesc($0)
+	while(getline < ARGV[1]) {
+		if (match($0, /<h[23456]/) && substr($0, RSTART, 17) != "<h2 id=\"contents\"")
+			break;
+
+		if (match($0, /^<p>/))
+			parstar = 4;
+		if (match($0, /<\/p>$/))
+			parends = 3;
+
+		if (parcont || parstar == 4) {
+			if (appToDesc($0)) break;
+			parcont = RSTART == 0;
+			parstar = 1;
+			parends = 0;
+		}
 	}
-	close(ARGV[1]) # FILENAME isn't set during BEGIN
+	close(ARGV[1]); # FILENAME isn't set during BEGIN
 }
 
-"\\[DESCRIPTION\\]" {
-	gsub("\\[DESCRIPTION\\]", desc)
+/\[DESCRIPTION\]/ {
+	sub(/\[DESCRIPTION\]/, desc);
 }