From cda6afbc5c9b2fc2d84a0d96daf166b8c28e379c Mon Sep 17 00:00:00 2001
From: "Troy A. Griffitts" <scribe@crosswire.org>
Date: Thu, 25 May 2023 08:17:14 +0000
Subject: added regularization and sigma normalization

git-svn-id: https://www.crosswire.org/svn/sword-tools/trunk@563 07627401-56e2-0310-80f4-f8cd0041bdcd
---
 migratetags/matchers/gntmatcher.h | 13 ++++++++++++-
 1 file changed, 12 insertions(+), 1 deletion(-)

diff --git a/migratetags/matchers/gntmatcher.h b/migratetags/matchers/gntmatcher.h
index d2fcbd8..35c245d 100644
--- a/migratetags/matchers/gntmatcher.h
+++ b/migratetags/matchers/gntmatcher.h
@@ -1,15 +1,21 @@
 #include "matcher.h"
 #include <utf8greekaccents.h>
+#include <map>
 
 #ifndef gntmatcher_h
 #define gntmatcher_h
 
+using std::map;
+
 class GNTMatcher : public Matcher {
 	UTF8GreekAccents sanitizeGreekAccentFilter;
+	map<SWBuf, SWBuf> globalRegs;
 public:
 
 	GNTMatcher() : sanitizeGreekAccentFilter() {
 		sanitizeGreekAccentFilter.setOptionValue("off");
+		globalRegs["ΘΣ"] = "ΘΕΟΣ";
+		globalRegs["ΚΥ"] = "ΚΥΡΙΟΥ";
 	}
 
 // Compares 2 words and tries to give a percentage assurance of a match
@@ -109,7 +115,6 @@ virtual SWBuf sanitizeWord(const SWBuf &word) {
 	SWBuf t1 = word;
 	// remove greek accents
 	sanitizeGreekAccentFilter.processText(t1);
-	t1.toUpper();
 
 	// remove ignoreSeries characters
 	SWBuf o = t1;
@@ -122,8 +127,14 @@ virtual SWBuf sanitizeWord(const SWBuf &word) {
 		SWBuf checkChar;
 		getUTF8FromUniChar(ch, &checkChar);
 		if (checkChar != " " && strstr(ignoreSeries, checkChar.c_str())) continue;
+		if (checkChar == "ϲ") checkChar = "σ";
+		if (checkChar == "ς") checkChar = "σ";
 		t1.append(checkChar);
 	}
+	t1.toUpper();
+	if (globalRegs.find(t1) != globalRegs.end()) {
+		t1 = globalRegs[t1];
+	}
 	return t1;
 }
 
-- 
cgit