www

commit 2b0bebe7a462ae95c45653cf9594a18191b6c1bf
parent e5404f4938eaec2032a727cb9d19a25fa3f43ec4
Author: Simon Kornblith <simon@simonster.com>
Date:   Mon,  4 Sep 2006 18:16:50 +0000

closes #258, MARC translator should capitalize titles


Diffstat:
M
chrome/chromeFiles/content/scholar/xpcom/utilities.js
 | 
33
+++++++++++++++++++++++++++++++++
M
scrapers.sql
 | 
13
++++++++++---

2 files changed, 43 insertions(+), 3 deletions(-)
diff --git a/chrome/chromeFiles/content/scholar/xpcom/utilities.js b/chrome/chromeFiles/content/scholar/xpcom/utilities.js
@@ -133,6 +133,39 @@ Scholar.Utilities.prototype.itemTypeExists = function(type) {
 }
 
 /*
+ * Cleans a title, capitalizing the proper words and replacing " :" with ":"
+ */
+Scholar.Utilities.capitalizeSkipWords = ["but", "or", "yet", "so", "for", "and",
+"nor", "a", "an", "the", "at", "by", "from", "in", "into", "of", "on", "to",
+"with", "up", "down"];
+Scholar.Utilities.prototype.capitalizeTitle = function(title) {
+	title = title.replace(/ : /g, ": ");	
+	var words = title.split(" ");
+	
+	// always capitalize first
+	words[0] = words[0][0].toUpperCase() + words[0].substr(1);
+	if(words.length > 1) {
+		var lastWordIndex = words.length-1;
+		// always capitalize last
+		words[lastWordIndex] = words[lastWordIndex][0].toUpperCase() + words[lastWordIndex].substr(1);
+		
+		if(words.length > 2) {
+			for(var i=1; i<lastWordIndex; i++) {
+				// if not a skip word
+				if(Scholar.Utilities.capitalizeSkipWords.indexOf(words[i].toLowerCase()) == -1 ||
+				   words[i-1][words[i-1].length-1] == ":") {
+					words[i] = words[i][0].toUpperCase() + words[i].substr(1);
+				} else {
+					words[i] = words[i].toLowerCase();
+				}
+			}
+		}
+	}
+	
+	return words.join(" ");
+}
+
+/*
  * END SCHOLAR FOR FIREFOX EXTENSIONS
  */
 
diff --git a/scrapers.sql b/scrapers.sql
@@ -1,4 +1,4 @@
--- 74
+-- 75
 
 -- Set the following timestamp to the most recent scraper update date
 REPLACE INTO "version" VALUES ('repository', STRFTIME('%s', '2006-08-31 22:44:00'));
@@ -185,7 +185,7 @@ REPLACE INTO "translators" VALUES ('838d8849-4ffb-9f44-3d0d-aa8a0a079afe', '2006
 					if(title.substring(title.length-2) == " /") {
 						title = title.substring(0, title.length-2);
 					}
-					newItem.title = title;
+					newItem.title = Scholar.Utilities.capitalizeTitle(title);
 				} else if(match[1] == ''Author(s)'') {
 					var yearRegexp = /[0-9]{4}-([0-9]{4})?/;
 					
@@ -1029,7 +1029,7 @@ REPLACE INTO "translators" VALUES ('add7c71c-21f3-ee14-d188-caf9da12728b', '2006
 					newItem.ISBN = m[0];
 				} else if(field == "title") {
 					var titleParts = value.split(" / ");
-					newItem.title = titleParts[0];
+					newItem.title = Scholar.Utilities.capitalizeTitle(titleParts[0]);
 				} else if(field == "publication info") {
 					var pubParts = value.split(" : ");
 					newItem.place = pubParts[0];
@@ -3679,6 +3679,9 @@ function processOWC(doc) {
 				var spanTitle = spanTags[i].getAttribute("title");
 				var item = new Scholar.Item();
 				if(Scholar.Utilities.parseContextObject(spanTitle, item)) {
+					if(item.title) {
+						item.title = Scholar.Utilities.capitalizeTitle(item.title);
+					}
 					item.complete();
 					return true;
 				} else {
@@ -5934,6 +5937,10 @@ record.prototype.translate = function(item) {
 	this._associateDBField(item, "070", "ab", "callNumber");
 	this._associateDBField(item, "060", "ab", "callNumber");
 	this._associateDBField(item, "050", "ab", "callNumber");
+	
+	if(item.title) {
+		item.title = Scholar.Utilities.capitalizeTitle(item.title);
+	}
 }
 
 function doImport() {