Adds a translator for Informaworld.

This commit is contained in:
Michael Berkowitz 2008-02-01 16:54:50 +00:00
parent 63f63de37a
commit 27e0887588

View File

@ -22,7 +22,7 @@
-- Set the following timestamp to the most recent scraper update date
REPLACE INTO version VALUES ('repository', STRFTIME('%s', '2008-01-31 23:00:00'));
REPLACE INTO version VALUES ('repository', STRFTIME('%s', '2008-02-01 18:00:00'));
REPLACE INTO translators VALUES ('96b9f483-c44d-5784-cdad-ce21b984fe01', '1.0.0b4.r1', '', '2007-06-21 20:00:00', '1', '100', '4', 'Amazon.com', 'Sean Takats', '^https?://(?:www\.)?amazon',
'function detectWeb(doc, url) {
@ -2088,6 +2088,101 @@ function getData(ids){
Zotero.wait();
}');
REPLACE INTO translators VALUES ('1885b93c-cf37-4b25-aef5-283f42eada9d', '1.0.0b4.r5', '', '2008-02-01 18:00:00', '0', '100', '4', 'Informaworld', 'Michael Berkowitz', 'http://www.informaworld.com',
'function detectWeb(doc, url) {
if (url.indexOf("quicksearch") != -1) {
return "multiple";
} else if (doc.evaluate(''//a[substring(text(), 2, 8) = "Download"]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext()) {
if (doc.evaluate(''//div[@id="metahead"]/div/strong[text() = "Published in:"]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext()) {
var pubtype = doc.evaluate(''//img[substring(@title, 1, 17) = "Publication type:"]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext().title;
if (pubtype.match("journal")) {
return "journalArticle";
} else if (pubtype.match("book")) {
return "bookSection";
}
} else {
return "book";
}
} else if (url.indexOf("content=g") != -1 ||
doc.evaluate(''//div[@id="browse"]//tbody/tr/td[2]/a[2]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext() ||
doc.evaluate(''//div[@id="title"]//td[2]/div/strong/a'', doc, null, XPathResult.ANY_TYPE, null).iterateNext()) {
return "multiple";
} else {
return true;
}
}
',
'function doWeb(doc, url) {
var links = new Array();
if (detectWeb(doc, url) == "multiple") {
var items = new Object();
if (doc.evaluate(''//div[@id="quicksearch"]//tr/td/b/a'', doc, null, XPathResult.ANY_TYPE, null).iterateNext()) {
var xpath = ''//div[@id="quicksearch"]//tr/td/b/a'';
} else if (doc.evaluate(''//div[@id="title"]/table/tbody/tr[2]//strong/a'', doc, null, XPathResult.ANY_TYPE, null).iterateNext()) {
var xpath = ''//div[@id="title"]/table/tbody/tr[2]//strong/a'';
} else if (doc.evaluate(''//div[@id="browse"]//tbody/tr/td[2]/a[2]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext()) {
var xpath = ''//div[@id="browse"]//tbody/tr/td[2]/a[2]'';
} else if (doc.evaluate(''//div[@id="title"]//td[2]/div/strong/a'', doc, null, XPathResult.ANY_TYPE, null).iterateNext()) {
var xpath = ''//div[@id="title"]//td[2]/div/strong/a'';
}
var titles = doc.evaluate(xpath, doc, null, XPathResult.ANY_TYPE, null);
var title = titles.iterateNext();
while (title) {
items[title.href] = title.textContent;
title = titles.iterateNext();
}
items = Zotero.selectItems(items);
for (var i in items) {
links.push(i);
}
} else {
links = [url];
}
Zotero.debug(links);
Zotero.Utilities.processDocuments(links, function(newDoc) {
var xpath = ''//div[@id="metahead"]/div'';
var stuff = newDoc.evaluate(xpath, newDoc, null, XPathResult.ANY_TYPE, null);
var thing = stuff.iterateNext() ;
while (thing) {
if (thing.textContent.match(/DOI/)) {
var doi = Zotero.Utilities.trimInternal(thing.textContent).match(/:\s+(.*)/)[1];
}
thing = stuff.iterateNext();
}
var id = newDoc.location.href.match(/content=([\w\d]+)/);
var post = ''tab=citation&selecteditems='' + id[1].substr(1) + ''&content='' + id[1] + ''&citstyle=refworks&showabs=false&format=file'';
Zotero.Utilities.HTTP.doPost(''http://www.informaworld.com/smpp/content'', post, function(text) {
text = text.replace(/RT/, "TY");
text = text.replace(/VO/, "VL");
text = text.replace(/LK/, "UR");
text = text.replace(/YR/, "PY");
text = text.replace(/([A-Z][A-Z\d]\s)/g, "$1 - ")
var translator = Zotero.loadTranslator("import");
translator.setTranslator("32d59d2d-b65a-4da4-b0a3-bdd3cfb979e7");
translator.setString(text);
translator.setHandler("itemDone", function(obj, item) {
var type = text.match(/TY\s+\-\s+([^\n]*)/)[1];
Zotero.debug(type);
if (type == "Journal") {
item.itemType = "journalArticle";
} else if (type == "Book, Whole") {
item.itemType = "book";
} else if (type == "Book, Section") {
item.itemType = "bookSection";
}
if (doi) {
item.DOI = doi;
}
item.complete();
});
translator.translate();
});
}, function() {Zotero.done;});
}');
REPLACE INTO translators VALUES ('f880bf79-d42f-4337-b0d2-7a7de4a48b7d', '1.0.0b4.r5', '', '2008-01-31 20:00:00', '0', '100', '6', 'Library Catalog (X-OPAC)', 'Michael Berkowitz', '(xopac|hylib)',
'function detectWeb(doc, url) {
if (url.indexOf("&nd=") != -1) {