suzanne.soy/zotero
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

-Adds European Educational Research Journal translator.

Browse Source
This commit is contained in:
Michael Berkowitz 2008-05-01 19:43:33 +00:00
parent 5a4bbd2651
commit d976f40125
1 changed files with 76 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

77
scrapers.sql
View File

@ -22,7 +22,7 @@
-- Set the following timestamp to the most recent scraper update date
REPLACE INTO version VALUES ('repository', STRFTIME('%s', '2008-05-01 16:30:00'));
REPLACE INTO version VALUES ('repository', STRFTIME('%s', '2008-05-01 20:30:00'));
REPLACE INTO translators VALUES ('96b9f483-c44d-5784-cdad-ce21b984fe01', '1.0.0b4.r1', '', '2008-03-21 20:00:00', '1', '100', '4', 'Amazon.com', 'Sean Takats and Michael Berkowitz', '^https?://(?:www\.)?amazon',
'function detectWeb(doc, url) {
@ -1089,6 +1089,81 @@ REPLACE INTO translators VALUES ('88915634-1af6-c134-0171-56fd198235ed', '1.0.0b
Zotero.wait();
}');
REPLACE INTO translators VALUES ('2e304579-dd7b-4770-85e9-0d724c9b49a5', '1.0.0b4.r5', '', '2008-05-01 20:30:00', '0', '100', '4', 'European Educational Research Journal', 'Michael Berkowitz', 'http://www.wwwords.co.uk/eerj/',
'function detectWeb(doc, url) {
if (doc.evaluate(''//div[@id="maincontent"]/table[*//p[@class="articletitle"]]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext()) {
return "multiple";
}
}',
'function titleCase(str) {
var skipWords = ["but", "or", "yet", "so", "for", "and", "nor", "a", "an", "the", "at", "by", "from", "in", "into", "of", "on", "to", "with", "up", "down", "as"];
var words = str.toLowerCase().split(/\s+/);
var newstr = "";
for each (var word in words) {
if (skipWords.indexOf(word.replace(/[^a-zA-Z]+/, "")) != -1) {
newstr += " " + word;
} else if (word.indexOf("-") != -1) {
newword = word.split("-");
newstr += " " + newword[0][0].toUpperCase() + newword[0].substr(1) + "-" + newword[1][0].toUpperCase() + newword[1].substr(1);
} else {
newstr += " " + word[0].toUpperCase() + word.substr(1);
}
}
return Zotero.Utilities.trimInternal(newstr);
}
function doWeb(doc, url) {
var namespace = doc.documentElement.namespaceURI;
var nsResolver = namespace ? function(prefix) {
if (prefix == ''x'') return namespace; else return false;
} : null;
var items = new Object();
var titles = doc.evaluate(''//p[@class="articletitle"]'', doc, nsResolver, XPathResult.ANY_TYPE, null);
var title;
while (title = titles.iterateNext()) {
var text = Zotero.Utilities.trimInternal(title.textContent);
items[text] = text;
}
items = Zotero.selectItems(items);
Zotero.debug(items);
var articles = doc.evaluate(''//div[@id="maincontent"]/table[*//p[@class="articletitle"]]'', doc, nsResolver, XPathResult.ANY_TYPE, null);
var art;
while (art = articles.iterateNext()) {
var title = Zotero.Utilities.trimInternal(doc.evaluate(''.//p[@class="articletitle"]'', art, nsResolver, XPathResult.ANY_TYPE, null).iterateNext().textContent);
if (items[title]) {
var pdfurl = doc.evaluate(''.//a[contains(text(), "FULL TEXT")]'', art, nsResolver, XPathResult.ANY_TYPE, null).iterateNext().href;
var item = new Zotero.Item("journalArticle");
item.publicationTitle = "European Educational Research Journal";
item.ISSN = "1474-9041";
item.url = url;
item.title = title;
var voliss = doc.title.match(/\-\s+(.*)$/)[1];
voliss = voliss.match(/Volume\s+(\d+)\s+Issue\s+(\d+)\s+\((\d+)\)/);
item.volume = voliss[1];
item.issue = voliss[2];
item.date = voliss[3];
var authors = doc.evaluate(''.//tr[2]/td'', art, nsResolver, XPathResult.ANY_TYPE, null).iterateNext().textContent;
var ibits = doc.evaluate(''.//tr[2]/td//i'', art, nsResolver, XPathResult.ANY_TYPE, null);
var ibit = "";
var bit;
while (bit = ibits.iterateNext()) {
authors = authors.replace(bit.textContent, ",");
}
authors = authors.split(/\s*(,|&)\s*/);
for each (var aut in authors) {
if (aut.match(/\w/)) {
aut = titleCase(Zotero.Utilities.trimInternal(aut));
item.creators.push(Zotero.Utilities.cleanAuthor(aut, "author"));
}
}
item.attachments = [{url:pdfurl, title:"EERJ Full Text PDF", mimeType:"application/pdf"}];
item.complete();
}
}
}');
REPLACE INTO translators VALUES ('bdaac15c-b0ee-453f-9f1d-f35d00c7a994', '1.0.0b4.r5', '', '2008-05-01 16:30:00', '0', '100', '4', 'AMS Journals', 'Michael Berkowitz', 'http://www.ams.org/',
'function detectWeb(doc, url) {
if (url.match(/jour(nals|search)/)) {