mirrors/zotero
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages 1 Wiki Activity Actions

Search result scraping for GEAC catalogs

Browse source
This commit is contained in:
Simon Kornblith 2006-06-23 21:27:32 +00:00
parent 2b58ead7aa
commit 6f19b215f5
1 changed files with 27 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

30
scrapers.sql
View file

@ -1382,18 +1382,39 @@ for(i in uris) {
wait();'); wait();');
REPLACE INTO "scrapers" VALUES('c0e6fda6-0ecd-e4f4-39ca-37a4de436e15', '2006-06-18 11:19:00', 'GEAC Scraper', 'Simon Kornblith', '/(?:Geac)?FETCH[\:\?].*[&:]next=html/(?:record\.html|geacnffull\.html)', NULL, REPLACE INTO "scrapers" VALUES('c0e6fda6-0ecd-e4f4-39ca-37a4de436e15', '2006-06-18 11:19:00', 'GEAC Scraper', 'Simon Kornblith', '/(?:GeacQUERY|(?:Geac)?FETCH[\:\?].*[&:]next=html/(?:record\.html|geacnffull\.html))', NULL,
'var prefixRDF = ''http://www.w3.org/1999/02/22-rdf-syntax-ns#''; 'var prefixRDF = ''http://www.w3.org/1999/02/22-rdf-syntax-ns#'';
var prefixDC = ''http://purl.org/dc/elements/1.1/''; var prefixDC = ''http://purl.org/dc/elements/1.1/'';
var prefixDCMI = ''http://purl.org/dc/dcmitype/''; var prefixDCMI = ''http://purl.org/dc/dcmitype/'';
var prefixDummy = ''http://chnm.gmu.edu/firefox-scholar/''; var prefixDummy = ''http://chnm.gmu.edu/firefox-scholar/'';
var uri = doc.location.href; var uri = doc.location.href;
var uris = new Array();
if(uri.indexOf("/GeacQUERY") > 0) {
var items = utilities.getItemArray(doc, doc, "(?:Geac)?FETCH[\:\?].*[&:]next=html/(?:record\.html|geacnffull\.html)");
items = utilities.selectItems(items);
if(!items) {
return true;
}
var uris = new Array();
for(i in items) {
var newUri = i.replace(/([:&])next=html\/geacnffull.html/, "$1next=html/marc.html");
newUri = newUri.replace(/([:&])next=html\/record.html/, "$1next=html/marc.html");
uris.push(newUri);
}
} else {
var newUri = uri.replace(/([:&])next=html\/geacnffull.html/, "$1next=html/marc.html"); var newUri = uri.replace(/([:&])next=html\/geacnffull.html/, "$1next=html/marc.html");
newUri = newUri.replace(/([:&])next=html\/record.html/, "$1next=html/marc.html"); newUri = newUri.replace(/([:&])next=html\/record.html/, "$1next=html/marc.html");
uris.push(newUri);
}
utilities.loadDocument(newUri, browser, function(newBrowser) { utilities.processDocuments(browser, null, uris, function(newBrowser) {
newDoc = newBrowser.contentDocument; var newDoc = newBrowser.contentDocument;
var uri = newDoc.location.href;
var namespace = newDoc.documentElement.namespaceURI; var namespace = newDoc.documentElement.namespaceURI;
var nsResolver = namespace ? function(prefix) { var nsResolver = namespace ? function(prefix) {
@ -1436,8 +1457,7 @@ utilities.loadDocument(newUri, browser, function(newBrowser) {
} }
utilities.importMARCRecord(record, uri, model); utilities.importMARCRecord(record, uri, model);
done(); }, function() { done(); }, function() {});
}, function() {});
wait();'); wait();');