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

-ACM translator should more consistently save abstracts.

Browse Source
This commit is contained in:
Michael Berkowitz 2008-05-19 16:15:20 +00:00
parent b50126776f
commit ec46106617
1 changed files with 6 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
scrapers.sql
View File

@ -22,7 +22,7 @@
-- Set the following timestamp to the most recent scraper update date -- Set the following timestamp to the most recent scraper update date
REPLACE INTO version VALUES ('repository', STRFTIME('%s', '2008-05-19 16:30:00')); REPLACE INTO version VALUES ('repository', STRFTIME('%s', '2008-05-19 17:30:00'));
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', 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) { 'function detectWeb(doc, url) {
@ -17021,7 +17021,7 @@ function doWeb(doc, url) {
} }
}'); }');
REPLACE INTO translators VALUES ('e78d20f7-488-4023-831-dfe39679f3f', '1.0.0b3.r1', '', '2008-03-04 20:00:00', '1', '100', '4', 'ACM', 'Simon Kornblith', 'https?://[^/]*portal\.acm\.org[^/]*/(?:results\.cfm|citation\.cfm)', REPLACE INTO translators VALUES ('e78d20f7-488-4023-831-dfe39679f3f', '1.0.0b3.r1', '', '2008-05-19 17:30:00', '1', '100', '4', 'ACM', 'Simon Kornblith and Michael Berkowitz', 'https?://[^/]*portal\.acm\.org[^/]*/(?:results\.cfm|citation\.cfm)',
'function detectWeb(doc, url) { 'function detectWeb(doc, url) {
if(url.indexOf("/results.cfm") != -1) { if(url.indexOf("/results.cfm") != -1) {
var items = Zotero.Utilities.getItemArray(doc, doc, ''^https?://[^/]+/citation.cfm\\?[^#]+$''); var items = Zotero.Utilities.getItemArray(doc, doc, ''^https?://[^/]+/citation.cfm\\?[^#]+$'');
@ -17057,10 +17057,11 @@ function scrape(doc) {
null).iterateNext().getAttribute("onClick"); null).iterateNext().getAttribute("onClick");
var m = onClick.match(/''([^'']+)''/); var m = onClick.match(/''([^'']+)''/);
var abstract = doc.evaluate(''//div[@class="abstract"]/p[@class="abstract"]'', doc, null, var abstract = doc.evaluate(''//div[@class="abstract"]/p[@class="abstract"]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext();
XPathResult.ANY_TYPE, null).iterateNext(); if (!abstract.textContent.match(/\w+/)) {
var abstract = doc.evaluate(''//div[@class="abstract"]/p[2]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext();
}
if(abstract) abstract = Zotero.Utilities.cleanString(abstract.textContent); if(abstract) abstract = Zotero.Utilities.cleanString(abstract.textContent);
var snapshot = doc.location.href; var snapshot = doc.location.href;
var attachments = new Array(); var attachments = new Array();