protected ArticleFiles processUrl(CachedUrl cu, Matcher mat) {
ArticleFiles af = new ArticleFiles();
af.setFullTextCu(cu);
af.setRoleCu(ArticleFiles.ROLE_ABSTRACT, cu);
// XXX Full text PDF link embedded in page, cannot guess URL
return af;
}
public void testArticleCountAndType() throws Exception {
int expCount = 28;
PluginTestUtil.crawlSimAu(sau);
String pat1 = "branch(\\d+)/(\\d+file\\.html)";
String rep1 = "aps/journal/v123/n$1/full/$2";
PluginTestUtil.copyAu(sau, nau, ".*[^.][^p][^d][^f]$", pat1, rep1);
String pat2 = "branch(\\d+)/(\\d+file\\.pdf)";
String rep2 = "aps/journal/v123/n$1/pdf/$2";
PluginTestUtil.copyAu(sau, nau, ".*\\.pdf$", pat2, rep2);
// Remove some URLs
int deleted = 0;
for (Iterator it = nau.getAuCachedUrlSet().contentHashIterator(); it.hasNext(); ) {
CachedUrlSetNode cusn = (CachedUrlSetNode) it.next();
if (cusn instanceof CachedUrl) {
CachedUrl cu = (CachedUrl) cusn;
String url = cu.getUrl();
if (url.contains("/journal/")
&& (url.endsWith("1file.html") || url.endsWith("2file.pdf"))) {
deleteBlock(cu);
++deleted;
}
}
}
assertEquals(8, deleted);
Iterator<ArticleFiles> it = nau.getArticleIterator();
int count = 0;
int countHtmlOnly = 0;
int countPdfOnly = 0;
while (it.hasNext()) {
ArticleFiles af = it.next();
log.info(af.toString());
CachedUrl cu = af.getFullTextCu();
String url = cu.getUrl();
assertNotNull(cu);
String contentType = cu.getContentType();
log.debug("count " + count + " url " + url + " " + contentType);
count++;
if (af.getRoleUrl(ArticleFiles.ROLE_FULL_TEXT_PDF) == null) {
++countHtmlOnly;
}
if (af.getRoleUrl(ArticleFiles.ROLE_FULL_TEXT_PDF) == url) {
++countPdfOnly;
}
}
log.debug("Article count is " + count);
assertEquals(expCount, count);
assertEquals(4, countHtmlOnly);
assertEquals(4, countPdfOnly);
}
/*
* In order to find full text PDF you need to find the citation_pdf_url meta tag in the
* abstract html pull out the pdf url normalize it (reorder params...) and find the matching
* cached URL
*/
protected ArticleFiles processAbstract(CachedUrl absCu, Matcher absMat) {
NodeList nl = null;
ArticleFiles af = new ArticleFiles();
if (absCu != null && absCu.hasContent()) {
// TEMPORARY: set absCU as default full text CU in case there is
// no PDF CU with content; the current metadata manager currently
// uses only the full text CU, but this will change with the new
// metadata schema that can have multiple CUs for an article.
af.setFullTextCu(absCu);
af.setRoleCu(ArticleFiles.ROLE_ABSTRACT, absCu);
try {
InputStreamSource is =
new InputStreamSource(new Stream(absCu.getUnfilteredInputStream()));
Page pg = new Page(is);
Lexer lx = new Lexer(pg);
Parser parser = new Parser(lx);
Lexer.STRICT_REMARKS = false;
NodeFilter nf =
new NodeFilter() {
public boolean accept(Node node) {
if (!(node instanceof MetaTag)) return false;
MetaTag meta = (MetaTag) node;
if (!"citation_pdf_url".equalsIgnoreCase(meta.getMetaTagName())) return false;
return true;
}
};
nl = parser.extractAllNodesThatMatch(nf);
} catch (ParserException e) {
log.debug("Unable to parse abstract page html", e);
} catch (UnsupportedEncodingException e) {
log.debug("Bad encoding in abstact page html", e);
} finally {
absCu.release();
}
}
try {
if (nl != null) {
if (nl.size() > 0) {
// minimally encode URL to prevent URL constructor
// from stripping trailing spaces
String pdfUrlStr = ((MetaTag) nl.elementAt(0)).getMetaContent();
URL pdfUrl = new URL(UrlUtil.minimallyEncodeUrl(pdfUrlStr));
List<String> paramList = new ArrayList<String>();
paramList.add("fileType");
paramList.add("fileId");
paramList.add("fileName");
pdfUrl = reArrangeUrlParams(pdfUrl, paramList);
if (!pdfUrl.getHost().startsWith("www.")) {
pdfUrl = new URL(pdfUrl.getProtocol(), "www." + pdfUrl.getHost(), pdfUrl.getFile());
}
// note: must leave URL encoded because that's how we store URLs
CachedUrl pdfCu = au.makeCachedUrl(pdfUrl.toString());
if (pdfCu != null && pdfCu.hasContent()) {
// replace absCU with pdfCU if exists and has content
af.setFullTextCu(pdfCu);
af.setRoleCu(ArticleFiles.ROLE_FULL_TEXT_PDF, pdfCu);
}
}
}
} catch (MalformedURLException e) {
log.debug("Badly formatted pdf url link", e);
} catch (IllegalArgumentException e) {
log.debug("Badly formatted pdf url link", e);
}
return af;
}