private void deleteBlock(CachedUrl cu) throws IOException {
log.info("deleting " + cu.getUrl());
CachedUrlSetSpec cuss = new SingleNodeCachedUrlSetSpec(cu.getUrl());
ArchivalUnit au = cu.getArchivalUnit();
CachedUrlSet cus = au.makeCachedUrlSet(cuss);
NodeManager nm = au.getPlugin().getDaemon().getNodeManager(au);
nm.deleteNode(cus);
}
public void testFunctionalFromTarHierarchy() throws Exception {
log.debug3("in testFromTarHierarchy");
// load the tarballs
InputStream file_input = null;
try {
file_input = getResourceAsStream(realTARFile_A);
// UrlCacher uc = au.makeUrlCacher(TAR_A_BASE);
// uc.storeContent(file_input, tarHeader);
UrlCacher uc = tarAu.makeUrlCacher(new UrlData(file_input, tarHeader, TAR_A_BASE));
uc.storeContent();
IOUtil.safeClose(file_input);
file_input = getResourceAsStream(realTARFile_B);
// uc = au.makeUrlCacher(TAR_B_BASE);
// uc.storeContent(file_input, tarHeader);
uc = tarAu.makeUrlCacher(new UrlData(file_input, tarHeader, TAR_B_BASE));
uc.storeContent();
IOUtil.safeClose(file_input);
} catch (IOException e) {
// TODO Auto-generated catch block
e.printStackTrace();
} finally {
IOUtil.safeClose(file_input);
}
CachedUrlSet cus = tarAu.getAuCachedUrlSet();
for (CachedUrl cu : cus.getCuIterable()) {
log.debug3("AU - cu is: " + cu.getUrl());
cu.release();
}
// We need to start from the level of the ArticleMetadataExtractor
MyListEmitter emitter = new MyListEmitter();
ArticleMetadataExtractor amEx =
new ElsevierDeferredArticleMetadataExtractor(ArticleFiles.ROLE_ARTICLE_METADATA);
Iterator<ArticleFiles> it = tarAu.getArticleIterator(MetadataTarget.Any());
while (it.hasNext()) {
ArticleFiles af = it.next();
log.debug3("Metadata test - articlefiles " + af.toString());
// CachedUrl cu = af.getFullTextCu();
CachedUrl cu = af.getRoleCu(ArticleFiles.ROLE_ARTICLE_METADATA);
log.debug3("metadata cu is " + cu.getUrl());
// List<ArticleMetadata> mdlist = mle.extract(MetadataTarget.Any(), cu);
amEx.extract(MetadataTarget.Any(), af, emitter);
List<ArticleMetadata> returnList = emitter.getAmList();
assertNotNull(returnList);
log.debug3("size of returnList is " + returnList.size());
Iterator<ArticleMetadata> mdIt = returnList.iterator();
ArticleMetadata mdRecord = null;
while (mdIt.hasNext()) {
mdRecord = (ArticleMetadata) mdIt.next();
validateCompleteMetadataRecord(mdRecord);
}
}
}
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);
}
protected ArticleFiles processFullTextPdf(CachedUrl cu, Matcher mat) {
ArticleFiles af = new ArticleFiles();
af.setRoleCu(ArticleFiles.ROLE_FULL_TEXT_PDF, cu);
String abstractUrl = mat.replaceAll("$1aspx");
CachedUrl abstractCu = cu.getArchivalUnit().makeCachedUrl(abstractUrl);
if (abstractCu.hasContent()) af.setFullTextCu(abstractCu);
else af.setFullTextCu(cu);
log.debug3("returning full text: " + af.getFullTextUrl());
return af;
}
@Override
public void extract(MetadataTarget target, CachedUrl cu, Emitter emitter) throws IOException {
ArticleMetadata am = new SimpleHtmlMetaTagMetadataExtractor().extract(target, cu);
am.cook(tagMap);
String url = am.get(MetadataField.FIELD_ACCESS_URL);
ArchivalUnit au = cu.getArchivalUnit();
if (url == null || url.isEmpty() || !au.makeCachedUrl(url).hasContent()) {
url = cu.getUrl();
}
am.replace(
MetadataField.FIELD_ACCESS_URL,
HttpToHttpsUtil.AuUtil.normalizeHttpHttpsFromBaseUrl(au, url));
emitter.emitMetadata(cu, am);
}
@Override
public ArticleMetadata extract(MetadataTarget target, CachedUrl cu) throws IOException {
ArticleMetadata am = super.extract(target, cu);
am.cook(tagMap);
String url = am.get(MetadataField.FIELD_ACCESS_URL);
if (url != null && !url.isEmpty()) {
CachedUrl val = cu.getArchivalUnit().makeCachedUrl(url);
if (!val.hasContent()) {
am.replace(MetadataField.FIELD_ACCESS_URL, cu.getUrl());
}
} else {
am.replace(MetadataField.FIELD_ACCESS_URL, cu.getUrl());
}
return am;
}
@Override
public ArticleMetadata extract(MetadataTarget target, CachedUrl cu) throws IOException {
log.debug3("Metadata - cachedurl cu:" + cu.getUrl());
ArticleMetadata am = super.extract(target, cu);
am.cook(tagMap);
return am;
} // extract
@Override
protected ArticleFiles createArticleFiles(CachedUrl cu) {
String url = cu.getUrl();
Matcher mat = ABSTRACT_PATTERN.matcher(url);
if (mat.find()) {
return processAbstract(cu, mat);
}
log.warning("Mismatch between article iterator factory and article iterator: " + url);
return null;
}
@Override
protected ArticleFiles createArticleFiles(CachedUrl cu) {
String url = cu.getUrl();
log.debug3("iterating url: " + url);
Matcher mat = PATTERN.matcher(url);
if (mat.find()) {
return processFullTextPdf(cu, mat);
}
log.warning("Mismatch between article iterator factory and article iterator: " + url);
return null;
}
/*
* hasArticleMetadata(CachedUrl cu)
* Given the CachedUrl for the potential abstract file, using the existing
* SimpleHtmlMetaTagMetadataExtractor to parse the file and
* retrieve any contained metadata. If a doi or author exists, it's an article
* NOT defining the Metadata Extractor here!
*/
private boolean hasArticleMetadata(CachedUrl cu) {
MetadataTarget at = new MetadataTarget(MetadataTarget.PURPOSE_ARTICLE);
ArticleMetadata am;
SimpleHtmlMetaTagMetadataExtractor ext = new SimpleHtmlMetaTagMetadataExtractor();
if (cu != null && cu.hasContent()) {
try {
at.setFormat("text/html");
am = ext.extract(at, cu);
if ((am.containsRawKey("bepress_citation_journal_title"))
|| (am.containsRawKey("bepress_citation_abstract_html_url"))
|| (am.containsRawKey("bepress_citation_doi"))
|| (am.containsRawKey("bepress_citation_author"))) {
return true;
}
} catch (IOException e) {
e.printStackTrace();
}
}
return false; // no reasonable metadata, probably a toc
}
/*
* This is comlicated. MOST AUs have articles that live below and issue level TOC
* that is,
* <blah>/<journal_id>/vol#/iss#/ is a toc with no relevant metadata
* <blah>/<journal_id>/vol#/iss#/xxx is an article with metadata
* (eg Economist Voice V1)
* BUT
* in some AUs there are issues with only 1 article, in which case
* <blah>/<journal_id>/vol#/iss#/ is an abstract with metadata
* (eg Rhodes Cook V4)
* and a few AUs with a mixture
* (eg Forum for Health Economics V5)
* So to identify ALL articles, we'll also have to capture issue level items and then look
* at the html and if it has article metadata in it, count it as an article.
*
*/
@Override
protected ArticleFiles createArticleFiles(CachedUrl cu) {
String url = cu.getUrl();
Matcher mat = pattern.matcher(url);
if (mat.find()) {
// we matched, but could this pattern potentially be a toc?
Matcher tocmat = TOC_pattern.matcher(url);
// if we could be a TOC then we must have metadata to be considered an article
if (tocmat.find()) {
if (hasArticleMetadata(cu)) {
return processUrl(cu, mat);
}
} else {
// we're not a potential TOC, so treat this as an article without checking
return processUrl(cu, mat);
}
return null; // this was a TOC, not an article
}
log.warning("Mismatch between article iterator factory and article iterator: " + url);
return null;
}
/*
* 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;
}
