@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 this to do some additional attempts to get valid data before emitting
@Override
public void extract(MetadataTarget target, CachedUrl cu, FileMetadataExtractor.Emitter emitter)
throws IOException, PluginException {
ArticleMetadata am = extract(target, cu);
/*
* if, due to overcrawl, we got to a page that didn't have anything
* valid, eg "this page not found" html page
* don't emit empty metadata (because defaults would get put in
* Must do this after cooking, because it checks size of cooked info
*/
if (am.isEmpty()) {
return;
}
/*
* RIS data can be variable. We don't have any way to add priority to
* the cooking of data, so fallback to alternate values manually
*
* There are differences between books and journals, so fork for titles
* and metadata check
*/
if (am.get(MetadataField.FIELD_DATE) == null) {
if (am.getRaw("Y1") != null) { // if DA wasn't there, use Y1
am.put(MetadataField.FIELD_DATE, am.getRaw("Y1"));
}
}
/*
* Determine if this is a book item or a journal item.
* set the appropriate article type once the daemon passes along the TY
*/
String ris_type = am.getRaw("TY");
if (ris_type == null) {
// pre 1.69, do an alternate check because TY wasn't passed through
ris_type = "JOUR"; // set a default
if (am.get(MetadataField.FIELD_ISBN) != null) {
// it is a bad value, but it was recognized as an isbn because of TY type
ris_type = "BOOK"; // it could be a chapter but until TY is passed through...
}
}
// Modify or try alternate RIS tag values based after cooking
postCookProcess(cu, am, ris_type);
// Only emit if this item is likely to be from this AU
// protect against counting overcrawled articles by checking against
// values from the TDB file - differentiate between book items and journal itesm
ArchivalUnit au = cu.getArchivalUnit();
if (ris_type.contains("BOOK") || ris_type.contains("CHAP")) {
if (!BaseAtyponMetadataUtil.metadataMatchesBookTdb(au, am)) {
return;
}
} else {
// JOURNAL default is to assume it's a journal for backwards compatibility
if (!BaseAtyponMetadataUtil.metadataMatchesTdb(au, am)) {
return;
}
}
/*
* Fill in DOI, publisher, other information available from
* the URL or TDB
* CORRECT the access.url if it is not in the AU
*/
BaseAtyponMetadataUtil.completeMetadata(cu, am);
emitter.emitMetadata(cu, am);
}
/*
* isolate the modifications done on the AM after the initial extraction
* in order to allow child plugins to do override this and do
* additional work before calling the pre-emit checking...
* ArticleMetadata - passed in information from extract/cook
* ris_type - the TY value or its inferred type (basically, book or journal)
*/
protected void postCookProcess(CachedUrl cu, ArticleMetadata am, String ris_type) {
/*
* RIS data can be variable. We don't have any way to add priority to
* the cooking of data, so fallback to alternate values manually
*/
if (am.get(MetadataField.FIELD_DATE) == null) {
if (am.getRaw("Y1") != null) { // if DA wasn't there, use Y1
am.put(MetadataField.FIELD_DATE, am.getRaw("Y1"));
}
}
/*
* There are differences between books and journals, so fork for titles
* and metadata check
*/
if (ris_type.contains("BOOK") || ris_type.contains("CHAP")) {
// BOOK in some form
// T1 is the primary title - of the chapter for a book chapter, or book for a complete book
// T2 is the next title up - of the book for a chapter, of the series for a book
// T3 is the uppermost - of the series for a chapter
// sometimes they use TI instead of T1...
if (am.get(MetadataField.FIELD_ARTICLE_TITLE) == null) {
if (am.getRaw("TI") != null) { // if T1 wasn't there, use TI
am.put(MetadataField.FIELD_ARTICLE_TITLE, am.getRaw("TI"));
}
}
if (ris_type.contains("CHAP")) {
// just one chapter - set the article type correctly
am.put(MetadataField.FIELD_ARTICLE_TYPE, MetadataField.ARTICLE_TYPE_BOOKCHAPTER);
if ((am.get(MetadataField.FIELD_PUBLICATION_TITLE) == null)
&& (am.getRaw("T2") != null)) {
// the publication and the article titles are just the name of the book
am.put(MetadataField.FIELD_PUBLICATION_TITLE, am.getRaw("T2"));
}
if ((am.get(MetadataField.FIELD_SERIES_TITLE) == null) && (am.getRaw("T3") != null)) {
// the publication and the article titles are just the name of the book
am.put(MetadataField.FIELD_SERIES_TITLE, am.getRaw("T3"));
}
} else {
// We're a full book volume - articletitle = publicationtitle
am.put(MetadataField.FIELD_ARTICLE_TYPE, MetadataField.ARTICLE_TYPE_BOOKVOLUME);
if (am.get(MetadataField.FIELD_PUBLICATION_TITLE) == null) {
// the publication and the article titles are just the name of the book
am.put(
MetadataField.FIELD_PUBLICATION_TITLE, am.get(MetadataField.FIELD_ARTICLE_TITLE));
}
// series title can be from T2
if ((am.get(MetadataField.FIELD_SERIES_TITLE) == null) && (am.getRaw("T2") != null)) {
// the publication and the article titles are just the name of the book
am.put(MetadataField.FIELD_SERIES_TITLE, am.getRaw("T2"));
}
}
} else {
// JOURNAL default is to assume it's a journal for backwards compatibility
if (am.get(MetadataField.FIELD_PUBLICATION_TITLE) == null) {
if (am.getRaw("T2") != null) {
am.put(MetadataField.FIELD_PUBLICATION_TITLE, am.getRaw("T2"));
} else if (am.getRaw("JO") != null) {
am.put(
MetadataField.FIELD_PUBLICATION_TITLE,
am.getRaw("JO")); // might be unabbreviated version
}
}
}
}