public InputStream createFilteredInputStream(ArchivalUnit au, InputStream in, String encoding)
throws PluginException {
HtmlTransform[] transforms =
new HtmlTransform[] {
// Filter out <a target="_blank">...</a>
HtmlNodeFilterTransform.exclude(
HtmlNodeFilters.tagWithAttribute("a", "target", "_blank")),
};
return new HtmlFilterInputStream(in, encoding, new HtmlCompoundTransform(transforms));
}
public InputStream createFilteredInputStream(ArchivalUnit au, InputStream in, String encoding) {
NodeFilter[] filters =
new NodeFilter[] {
// filter out script
new TagNameFilter("script"),
// Menu and related articles/other issues
HtmlNodeFilters.tagWithAttribute("div", "class", "A_Left_Column"),
// Footer menu that seems to currently be blank
HtmlNodeFilters.tagWithAttribute("div", "class", "A_Foot_Menu"),
// Copyright
HtmlNodeFilters.tagWithAttribute("div", "class", "A_Foot_Copy"),
// Lazy HTML these are inserted everywhere in an attempt to get the layout to work
HtmlNodeFilters.tagWithAttribute("div", "style", "clear:both;"),
// search box with broken layout. they may try to fix this
HtmlNodeFilters.tagWithAttribute("div", "id", "search"),
// another blank navigation div
HtmlNodeFilters.tagWithAttribute("div", "id", "Sub_Top_Nav"),
};
return new HtmlFilterInputStream(
in, encoding, HtmlNodeFilterTransform.exclude(new OrFilter(filters)));
}
@Override
public InputStream createFilteredInputStream(ArchivalUnit au, InputStream in, String encoding) {
NodeFilter[] includeNodes =
new NodeFilter[] {
// manifest pages
// <ul> and <li> without attributes (unlike TOC/full/abs/ref breadcrumbs)
new NodeFilter() {
@Override
public boolean accept(Node node) {
if (HtmlNodeFilters.tagWithAttributeRegex("a", "href", "/toc/").accept(node)) {
Node liParent = node.getParent();
if (liParent instanceof Bullet) {
Bullet li = (Bullet) liParent;
Vector liAttr = li.getAttributesEx();
if (liAttr != null && liAttr.size() == 1) {
Node ulParent = li.getParent();
if (ulParent instanceof BulletList) {
BulletList ul = (BulletList) ulParent;
Vector ulAttr = ul.getAttributesEx();
return ulAttr != null && ulAttr.size() == 1;
}
}
}
} else if (HtmlNodeFilters.tagWithAttributeRegex("a", "href", "/doi/book/")
.accept(node)) {
// book manifest page has single doi/book ref whose parent is just the <body>
// element
// http://emeraldinsight.com/clockss/eisbn/9780080549910
Node liParent = node.getParent();
if (liParent instanceof BodyTag) {
return true;
}
}
return false;
}
},
// book - landing page main contents chapter list and synopsis)
// http://emeraldinsight.com/doi/book/10.1108/9780080549910
HtmlNodeFilters.tagWithAttributeRegex("div", "class", "literatumBookDetailsWidget"),
// toc - contents only
// http://www.emeraldinsight.com/toc/aaaj/26/8
HtmlNodeFilters.tagWithAttributeRegex("div", "class", "literatumTocWidget"),
// abs, full, ref - contents only
// http://www.emeraldinsight.com/doi/full/10.1108/AAAJ-05-2013-1360
HtmlNodeFilters.tagWithAttributeRegex(
"div", "class", "literatumPublicationContentWidget"),
// showCitFormats
// http://www.emeraldinsight.com/action/
// showCitFormats?doi=10.1108%2F09513571311285621
HtmlNodeFilters.tagWithAttributeRegex("div", "class", "downloadCitationsWidget"),
// showPopup - generated by BaseAtyponHtmlLinkExtractorFactory
// http://www.emeraldinsight.com/action/showPopup?citid=citart1
// &id=FN_fn1&doi=10.1108%2FAAAJ-02-2012-00947
HtmlNodeFilters.tagWithAttributeRegex("body", "class", "popupBody")
};
// handled by parent: script, sfxlink, stylesheet, pdfplus file sise
// <head> tag, <li> item has the text "Cited by", accessIcon,
NodeFilter[] excludeNodes =
new NodeFilter[] {
// toc, abs, full, ref - Reprints and Permissions
HtmlNodeFilters.tagWithAttributeRegex("a", "class", "rightsLink"),
// toc - above the first toc entry with Track Citations
HtmlNodeFilters.tagWithAttributeRegex("div", "class", "toc-actions"),
// abs, full, ref - downloads count
HtmlNodeFilters.tagWithAttributeRegex("div", "class", "downloadsCount"),
// full - section choose pulldown appeared in multiple sections
// http://www.emeraldinsight.com/doi/full/10.1108/AAAJ-02-2013-1228
HtmlNodeFilters.tagWithAttribute("div", "class", "sectionJumpTo"),
// abs, full, ref - Article Options and Tools
HtmlNodeFilters.allExceptSubtree(
HtmlNodeFilters.tagWithAttributeRegex("div", "class", "options"),
HtmlNodeFilters.tagWithAttributeRegex("a", "href", "/action/showCitFormats\\?")),
// abs, full, ref - random html - potential problem
HtmlNodeFilters.tagWithAttribute("span", "class", "Z3988"),
// full, ref - references section - Crossref/ISI/Abstract/Infotrieve
// separated by a comma. Not easy to remove the comma, so hash out
// class citation
HtmlNodeFilters.tagWithAttribute("div", "class", "citation"),
// TOC - in case icon options change
HtmlNodeFilters.tagWithAttributeRegex("div", "class", "icon-key"),
// on the full/abs/ref pages there are little style definitions that
HtmlNodeFilters.tagWithAttributeRegex("style", "type", "text/css"),
};
return super.createFilteredInputStream(au, in, encoding, includeNodes, excludeNodes);
}
public InputStream createFilteredInputStream(ArchivalUnit au, InputStream in, String encoding)
throws PluginException {
NodeFilter[] filters =
new NodeFilter[] {
// Filter out <div id="footer">...</div>
HtmlNodeFilters.tagWithAttribute("div", "id", "footer"),
// Filter out <div id="top-ad-alignment">...</div>
HtmlNodeFilters.tagWithAttribute("div", "id", "top-ad-alignment"),
// Filter out <div id="top-ad">...</div>
HtmlNodeFilters.tagWithAttribute("div", "id", "top-ad"),
// Filter out <div id="ident">...</div>
HtmlNodeFilters.tagWithAttribute("div", "id", "ident"),
// Filter out <div id="ad">...</div>
HtmlNodeFilters.tagWithAttribute("div", "id", "ad"),
// Filter out <div id="vertical-ad">...</div>
HtmlNodeFilters.tagWithAttribute("div", "id", "vertical-ad"),
// Filter out <div class="right-col-download">...</div>
HtmlNodeFilters.tagWithAttribute("div", "class", "right-col-download"),
// Filter out <div id="cart-navbar">...</div>
HtmlNodeFilters.tagWithAttribute("div", "id", "cart-navbar"),
// // Filter out <div class="heading-macfix article-access-options">...</div>
// HtmlNodeFilters.tagWithAttribute("div", "class", "heading-macfix
// article-access-options"),
// Filter out <div id="baynote-recommendations">...</div>
HtmlNodeFilters.tagWithAttribute("div", "id", "baynote-recommendations"),
// Filter out <div id="bookmarks-container">...</div>
HtmlNodeFilters.tagWithAttribute("div", "id", "bookmarks-container"),
// Filter out <div id="llb">...</div>
HtmlNodeFilters.tagWithAttribute("div", "id", "llb"),
// Filter out <a href="...">...</a> where the href value includes "exitTargetId" as a
// parameter
HtmlNodeFilters.tagWithAttributeRegex("a", "href", "[\\?&]exitTargetId="),
// Filter out <input name="exitTargetId">
HtmlNodeFilters.tagWithAttribute("input", "name", "exitTargetId"),
};
return new HtmlFilterInputStream(
in, encoding, HtmlNodeFilterTransform.exclude(new OrFilter(filters)));
}
/**
* BaseAtyponHtmlCrawlFilterFactory The basic AtyponHtmlCrawlFilterFactory Child plugins can extend
* this class and add publisher specific crawl filters, if necessary. Common crawl filters can be
* easily added and be available to children. Otherwise, this can be used by child plugins if no
* other crawl filters are needed.
*/
public class BaseAtyponHtmlCrawlFilterFactory implements FilterFactory {
protected static final Pattern corrections =
Pattern.compile(
"Original Article|Corrigendum|Correction|Errata|Erratum", Pattern.CASE_INSENSITIVE);
protected static NodeFilter[] baseAtyponFilters =
new NodeFilter[] {
HtmlNodeFilters.tagWithAttribute("div", "class", "citedBySection"),
// Since overcrawling is a constant problem for Atypon, put common
// next article-previous article link for safety;
// AIAA, AMetSoc, ASCE, Ammons, APHA, SEG,Siam,
HtmlNodeFilters.tagWithAttribute("a", "class", "articleToolsNav"),
// BIR, Maney, Endocrine - also handles next/prev issue - also for issues
HtmlNodeFilters.tagWithAttributeRegex("td", "class", "journalNavRightTd"),
HtmlNodeFilters.tagWithAttributeRegex("td", "class", "journalNavLeftTd"),
// BQ, BioOne, Edinburgh, futurescience, nrc
// all handle next/prev article link in plugin
// T&F doesn't have prev/next article links
// breadcrumb or other link back to TOC from article page
// AMetSoc, Ammons, APHA, NRC,
HtmlNodeFilters.tagWithAttribute("div", "id", "breadcrumbs"),
// ASCE, BiR, Maney, SEG, SIAM, Endocrine
HtmlNodeFilters.tagWithAttributeRegex("ul", "class", "^(linkList )?breadcrumbs$"),
// on TOC next-prev issue
// AIAA, AMetSoc, Ammons, APHA,
HtmlNodeFilters.tagWithAttribute("div", "id", "nextprev"),
// ASCE, SEG, SIAM
HtmlNodeFilters.tagWithAttribute("div", "id", "prevNextNav"),
// on TOC left column with listing of all volumes/issues
HtmlNodeFilters.tagWithAttribute("ul", "class", "volumeIssues"),
// have started finding cases of direct in-publication links within references
// there are a variety of ways these blocks are identified, but
// these are unlikely to be used anywhere else so put in parent
// emerald, AIAA
HtmlNodeFilters.tagWithAttribute("div", "class", "references"),
// ASCE
HtmlNodeFilters.tagWithAttribute("li", "class", "reference"),
// maney, future-science (also in child...will remove later)
HtmlNodeFilters.tagWithAttribute("table", "class", "references"),
// Not all Atypon plugins necessarily need this but MANY do and it is
// an insidious source of over crawling
new NodeFilter() {
@Override
public boolean accept(Node node) {
if (!(node instanceof LinkTag)) return false;
String allText = ((CompositeTag) node).toPlainTextString();
return corrections.matcher(allText).find();
}
},
};
/**
* Create an array of NodeFilters that combines the atyponBaseFilters with the given array
*
* @param nodes The array of NodeFilters to add
*/
private NodeFilter[] addTo(NodeFilter[] nodes) {
NodeFilter[] result = Arrays.copyOf(baseAtyponFilters, baseAtyponFilters.length + nodes.length);
System.arraycopy(nodes, 0, result, baseAtyponFilters.length, nodes.length);
return result;
}
/**
* Create a FilteredInputStream that excludes the the atyponBaseFilters
*
* @param au The archival unit
* @param in Incoming input stream
* @param encoding The encoding
*/
public InputStream createFilteredInputStream(ArchivalUnit au, InputStream in, String encoding)
throws PluginException {
return new HtmlFilterInputStream(
in, encoding, HtmlNodeFilterTransform.exclude(new OrFilter(baseAtyponFilters)));
}
/**
* Create a FilteredInputStream that excludes the the atyponBaseFilters and moreNodes
*
* @param au The archival unit
* @param in Incoming input stream
* @param encoding The encoding
* @param moreNodes An array of NodeFilters to be excluded with atyponBaseFilters
*/
public InputStream createFilteredInputStream(
ArchivalUnit au, InputStream in, String encoding, NodeFilter[] moreNodes)
throws PluginException {
NodeFilter[] bothFilters = addTo(moreNodes);
return new HtmlFilterInputStream(
in, encoding, HtmlNodeFilterTransform.exclude(new OrFilter(bothFilters)));
}
}
@Override
public InputStream createFilteredInputStream(ArchivalUnit au, InputStream in, String encoding) {
NodeFilter[] filters =
new NodeFilter[] {
// handled by parent: script, sfxlink, stylesheet
HtmlNodeFilters.tag("noscript"),
// toc - first top block ad
// http://www.birpublications.org/toc/bjr/87/1044
HtmlNodeFilters.tagWithAttributeRegex("div", "class", "literatumAd"),
// page header: login, register, etc., and journal menu such as
// subscribe, alerts, ...
HtmlNodeFilters.tagWithAttributeRegex("div", "id", "pageHeader"),
// page footer
HtmlNodeFilters.tagWithAttributeRegex("div", "id", "pageFooter"),
// toc - BJR logo image right below pageHeader
HtmlNodeFilters.tagWithAttributeRegex("div", "class", "^widget general-image"),
// toc, abs, full, ref - menu above breadcrumbs
HtmlNodeFilters.tagWithAttributeRegex("div", "class", "menuXml"),
// toc - free.gif image tied to an abs
HtmlNodeFilters.tagWithAttributeRegex("img", "src", "free.gif"),
// toc - access icon container
HtmlNodeFilters.tagWithAttribute("td", "class", "accessIconContainer"),
// toc - pulldown with sections - may add citedby later
HtmlNodeFilters.tagWithAttribute("div", "class", "publicationTooldropdownContainer"),
// toc - right column, current issue
HtmlNodeFilters.tagWithAttributeRegex("div", "class", "literatumBookIssueNavigation"),
// toc, abs - share social media
HtmlNodeFilters.tagWithAttributeRegex("div", "class", "general-bookmark-share"),
// toc - right column impact factor block - no unique name found
HtmlNodeFilters.tagWithAttributeRegex(
"div",
"class",
"widget\\s+layout-one-column\\s+none\\s+widget-regular\\s+widget-border-toggle"),
// ref - this seems unused but may get turned on
// http://www.birpublications.org/doi/ref/10.1259/bjr.20130571
HtmlNodeFilters.tagWithAttribute("div", "id", "MathJax_Message"),
// full - section choose pulldown appeared in multiple sections
// http://www.birpublications.org/doi/full/10.1259/dmfr.20120050
HtmlNodeFilters.tagWithAttribute("div", "class", "sectionJumpTo"),
// toc, abs, full, text and ref right column - most read
// http://www.birpublications.org/toc/bjr/88/1052
HtmlNodeFilters.tagWithAttributeRegex("div", "class", "literatumMostReadWidget"),
// abs - right column all literatumArticleToolsWidget
// except Download Citation
// http://www.birpublications.org/doi/abs/10.1259/bjr.20140472
HtmlNodeFilters.allExceptSubtree(
HtmlNodeFilters.tagWithAttributeRegex("div", "class", "literatumArticleToolsWidget"),
HtmlNodeFilters.tagWithAttributeRegex("a", "href", "/action/showCitFormats\\?")),
};
// super.createFilteredInputStream adds bir filter to the baseAtyponFilters
// and returns the filtered input stream using an array of NodeFilters that
// combine the two arrays of NodeFilters.
return super.createFilteredInputStream(au, in, encoding, filters);
}
