it.unimi.di.big.mg4j.document.WikipediaDocumentSequence.java Source code

Java tutorial

Introduction

Here is the source code for it.unimi.di.big.mg4j.document.WikipediaDocumentSequence.java

Source

package it.unimi.di.big.mg4j.document;

/*       
 * MG4J: Managing Gigabytes for Java (big)
 *
 * Copyright (C) 2013-2015 Sebastiano Vigna 
 *
 *  This library is free software; you can redistribute it and/or modify it
 *  under the terms of the GNU Lesser General Public License as published by the Free
 *  Software Foundation; either version 3 of the License, or (at your option)
 *  any later version.
 *
 *  This library is distributed in the hope that it will be useful, but
 *  WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
 *  or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU Lesser General Public License
 *  for more details.
 *
 *  You should have received a copy of the GNU Lesser General Public License
 *  along with this program; if not, see <http://www.gnu.org/licenses/>.
 *
 */

import info.bliki.wiki.filter.Encoder;
import info.bliki.wiki.filter.HTMLConverter;
import info.bliki.wiki.filter.PlainTextConverter;
import info.bliki.wiki.model.WikiModel;
import it.unimi.di.big.mg4j.tool.URLMPHVirtualDocumentResolver;
import it.unimi.di.big.mg4j.tool.VirtualDocumentResolver;
import it.unimi.di.big.mg4j.util.parser.callback.AnchorExtractor;
import it.unimi.di.big.mg4j.util.parser.callback.AnchorExtractor.Anchor;
import it.unimi.dsi.big.io.FileLinesCollection;
import it.unimi.dsi.big.util.ShiftAddXorSignedStringMap;
import it.unimi.dsi.big.util.StringMap;
import it.unimi.dsi.big.webgraph.ImmutableGraph;
import it.unimi.dsi.bits.TransformationStrategies;
import it.unimi.dsi.bits.TransformationStrategy;
import it.unimi.dsi.fastutil.io.BinIO;
import it.unimi.dsi.fastutil.io.FastBufferedInputStream;
import it.unimi.dsi.fastutil.objects.AbstractObject2LongFunction;
import it.unimi.dsi.fastutil.objects.Object2IntOpenHashMap;
import it.unimi.dsi.fastutil.objects.Object2LongFunction;
import it.unimi.dsi.fastutil.objects.Object2LongLinkedOpenHashMap;
import it.unimi.dsi.fastutil.objects.Object2ObjectOpenHashMap;
import it.unimi.dsi.fastutil.objects.ObjectArrayList;
import it.unimi.dsi.fastutil.objects.ObjectBigList;
import it.unimi.dsi.fastutil.objects.ObjectOpenHashSet;
import it.unimi.dsi.fastutil.objects.Reference2ObjectMap;
import it.unimi.dsi.fastutil.objects.Reference2ObjectOpenHashMap;
import it.unimi.dsi.io.FastBufferedReader;
import it.unimi.dsi.io.WordReader;
import it.unimi.dsi.lang.MutableString;
import it.unimi.dsi.lang.ObjectParser;
import it.unimi.dsi.logging.ProgressLogger;
import it.unimi.dsi.sux4j.mph.MWHCFunction;
import it.unimi.dsi.util.TextPattern;

import java.io.FileInputStream;
import java.io.IOException;
import java.io.InputStream;
import java.io.InputStreamReader;
import java.io.Serializable;
import java.text.ParseException;
import java.text.SimpleDateFormat;
import java.util.Map;
import java.util.Map.Entry;
import java.util.concurrent.ArrayBlockingQueue;

import javax.xml.parsers.ParserConfigurationException;
import javax.xml.parsers.SAXParser;
import javax.xml.parsers.SAXParserFactory;

import org.apache.commons.compress.compressors.bzip2.BZip2CompressorInputStream;
import org.apache.commons.io.IOUtils;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.xml.sax.Attributes;
import org.xml.sax.InputSource;
import org.xml.sax.SAXException;
import org.xml.sax.helpers.DefaultHandler;

import com.google.common.base.Charsets;
import com.google.common.collect.ImmutableList;
import com.google.common.collect.ImmutableSet;
import com.google.common.collect.Iterables;
import com.google.common.html.HtmlEscapers;
import com.martiansoftware.jsap.FlaggedOption;
import com.martiansoftware.jsap.JSAP;
import com.martiansoftware.jsap.JSAPException;
import com.martiansoftware.jsap.JSAPResult;
import com.martiansoftware.jsap.Parameter;
import com.martiansoftware.jsap.SimpleJSAP;
import com.martiansoftware.jsap.Switch;
import com.martiansoftware.jsap.UnflaggedOption;

/** A class exhibiting a standard Wikipedia XML dump as a {@link DocumentSequence}. 
 * 
 * <P><strong>Warning</strong>: this class has no connection whatsoever with
 * {@link WikipediaDocumentCollection}.
 * 
 * <p>The purpose of this class is making the indexing of Wikipedia and of its entity
 * graph starting from a pristine Wikipedia XML dump reasonably easy. There are a few
 * steps involved, mainly due to the necessity of working out redirects, but the whole
 * procedure can be carried out with very little resources. The class uses the
 * {@link WikiModel#toHtml(String, Appendable, String, String)} method to convert
 * the Wikipedia format into HTML, and then passes the result to a standard {@link HtmlDocumentFactory}
 * (suggestion on alternative conversion methods are welcome).
 * A few additional fields are handled by {@link WikipediaHeaderFactory}.
 * 
 * <p>Note that no properties are passed to the underlying {@link HtmlDocumentFactory}: if you want
 * to set the anchor properties (see {@link HtmlDocumentFactory.MetadataKeys}), you need to use
 * {@linkplain #WikipediaDocumentSequence(String, boolean, String, boolean, boolean, boolean, int, int, int) a quite humongous constructor}.
 * 
 * <h3>How to index Wikipedia</h3>
 * 
 * <p>As a first step, download the Wikipedia XML dump (it's the &ldquo;pages-articles&rdquo; file;
 * it should start with a <code>mediawiki</code> opening tag). This class can process the
 * file in its compressed form, but we suggest to uncompress it using <code>bunzip2</code>,
 * as processing is an order of magnitude faster. (Note that the following process will exclude namespaced
 * pages such as <code>Template:<var>something</var></code> and all templates, such as infoboxes, etc.; if you want to include them, you must 
 * use a different {@linkplain #WikipediaDocumentSequence(String, boolean, String, boolean, boolean, boolean) constructor}.)
 * 
 * 
 * <p>The first step is extracting metadata (in particular, the URLs that are necessary to
 * index correctly the anchor text). We do not suggest specific Java options, but try to use
 * as much memory as you can.
 * <pre>
 * java it.unimi.di.big.mg4j.tool.ScanMetadata \
 *   -o "it.unimi.di.big.mg4j.document.WikipediaDocumentSequence(enwiki-latest-pages-articles.xml,false,http://en.wikipedia.org/wiki/,false)" \
 *   -u enwiki.uris -t enwiki.titles
 * </pre>
 * 
 * <p>Note that we used the {@link ObjectParser}-based constructor of this class, which makes it possible to create
 * a {@link WikipediaDocumentSequence} instance parsing a textual specification (see the 
 * {@linkplain #WikipediaDocumentSequence(String, boolean, String, boolean) constructor}
 * documentation for details about the parameters).
 * 
 * <p>The second step consists in building a first {@link VirtualDocumentResolver} which, however, 
 * does not comprise redirect information:
 * <pre>
 * java it.unimi.di.big.mg4j.tool.URLMPHVirtualDocumentResolver -o enwiki.uris enwiki.vdr
 * </pre>
 * 
 * <p>Now we need to use the <i>ad hoc</i> main method of this class to rescan the collection, gather the redirect
 * information and merge it with our current resolver:
 * <pre>
 * java it.unimi.di.big.mg4j.document.WikipediaDocumentSequence \
 *   enwiki-latest-pages-articles.xml http://en.wikipedia.org/wiki/ enwiki.uris enwiki.vdr enwikired.vdr
 * </pre>
 * 
 * <p>During this phase a quite large number of warnings about <em>failed redirects</em> might appear. This is normal,
 * in particular if you do not index template pages. If you suspect an actual bug, try first to index template pages,
 * too. Failed redirects should be in the order of few thousands, and all due to internal inconsistencies of
 * the dump: to check that this is the case, check whether the target of a failed redirect appears as a page
 * title (it shouldn't).
 *   
 * <p>We have now all information required to build a complete index (we use the Porter2 stemmer in this example):
 * <pre>
 * java it.unimi.di.big.mg4j.tool.IndexBuilder \
 *   -o "it.unimi.di.big.mg4j.document.WikipediaDocumentSequence(enwiki.xml,false,http://en.wikipedia.org/wiki/,true)" \ 
 *   --all-fields -v enwikired.vdr -t EnglishStemmer enwiki
 * </pre>
 * 
 * <p>Finally, we can build the entity graph using a bridge class that exposes any {@link DocumentSequence} with a virtual
 * field as an {@link ImmutableGraph} of the <a href="http://webgraph.di.unimi.it/">WebGraph framework</a> (the nodes will be in one-to-one correspondence with the documents
 * returned by the index):
 * <pre>
 * java it.unimi.dsi.big.webgraph.BVGraph \
 *   -s "it.unimi.di.big.mg4j.util.DocumentSequenceImmutableSequentialGraph(\"it.unimi.di.big.mg4j.document.WikipediaDocumentSequence(enwiki.xml,false,http://en.wikipedia.org/wiki/,true)\",anchor,enwikired.vdr)" \ 
 *   enwiki
 * </pre>
 * 
 * <h2>Additional fields</h2>
 * 
 * <p>The additional fields generated by this class (some of which are a bit hacky) are:
 *
 * <dl>
 * <dt><code>title</code>
 * <dd>the title of the Wikipedia page;
 * <dt><code>id</code>
 * <dd>a payload index containing the Wikipedia identifier of the page;
 * <dt><code>lastedit</code>
 * <dd>a payload index containing the last edit of the page;
 * <dt><code>category</code>
 * <dd>a field containing the categories of the page, separated by an artificial marker <code>OXOXO</code> (so when you look for a category as a phrase you
 * don't get false cross-category positives);
 * <dt><code>firstpar</code>
 * <dd>a heuristically generated first paragraph of the page, useful for identification beyond the title;
 * <dt><code>redirects</code>
 * <dd>a virtual field treating the link of the page with its title and any redirect link to the page as an anchor: in practice, the
 * field contains all names under which the page is known in Wikipedia. 
 * </dl>
 * 
 * <p>Note that for each link in a disambiguation page this class will generate a fake link with the same target, but
 * the title of the disambiguation page as text. This is in the same spirit of the <code>redirects</code> field&mdash;we enrich
 * the HTML <code>anchor</code> field with useful information without altering the generated graph.
 */

public class WikipediaDocumentSequence extends AbstractDocumentSequence implements Serializable {
    private static final Logger LOGGER = LoggerFactory.getLogger(WikipediaDocumentSequence.class);
    private static final long serialVersionUID = 1L;

    private static final TextPattern CATEGORY_START = new TextPattern("[[Category:");
    private static final TextPattern BRACKETS_CLOSED = new TextPattern("]]");
    private static final TextPattern BRACES_CLOSED = new TextPattern("}}");
    private static final TextPattern DISAMBIGUATION = new TextPattern("{{disambiguation");
    private static final TextPattern BRACKETS_OPEN = new TextPattern("[[");
    private static final char[] END_OF_DISAMBIGUATION_LINK = new char[] { '|', ']' };

    /** A marker used to denote end of input. */
    private static final DocumentAndFactory END = new DocumentAndFactory(null, null);
    /** The prototype {@link CompositeDocumentFactory} used to parse Wikipedia pages. */
    private final DocumentFactory factory;
    /** Whether the input is compressed with <code>bzip2</code>. */
    private final boolean bzipped;
    /** Whether to parse text (e.g., we do not parse text when computing titles/URIs). */
    private final boolean parseText;
    /** Whether to keep in the index namespace pages. */
    private final boolean keepNamespaced;
    /** The Wikipedia XML dump. */
    private final String wikipediaXmlDump;
    /** The base URL for pages (e.g., <code>http://en.wikipedi.org/wiki/</code>). */
    private final String baseURL;
    /** {@link #baseURL} concatenated with <code>${title}</code>. */
    private final String linkBaseURL;
    /** {@link #baseURL} concatenated with <code>${image}</code>. */
    private final String imageBaseURL;
    /** The set of namespaces specified in {@link #wikipediaXmlDump}. */
    private ImmutableSet<MutableString> nameSpaces;
    /** This list (whose access must be synchronized) accumulates virtual text (anchors) generated by redirects.
     * It is filled when meeting redirect pages, and it is emptied at the first non-redirect page (the page in which the list
     * is emptied is immaterial). Note that because of this setup, if there are some redirect 
     * pages that are not followed by any indexed page the anchors of those redirects won't be processed at all. 
     * If this is a problem, just add a fake empty page at the end. */
    private final ObjectArrayList<Anchor> redirectAnchors = new ObjectArrayList<Anchor>();
    /** The Bliki model used to parse pages. */
    private final MyWikiModel wikiModel;

    public static enum MetadataKeys {
        ID, LASTEDIT, CATEGORY, FIRSTPAR,
        /** This key is used internally by {@link WikipediaHeaderFactory} and is associated with the list of redirect anchors. */
        REDIRECT
    };

    /** A factory responsible for special Wikipedia fields (see the {@linkplain WikipediaDocumentSequence class documentation}). It
     * will be {@linkplain CompositeDocumentFactory composed} with an {@link HtmlDocumentFactory}. */
    public static final class WikipediaHeaderFactory extends AbstractDocumentFactory {
        private static final long serialVersionUID = 1L;
        private static final Object2IntOpenHashMap<String> FIELD_2_INDEX = new Object2IntOpenHashMap<String>(
                new String[] { "title", "id", "lastedit", "category", "firstpar", "redirect" },
                new int[] { 0, 1, 2, 3, 4, 5 });
        static {
            FIELD_2_INDEX.defaultReturnValue(-1);
        }

        private final WordReader wordReader = new FastBufferedReader();

        @Override
        public int numberOfFields() {
            return 6;
        }

        @Override
        public String fieldName(int field) {
            switch (field) {
            case 0:
                return "title";
            case 1:
                return "id";
            case 2:
                return "lastedit";
            case 3:
                return "category";
            case 4:
                return "firstpar";
            case 5:
                return "redirect";
            default:
                throw new IllegalArgumentException();
            }
        }

        @Override
        public int fieldIndex(String fieldName) {
            return FIELD_2_INDEX.getInt(fieldName);
        }

        @Override
        public FieldType fieldType(int field) {
            switch (field) {
            case 0:
                return FieldType.TEXT;
            case 1:
                return FieldType.INT;
            case 2:
                return FieldType.DATE;
            case 3:
                return FieldType.TEXT;
            case 4:
                return FieldType.TEXT;
            case 5:
                return FieldType.VIRTUAL;
            default:
                throw new IllegalArgumentException();
            }
        }

        @Override
        public Document getDocument(final InputStream unusedRawContent,
                final Reference2ObjectMap<Enum<?>, Object> metadata) throws IOException {
            return new AbstractDocument() {

                @Override
                public WordReader wordReader(int field) {
                    return wordReader; // Fixed, for the time being.
                }

                @Override
                public CharSequence uri() {
                    return (CharSequence) metadata.get(PropertyBasedDocumentFactory.MetadataKeys.URI);
                }

                @Override
                public CharSequence title() {
                    return (CharSequence) metadata.get(PropertyBasedDocumentFactory.MetadataKeys.TITLE);
                }

                @Override
                public Object content(final int field) throws IOException {
                    switch (field) {
                    case 0:
                        return new FastBufferedReader(
                                (MutableString) metadata.get(PropertyBasedDocumentFactory.MetadataKeys.TITLE));
                    case 1:
                        return metadata.get(MetadataKeys.ID);
                    case 2:
                        return metadata.get(MetadataKeys.LASTEDIT);
                    case 3:
                        return new FastBufferedReader((MutableString) metadata.get(MetadataKeys.CATEGORY));
                    case 4:
                        return new FastBufferedReader((MutableString) metadata.get(MetadataKeys.FIRSTPAR));
                    case 5:
                        @SuppressWarnings("unchecked")
                        final ObjectArrayList<Anchor> redirectAnchors = (ObjectArrayList<Anchor>) metadata
                                .get(MetadataKeys.REDIRECT);
                        ImmutableList<Anchor> result;

                        synchronized (redirectAnchors) {
                            redirectAnchors.add(new Anchor(
                                    (MutableString) metadata.get(PropertyBasedDocumentFactory.MetadataKeys.URI),
                                    (MutableString) metadata.get(PropertyBasedDocumentFactory.MetadataKeys.TITLE)));
                            result = ImmutableList.copyOf(redirectAnchors);
                            redirectAnchors.clear();
                        }
                        return result;
                    default:
                        throw new IllegalArgumentException();
                    }
                }
            };
        }

        @Override
        public DocumentFactory copy() {
            return new WikipediaHeaderFactory();
        }

    }

    /** An extension of {@link WikiModel} that is able, optionally to skip templates. */
    private static class MyWikiModel extends WikiModel {
        private final boolean keepTemplates;

        public MyWikiModel(String imageBaseURL, String linkBaseURL, final boolean keepTemplates) {
            super(imageBaseURL, linkBaseURL);
            this.keepTemplates = keepTemplates;
        }

        @Override
        public void substituteTemplateCall(String templateName, Map<String, String> parameterMap, Appendable writer)
                throws IOException {
            if (keepTemplates)
                super.substituteTemplateCall(templateName, parameterMap, writer);
        }
    }

    /** Builds a new Wikipedia document sequence using default anchor settings and discarding namespaced pages and templates.
     * 
     * @param file the file containing the Wikipedia dump.
     * @param bzipped whether {@code file} is compressed with <code>bzip2</code>.
     * @param baseURL a base URL for links (e.g., for the English Wikipedia, <code>http://en.wikipedia.org/wiki/</code>);
     * note that if it is nonempty this string <strong>must</strong> terminate with a slash.
     * @param parseText whether to parse the text (this parameter is only set to false during metadata-scanning
     * phases to speed up the scanning process).
     */
    public WikipediaDocumentSequence(final String file, final boolean bzipped, final String baseURL,
            final boolean parseText) {
        this(file, bzipped, baseURL, parseText, false, false);
    }

    /** Builds a new Wikipedia document sequence using default anchor settings.
     * 
     * @param file the file containing the Wikipedia dump.
     * @param bzipped whether {@code file} is compressed with <code>bzip2</code>.
     * @param baseURL a base URL for links (e.g., for the English Wikipedia, <code>http://en.wikipedia.org/wiki/</code>);
     * note that if it is nonempty this string <strong>must</strong> terminate with a slash.
     * @param parseText whether to parse the text (this parameter is only set to false during metadata-scanning
     * phases to speed up the scanning process).
     * @param keepNamespaced whether to keep namespaced pages (e.g., <code>Template:<var>something</var></code> pages).
     */
    public WikipediaDocumentSequence(final String file, final boolean bzipped, final String baseURL,
            final boolean parseText, final boolean keepNamespaced, final boolean keepTemplates) {
        this(file, bzipped, baseURL, parseText, keepNamespaced, keepTemplates,
                HtmlDocumentFactory.DEFAULT_MAXPREANCHOR, HtmlDocumentFactory.DEFAULT_MAXANCHOR,
                HtmlDocumentFactory.DEFAULT_MAXPOSTANCHOR);
    }

    /** Builds a new Wikipedia document sequence.
     * 
     * @param file the file containing the Wikipedia dump.
     * @param bzipped whether {@code file} is compressed with <code>bzip2</code>.
     * @param baseURL a base URL for links (e.g., for the English Wikipedia, <code>http://en.wikipedia.org/wiki/</code>);
     * note that if it is nonempty this string <strong>must</strong> terminate with a slash.
     * @param parseText whether to parse the text (this parameter is only set to false during metadata-scanning
     * phases to speed up the scanning process).
     * @param keepNamespaced whether to keep namespaced pages (e.g., <code>Template:<var>something</var></code> pages).
     * @param keepTemplates whether to keep templates (e.g., infoboxes, taxoboxes, etc.); we suggest to pass false if you're building a Wikipedia graph.
     * @param maxPreAnchor maximum number of character before an anchor.
     * @param maxAnchor maximum number of character in an anchor.
     * @param maxPostAnchor maximum number of characters after an anchor.
     */
    public WikipediaDocumentSequence(final String file, final boolean bzipped, final String baseURL,
            final boolean parseText, final boolean keepNamespaced, final boolean keepTemplates,
            final int maxPreAnchor, final int maxAnchor, final int maxPostAnchor) {
        this.wikipediaXmlDump = file;
        this.bzipped = bzipped;
        this.baseURL = baseURL;
        this.parseText = parseText;
        this.keepNamespaced = keepNamespaced;
        this.factory = new CompositeDocumentFactory(
                new DocumentFactory[] { new WikipediaHeaderFactory(),
                        new HtmlDocumentFactory(new Reference2ObjectOpenHashMap<Enum<?>, Object>(
                                new Enum[] { HtmlDocumentFactory.MetadataKeys.MAXPREANCHOR,
                                        HtmlDocumentFactory.MetadataKeys.MAXANCHOR,
                                        HtmlDocumentFactory.MetadataKeys.MAXPOSTANCHOR },
                                new Integer[] { Integer.valueOf(maxPreAnchor), Integer.valueOf(maxAnchor),
                                        Integer.valueOf(maxPostAnchor) })) },
                new String[] { "title", "id", "lastedit", "category", "firstpar", "redirect", "text", "dummy",
                        "anchor" });
        linkBaseURL = baseURL + "${title}";
        imageBaseURL = baseURL + "${image}";
        wikiModel = new MyWikiModel(imageBaseURL, linkBaseURL, keepTemplates);
    }

    /** A string-based constructor to be used with an {@link ObjectParser}.
     *
     * @see #WikipediaDocumentSequence(String, boolean, String, boolean)
     */
    public WikipediaDocumentSequence(final String file, final String bzipped, final String baseURL,
            final String parseText) {
        this(file, Boolean.parseBoolean(bzipped), baseURL, Boolean.parseBoolean(parseText));
    }

    /** A string-based constructor to be used with an {@link ObjectParser}.
     *
     * @see #WikipediaDocumentSequence(String, boolean, String, boolean, boolean, boolean)
     */
    public WikipediaDocumentSequence(final String file, final String bzipped, final String baseURL,
            final String parseText, final String keepNamespaced, final String keepTemplates) {
        this(file, Boolean.parseBoolean(bzipped), baseURL, Boolean.parseBoolean(parseText),
                Boolean.parseBoolean(keepNamespaced), Boolean.parseBoolean(keepTemplates));
    }

    /** A string-based constructor to be used with an {@link ObjectParser}.
     *
     * @see #WikipediaDocumentSequence(String, boolean, String, boolean, boolean, boolean, int, int, int)
     */
    public WikipediaDocumentSequence(final String file, final String bzipped, final String baseURL,
            final String parseText, final String keepNamespaced, final String keepTemplates,
            final String maxBeforeAnchor, final String maxAnchor, final String maxPostAnchor) {
        this(file, Boolean.parseBoolean(bzipped), baseURL, Boolean.parseBoolean(parseText),
                Boolean.parseBoolean(keepNamespaced), Boolean.parseBoolean(keepTemplates),
                Integer.parseInt(maxBeforeAnchor), Integer.parseInt(maxAnchor), Integer.parseInt(maxPostAnchor));
    }

    private static final class DocumentAndFactory {
        public final Document document;
        public final DocumentFactory factory;

        public DocumentAndFactory(final Document document, final DocumentFactory documentFactory) {
            this.document = document;
            this.factory = documentFactory;
        }
    }

    @Override
    public DocumentIterator iterator() throws IOException {
        final SAXParserFactory saxParserFactory = SAXParserFactory.newInstance();
        saxParserFactory.setNamespaceAware(true);
        final MutableString nameSpaceAccumulator = new MutableString();
        final ObjectOpenHashSet<MutableString> nameSpacesAccumulator = new ObjectOpenHashSet<MutableString>();
        final ArrayBlockingQueue<DocumentFactory> freeFactories = new ArrayBlockingQueue<DocumentFactory>(16);
        for (int i = freeFactories.remainingCapacity(); i-- != 0;)
            freeFactories.add(this.factory.copy());
        final ArrayBlockingQueue<DocumentAndFactory> readyDocumentsAndFactories = new ArrayBlockingQueue<DocumentAndFactory>(
                freeFactories.size());

        final SAXParser parser;
        try {
            parser = saxParserFactory.newSAXParser();
        } catch (Exception e) {
            throw new RuntimeException(e.getMessage(), e);
        }
        final DefaultHandler handler = new DefaultHandler() {
            private final SimpleDateFormat dateFormat = new SimpleDateFormat("yyyy-MM-dd'T'HH:mm:ss'Z'");
            private boolean inText;
            private boolean inTitle;
            private boolean inId;
            private boolean inTimestamp;
            private boolean inNamespaceDef;
            private boolean redirect;
            private MutableString text = new MutableString();
            private MutableString title = new MutableString();
            private MutableString id = new MutableString();
            private MutableString timestamp = new MutableString();
            private final Reference2ObjectMap<Enum<?>, Object> metadata = new Reference2ObjectOpenHashMap<Enum<?>, Object>();
            {
                metadata.put(PropertyBasedDocumentFactory.MetadataKeys.ENCODING, "UTF-8");
                metadata.put(MetadataKeys.REDIRECT, redirectAnchors);
            }

            @Override
            public void startElement(String uri, String localName, String qName, Attributes attributes)
                    throws SAXException {
                if ("page".equals(localName)) {
                    redirect = inText = inTitle = inId = inTimestamp = false;
                    text.length(0);
                    title.length(0);
                    id.length(0);
                    timestamp.length(0);
                } else if ("text".equals(localName))
                    inText = true;
                else if ("title".equals(localName) && title.length() == 0)
                    inTitle = true; // We catch only the first id/title elements.
                else if ("id".equals(localName) && id.length() == 0)
                    inId = true;
                else if ("timestamp".equals(localName) && timestamp.length() == 0)
                    inTimestamp = true;
                else if ("redirect".equals(localName)) {
                    redirect = true;
                    if (attributes.getValue("title") != null)
                        // Accumulate the title of the page as virtual text of the redirect page.
                        synchronized (redirectAnchors) {
                            final String link = Encoder.encodeTitleToUrl(attributes.getValue("title"), true);
                            redirectAnchors.add(
                                    new AnchorExtractor.Anchor(new MutableString(baseURL.length() + link.length())
                                            .append(baseURL).append(link), title.copy()));
                        }
                } else if ("namespace".equals(localName)) {
                    // Found a new namespace
                    inNamespaceDef = true;
                    nameSpaceAccumulator.length(0);
                }
            }

            @Override
            public void endElement(String uri, String localName, String qName) throws SAXException {
                if ("namespace".equals(localName)) { // Collecting a namespace
                    if (nameSpaceAccumulator.length() != 0)
                        nameSpacesAccumulator.add(nameSpaceAccumulator.copy().toLowerCase());
                    return;
                }

                if ("namespaces".equals(localName)) { // All namespaces collected
                    nameSpaces = ImmutableSet.copyOf(nameSpacesAccumulator);
                    return;
                }

                if (!redirect) {
                    if ("title".equals(localName)) {
                        // Set basic metadata for the page
                        metadata.put(PropertyBasedDocumentFactory.MetadataKeys.TITLE, title.copy());
                        String link = Encoder.encodeTitleToUrl(title.toString(), true);
                        metadata.put(PropertyBasedDocumentFactory.MetadataKeys.URI,
                                new MutableString(baseURL.length() + link.length()).append(baseURL).append(link));
                        inTitle = false;
                    } else if ("id".equals(localName)) {
                        metadata.put(MetadataKeys.ID, Long.valueOf(id.toString()));
                        inId = false;
                    } else if ("timestamp".equals(localName)) {
                        try {
                            metadata.put(MetadataKeys.LASTEDIT, dateFormat.parse(timestamp.toString()));
                        } catch (ParseException e) {
                            throw new RuntimeException(e.getMessage(), e);
                        }
                        inTimestamp = false;
                    } else if ("text".equals(localName)) {
                        inText = false;
                        if (!keepNamespaced) {
                            // Namespaces are case-insensitive and language-dependent
                            final int pos = title.indexOf(':');
                            if (pos != -1 && nameSpaces.contains(title.substring(0, pos).toLowerCase()))
                                return;
                        }
                        try {
                            final MutableString html = new MutableString();
                            DocumentFactory freeFactory;
                            try {
                                freeFactory = freeFactories.take();
                            } catch (InterruptedException e) {
                                throw new RuntimeException(e.getMessage(), e);
                            }
                            if (parseText) {
                                if (DISAMBIGUATION.search(text) != -1) { // It's a disambiguation page.
                                    /* Roi's hack: duplicate links using the page title, so the generic name will end up as anchor text. */
                                    final MutableString newLinks = new MutableString();
                                    for (int start = 0, end; (start = BRACKETS_OPEN.search(text,
                                            start)) != -1; start = end) {
                                        end = start;
                                        final int endOfLink = text.indexOfAnyOf(END_OF_DISAMBIGUATION_LINK, start);
                                        // Note that we don't escape title because we are working at the Wikipedia raw text level.
                                        if (endOfLink != -1) {
                                            newLinks.append(text.array(), start, endOfLink - start).append('|')
                                                    .append(title).append("]]\n");
                                            end = endOfLink;
                                        }
                                        end++;
                                    }

                                    text.append(newLinks);
                                }
                                // We separate categories by OXOXO, so we don't get overflowing phrases.
                                final MutableString category = new MutableString();
                                for (int start = 0, end; (start = CATEGORY_START.search(text,
                                        start)) != -1; start = end) {
                                    end = BRACKETS_CLOSED.search(text, start += CATEGORY_START.length());
                                    if (end != -1)
                                        category.append(text.subSequence(start, end)).append(" OXOXO ");
                                    else
                                        break;
                                }
                                metadata.put(MetadataKeys.CATEGORY, category);

                                // Heuristics to get the first paragraph
                                metadata.put(MetadataKeys.FIRSTPAR, new MutableString());
                                String plainText = wikiModel.render(new PlainTextConverter(true), text.toString());
                                for (int start = 0; start < plainText.length(); start++) {
                                    //System.err.println("Examining " + plainText.charAt( start )  );
                                    if (Character.isWhitespace(plainText.charAt(start)))
                                        continue;
                                    if (plainText.charAt(start) == '{') {
                                        //System.err.print( "Braces " + start + " text: \"" + plainText.subSequence( start, start + 10 )  + "\" -> " );
                                        start = BRACES_CLOSED.search(plainText, start);
                                        //System.err.println( start + " text: \"" + plainText.subSequence( start, start + 10 ) + "\"" );
                                        if (start == -1)
                                            break;
                                        start++;
                                    } else if (plainText.charAt(start) == '[') {
                                        start = BRACKETS_CLOSED.search(plainText, start);
                                        if (start == -1)
                                            break;
                                        start++;
                                    } else {
                                        final int end = plainText.indexOf('\n', start);
                                        if (end != -1)
                                            metadata.put(MetadataKeys.FIRSTPAR,
                                                    new MutableString(plainText.substring(start, end)));
                                        break;
                                    }
                                }

                                try {
                                    wikiModel.render(new HTMLConverter(), text.toString(), html, false, true);
                                    final Map<String, String> categories = wikiModel.getCategories();
                                    // Put back category links in the page (they have been parsed by bliki and to not appear anymore in the HTML rendering)
                                    for (Entry<String, String> entry : categories.entrySet()) {
                                        final String key = entry.getKey();
                                        final String value = entry.getValue().trim();
                                        if (value.length() != 0) // There are empty such things
                                            html.append("\n<a href=\"").append(baseURL).append("Category:")
                                                    .append(Encoder.encodeTitleToUrl(key, true)).append("\">")
                                                    .append(HtmlEscapers.htmlEscaper().escape(key))
                                                    .append("</a>\n");
                                    }
                                } catch (Exception e) {
                                    LOGGER.error("Unexpected exception while parsing " + title, e);
                                }
                            }
                            readyDocumentsAndFactories.put(new DocumentAndFactory(
                                    freeFactory.getDocument(IOUtils.toInputStream(html, Charsets.UTF_8),
                                            new Reference2ObjectOpenHashMap<Enum<?>, Object>(metadata)),
                                    freeFactory));
                        } catch (InterruptedException e) {
                            throw new RuntimeException(e.getMessage(), e);
                        } catch (IOException e) {
                            throw new RuntimeException(e.getMessage(), e);
                        }
                    }
                }
            }

            @Override
            public void characters(char[] ch, int start, int length) throws SAXException {
                if (inText && parseText)
                    text.append(ch, start, length);
                if (inTitle)
                    title.append(ch, start, length);
                if (inId)
                    id.append(ch, start, length);
                if (inTimestamp)
                    timestamp.append(ch, start, length);
                if (inNamespaceDef) {
                    nameSpaceAccumulator.append(ch, start, length);
                    inNamespaceDef = false; // Dirty, but it works
                }
            }

            @Override
            public void ignorableWhitespace(char[] ch, int start, int length) throws SAXException {
                if (inText && parseText)
                    text.append(ch, start, length);
                if (inTitle)
                    title.append(ch, start, length);
            }
        };

        final Thread parsingThread = new Thread() {
            public void run() {
                try {
                    InputStream in = new FileInputStream(wikipediaXmlDump);
                    if (bzipped)
                        in = new BZip2CompressorInputStream(in);
                    parser.parse(
                            new InputSource(new InputStreamReader(new FastBufferedInputStream(in), Charsets.UTF_8)),
                            handler);
                    readyDocumentsAndFactories.put(END);
                } catch (Exception e) {
                    throw new RuntimeException(e.getMessage(), e);
                }
            }
        };

        parsingThread.start();

        return new AbstractDocumentIterator() {
            private DocumentFactory lastFactory;

            @Override
            public Document nextDocument() throws IOException {
                try {
                    final DocumentAndFactory documentAndFactory = readyDocumentsAndFactories.take();
                    if (lastFactory != null)
                        freeFactories.put(lastFactory);
                    if (documentAndFactory == END)
                        return null;
                    lastFactory = documentAndFactory.factory;
                    return documentAndFactory.document;
                } catch (InterruptedException e) {
                    throw new RuntimeException(e.getMessage(), e);
                }
            }
        };
    }

    @Override
    public DocumentFactory factory() {
        return factory;
    }

    /** A wrapper around a signed function that remaps entries exceeding a provided threshold using a specified target array. */
    public static final class SignedRedirectedStringMap extends AbstractObject2LongFunction<CharSequence>
            implements StringMap<CharSequence> {
        private static final long serialVersionUID = 1L;
        /** The number of documents. */
        private final long numberOfDocuments;
        /** A signed function function mapping valid keys to their ordinal position. */
        private Object2LongFunction<CharSequence> signedFunction;
        /** The value to be returned for keys whose ordinal position is greater than {@link #numberOfDocuments}. */
        private final long[] target;

        /** Creates a new signed redirected map.
         * 
         * @param numberOfDocuments the threshold after which the {@code target} array will be used to compute the output.
         * @param signedFunction the base signed function.
         * @param target an array providing the output for items beyond {@code numberOfDocuments}; it must be
         * long as the size of {@code signedFunction} minus {@code numberOfDocuments}.
         */
        public SignedRedirectedStringMap(final long numberOfDocuments,
                final Object2LongFunction<CharSequence> signedFunction, final long[] target) {
            this.numberOfDocuments = numberOfDocuments;
            this.signedFunction = signedFunction;
            this.target = target;
        }

        @Override
        public long getLong(Object key) {
            final long index = signedFunction.getLong(key);
            if (index == -1)
                return -1;
            if (index < numberOfDocuments)
                return index;
            return target[(int) (index - numberOfDocuments)];
        }

        @Override
        public boolean containsKey(Object key) {
            return signedFunction.getLong(key) != -1;
        }

        public long size64() {
            return numberOfDocuments;
        }

        @Override
        @Deprecated
        public int size() {
            return (int) Math.min(Integer.MAX_VALUE, size64());
        }

        @Override
        public ObjectBigList<? extends CharSequence> list() {
            return null;
        }
    }

    public static void main(final String arg[])
            throws ParserConfigurationException, SAXException, IOException, JSAPException, ClassNotFoundException {
        SimpleJSAP jsap = new SimpleJSAP(WikipediaDocumentSequence.class.getName(),
                "Computes the redirects of a Wikipedia dump and integrate them into an existing virtual document resolver for the dump.",
                new Parameter[] { new Switch("bzip2", 'b', "bzip2", "The file is compressed with bzip2"),
                        new Switch("iso", 'i', "iso",
                                "Use ISO-8859-1 coding internally (i.e., just use the lower eight bits of each character)."),
                        new FlaggedOption("width", JSAP.INTEGER_PARSER, Integer.toString(Long.SIZE),
                                JSAP.NOT_REQUIRED, 'w', "width",
                                "The width, in bits, of the signatures used to sign the function from URIs to their rank."),
                        new UnflaggedOption("file", JSAP.STRING_PARSER, JSAP.REQUIRED,
                                "The file containing the Wikipedia dump."),
                        new UnflaggedOption("baseURL", JSAP.STRING_PARSER, JSAP.REQUIRED,
                                "The base URL for the collection (e.g., http://en.wikipedia.org/wiki/)."),
                        new UnflaggedOption("uris", JSAP.STRING_PARSER, JSAP.REQUIRED,
                                "The URIs of the documents in the collection (generated by ScanMetadata)."),
                        new UnflaggedOption("vdr", JSAP.STRING_PARSER, JSAP.REQUIRED,
                                "The name of a precomputed virtual document resolver for the collection."),
                        new UnflaggedOption("redvdr", JSAP.STRING_PARSER, JSAP.REQUIRED,
                                "The name of the resulting virtual document resolver.") });

        JSAPResult jsapResult = jsap.parse(arg);
        if (jsap.messagePrinted())
            return;

        final SAXParserFactory saxParserFactory = SAXParserFactory.newInstance();
        saxParserFactory.setNamespaceAware(true);
        final Object2ObjectOpenHashMap<MutableString, String> redirects = new Object2ObjectOpenHashMap<MutableString, String>();
        final String baseURL = jsapResult.getString("baseURL");
        final ProgressLogger progressLogger = new ProgressLogger(LOGGER);
        progressLogger.itemsName = "redirects";
        progressLogger.start("Extracting redirects...");

        final SAXParser parser = saxParserFactory.newSAXParser();
        final DefaultHandler handler = new DefaultHandler() {
            private boolean inTitle;
            private MutableString title = new MutableString();

            @Override
            public void startElement(String uri, String localName, String qName, Attributes attributes)
                    throws SAXException {
                if ("page".equals(localName)) {
                    inTitle = false;
                    title.length(0);
                } else if ("title".equals(localName) && title.length() == 0)
                    inTitle = true; // We catch only the first title element.
                else if ("redirect".equals(localName) && attributes.getValue("title") != null) {
                    progressLogger.update();
                    redirects.put(title.copy(), attributes.getValue("title"));
                }
            }

            @Override
            public void endElement(String uri, String localName, String qName) throws SAXException {
                if ("title".equals(localName))
                    inTitle = false;
            }

            @Override
            public void characters(char[] ch, int start, int length) throws SAXException {
                if (inTitle)
                    title.append(ch, start, length);
            }

            @Override
            public void ignorableWhitespace(char[] ch, int start, int length) throws SAXException {
                if (inTitle)
                    title.append(ch, start, length);
            }
        };

        InputStream in = new FileInputStream(jsapResult.getString("file"));
        if (jsapResult.userSpecified("bzip2"))
            in = new BZip2CompressorInputStream(in);
        parser.parse(new InputSource(new InputStreamReader(new FastBufferedInputStream(in), Charsets.UTF_8)),
                handler);
        progressLogger.done();

        final Object2LongLinkedOpenHashMap<MutableString> resolved = new Object2LongLinkedOpenHashMap<MutableString>();
        final VirtualDocumentResolver vdr = (VirtualDocumentResolver) BinIO.loadObject(jsapResult.getString("vdr"));

        progressLogger.expectedUpdates = redirects.size();
        progressLogger.start("Examining redirects...");

        for (Map.Entry<MutableString, String> e : redirects.entrySet()) {
            final MutableString start = new MutableString().append(baseURL)
                    .append(Encoder.encodeTitleToUrl(e.getKey().toString(), true));
            final MutableString end = new MutableString().append(baseURL)
                    .append(Encoder.encodeTitleToUrl(e.getValue(), true));
            final long s = vdr.resolve(start);
            if (s == -1) {
                final long t = vdr.resolve(end);
                if (t != -1)
                    resolved.put(start.copy(), t);
                else
                    LOGGER.warn("Failed redirect: " + start + " -> " + end);
            } else
                LOGGER.warn("URL " + start + " is already known to the virtual document resolver");

            progressLogger.lightUpdate();
        }

        progressLogger.done();

        //System.err.println(resolved);

        final Iterable<MutableString> allURIs = Iterables
                .concat(new FileLinesCollection(jsapResult.getString("uris"), "UTF-8"), resolved.keySet());
        final long numberOfDocuments = vdr.numberOfDocuments();

        final TransformationStrategy<CharSequence> transformationStrategy = jsapResult.userSpecified("iso")
                ? TransformationStrategies.iso()
                : TransformationStrategies.utf16();

        BinIO.storeObject(new URLMPHVirtualDocumentResolver(new SignedRedirectedStringMap(numberOfDocuments,
                new ShiftAddXorSignedStringMap(allURIs.iterator(),
                        new MWHCFunction.Builder<CharSequence>().keys(allURIs).transform(transformationStrategy)
                                .build(),
                        jsapResult.getInt("width")),
                resolved.values().toLongArray())), jsapResult.getString("redvdr"));
    }
}