medlinecitation.java

一个自然语言处理的Java开源工具包。LingPipe目前已有很丰富的功能

/*
 * LingPipe v. 3.5
 * Copyright (C) 2003-2008 Alias-i
 *
 * This program is licensed under the Alias-i Royalty Free License
 * Version 1 WITHOUT ANY WARRANTY, without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the Alias-i
 * Royalty Free License Version 1 for more details.
 *
 * You should have received a copy of the Alias-i Royalty Free License
 * Version 1 along with this program; if not, visit
 * http://alias-i.com/lingpipe/licenses/lingpipe-license-1.txt or contact
 * Alias-i, Inc. at 181 North 11th Street, Suite 401, Brooklyn, NY 11211,
 * [phone].
 */

package com.aliasi.medline;

import com.aliasi.xml.DelegatingHandler;
import com.aliasi.xml.DelegateHandler;
import com.aliasi.xml.TextAccumulatorHandler;

import java.io.IOException;

import java.text.DateFormat;
import java.text.ParseException;
import java.text.SimpleDateFormat;

import java.util.ArrayList;
import java.util.Arrays;
import java.util.Date;
import java.util.Locale;

import org.xml.sax.Attributes;
import org.xml.sax.InputSource;
import org.xml.sax.SAXException;
import org.xml.sax.XMLReader;

import org.xml.sax.helpers.DefaultHandler;

/**
 * A <code>MedlineCitation</code> represents the content of a single
 * record in the 2008 MEDLINE database for the citation of an individual
 * article.
 *
 * <P>A citation contains a richly structured set of information.
 * Each citation contains an {@link Article}, which contains most of
 * the information drawn from the article itself.  The other
 * information has been added by the creators of MEDLINE and comes in
 * two flavors.  The first type is information about the MEDLINE
 * record itself, such as who &quot;owns&quot; it, when it was
 * created, completed and revised, etc.  The second type of
 * information is metadata about the article itself, such as names of
 * chemicals and database links, keywords, lists of MeSH terms, gene
 * names and database links, cross-references to comments or
 * corrections in other articles, etc.
 *
 * <P>In addition to the structured information in a citation the raw
 * XML is available as a string through the {@link #xmlString()}
 * method.
 *
 * @author  Bob Carpenter
 * @version 3.3
 * @since   LingPipe2.0
 */
public class MedlineCitation {

    private final String mNlmDcmsID;
    private final String mPMID;
    private final String mOwner;
    private final String mStatus;
    private final Date mDateCreated;
    private final Date mDateCompleted;
    private final Date mDateRevised;
    private final Chemical[] mChemicals;
    private final String[] mCitationSubsets;
    private final MeshHeading[] mMeshHeadings;
    private final String mNumberOfReferences;
    private final PersonalNameSubject[] mPersonalNameSubjects;
    private final KeywordList[] mKeywordLists;
    private final GeneralNote[] mGeneralNotes;
    private final String[] mGeneSymbols;
    private final String[] mSpaceFlightMissions;
    private final OtherID[] mOtherIDs;
    private final Investigator[] mInvestigators;
    private final OtherAbstract[] mOtherAbstracts;
    private final CommentOrCorrection[] mCommentOrCorrections;
    private final JournalInfo mJournalInfo;
    private final Article mArticle;
    private String mXMLString;

    MedlineCitation(String owner,
                    String status,
                    String nlmDcmsID,
                    String pmid,
                    Date dateCreated,
                    Date dateCompleted,
                    Date dateRevised,
                    Chemical[] chemicals,
                    String[] citationSubsets,
                    MeshHeading[] meshHeadings,
                    String numberOfReferences,
                    PersonalNameSubject[] personalNameSubjects,
                    KeywordList[] keywordLists,
                    GeneralNote[] generalNotes,
                    String[] geneSymbols,
                    String[] spaceFlightMissions,
                    OtherID[] otherIDs,
                    Investigator[] investigators,
                    OtherAbstract[] otherAbstracts,
                    CommentOrCorrection[] commentOrCorrections,
                    JournalInfo journalInfo,
                    Article article) {
        mOwner = owner;
        mStatus = status;
        mNlmDcmsID = nlmDcmsID;
        mPMID = pmid;
        mDateCreated = dateCreated;
        mDateCompleted = dateCompleted;
        mDateRevised = dateRevised;
        mChemicals = chemicals;
        mCitationSubsets = citationSubsets;
        mMeshHeadings = meshHeadings;
        mNumberOfReferences = numberOfReferences;
        mPersonalNameSubjects = personalNameSubjects;
        mKeywordLists = keywordLists;
        mGeneralNotes = generalNotes;
        mGeneSymbols = geneSymbols;
        mSpaceFlightMissions = spaceFlightMissions;
        mOtherIDs = otherIDs;
        mInvestigators = investigators;
        mOtherAbstracts = otherAbstracts;
        mCommentOrCorrections = commentOrCorrections;
        mJournalInfo = journalInfo;
        mArticle = article;
    }

    /**
     * Returns the XML underlying this citation as a string.  Note
     * that there will be no XML declaration in this string, nor
     * will there be a DTD reference.  All entities and defaults
     * provided by the DTD will be expanded into the string.
     *
     * @return The XML underlying this citation as a string.
     */
    public String xmlString() {
        return mXMLString;
    }

    /**
     * The owner is the group that created and validated this
     * citation.  Each citation has a unique owner, which takes on one
     * of the following possible values, shown with their
     * corresponding constants in {@link MedlineCitationSet}.
     *
     * <blockquote>
     * <table border='1' cellpadding='5'>
     * <tr><td><i>Value</i></td>
     *     <td><i>Constant in</i> <code>MedlineCitationSet</code></td>
     *     <td><i>Description</i></td><td><i>Notes</i></tr>
     * <tr><td>NLM</td><td>NLM_VALUE</td>
     *     <td>(U.S.) National Library of Medicine, Index Section</td>
     *     <td>Value for majority of citations</td></tr>
     * <tr><td>NASA</td><td>NASA_VALUE</td>
     *     <td>(U.S.) National Aeronautics and Space Administration</td>
     *     <td>&nbsp;</td></tr>
     * <tr><td>PIP</td><td>PIP_VALUE</td>
     *     <td>Population Information Program, Johns Hopkins</td>
     *     <td>Not current; only on older citations</td></tr>
     * <tr><td>KIE</td><td>KIE_VALUE</td>
     *     <td>Kennedy Institute of Ethics, Georgetown University</td>
     *     <td>&nbsp;</td></tr>
     * <tr><td>HSR</td><td>HSR_VALUE</td>
     *     <td>(U.S.) National Infomration Center on Health Services Reserach
     *         and Health Care Technology, National Library of Medicine</td>
     *     <td>&nbsp;</td></tr>
     * <tr><td>HMD</td><td>HMD_VALUE</td>
     *     <td>History of Medicine Division, National Library of Medicine</td>
     *     <td>&nbsp;</td></tr>
     * <tr><td>SIS</td><td>SIS_VALUE</td>
     *     <td>Specialized Information Services Division,
     *         National Library of Medicine</td>
     *     <td>Not yet used; reserved for future use.</td></tr>
     * <tr><td>NOTNLM</td><td>NOT_NLM_VALUE</td>
     *     <td>Not from NLM</td>
     *     <td>Will never be used for any distributed MEDLINE record</td></tr>
     * </table>
     * </blockquote>
     *
     * <P>These values may also appear as owners of general notes and
     * keyword lists.
     *
     * @return The owner of the group that created and validated this
     * citation.
     */
    public String owner() {
        return mOwner;
    }

    /**
     * Returns the status of this citation. The status is always
     * defined, and takes one of the values listed below.  As articles
     * progress through the pipeline, they are rereleased with new
     * status values; the later versions should replace the earlier
     * ones based on PubMed identifier.
     *
     * <blockquote>
     * <table border='1' cellpadding='5'>
     * <tr><td><i>Value</i></td>
     *     <td><i>Constant in</i> <code>MedlineCitationSet</code></td>
     *     <td><i>Notes</i></td></tr>
     * <tr><td>In-Data-Review</td><td>IN_DATA_REVIEW_VALUE</td>
     *     <td>Reference citation elements at the journal issue level have
     *         been checked against print or online versions.  Citations
     *         with this status lack completion dates.  They have not
     *         undergrone quality assurance nor had MeSH terms added.</td></tr>
     * <tr><td>In-Process</td><td>IN_PROCESS_VALUE</td>
     *     <td>Citation elements at the article level have been
     *         reviewed, including author names, title and pagination.
     *         Not every in-process article is promoted to MEDLINE; some
     *         are classified out of scope and receive PubMed-not-Medline
     *         as their final status</td></tr>
     * <tr><td>Completed</td><td>COMPLETED_VALUE</td>
     *     <td>These are 'true' MEDLINE records with
     *         date completed and typically with MeSH headings.
     *         This status remains under revisions.</td></tr>
     * <tr><td>PubMed-not-MEDLINE</td><td>PUBMED_NOT_MEDLINE_VALUE
     *     <td>These are articles have gone through data review,
     *         but were not chosen for inclusion.  They fall into
     *         the following categories:
     *         (1) citations that precede the data a journal was
     *         selected for MEDLINE, (2) citations that are out of scope,
     *         and (3) analytical summaries of articles published
     *         elsewhere.</td></tr>
     * <tr><td>Publisher</td><td>PUBLISHER_VALUE</td>
     *     <td>These are not distributed to licensees, and include
     *         a number of categories of yet-to-be processed, out of scope,
     *         and out of date articles.</td></tr>
     * <tr><td>MEDLINE</td><td>MEDLINE_VALUE</td>
     *     <td>2005 addition not yet documented by the NLM.</td></tr>
     * <tr><td>OLDMEDLINE</td><td>MEDLINE_VALUE</td>
     *     <td>2005 addition not yet documented by the NLM.</td></tr>
     * </table>
     * </blockquote>
     *
     * @return The status of this citation.
     */
    public String status() {
        return mStatus;
    }

    /**
     * Returns the NLM DCMS identifier for this document.
     * May return <code>null</code> if there is no identifier.
     *
     * <P>The PubMed identifier, available through the method {@link
     * #pmid()}, is the official identifier for MEDLINE documents.
     *
     * <P>Note that this identifier was introduced in 2005 and
     * was previously known as the MEDLINE ID.
     *
     * @return The NLM DCMS identifier for this document.
     */
    public String nlmDcmsID() {
        return mNlmDcmsID;
    }

    /**
     * Returns the PubMed identifier for this document.  This is now
     * the official identifier for MEDLINE documents.  Each document
     * has a unique PubMed identifier, which will be a one to eight
     * digit accession number without leading zeros.  The identifiers
     * of deleted records will not be reused.
     *
     * @return The PubMed identifier for this document.
     */
    public String pmid() {
        return mPMID;
    }

    /**
     * Returns the date on which this citation was created.  Note that
     * this is <i>not</i> the date on which the article was created,
     * but is the date on which the MEDLINE citation for the article
     * was created.  The value will be non-null because every citation
     * is required to have a creation date.  The date will only be
     * resolved to the level of day, month and year.
     *
     * <P>Note that this is also not the same as the PubMed Entrez
     * date, which is not distributed as part of MEDLINE.
     *
     * @return The date on which this citation was created.
     */
    public Date dateCreated() {
        return mDateCreated;
    }

    /**
     * Returns the date that the processing of this citation ended or
     * <code>null</code> if it is still in process.  Specifically,
     * citations are completed when all extra information has been
     * added and quality assurance has been completed.  If a record's
     * status is in-process, as indicated by method {@link #status()},
     * it lacks a date completed, so that this method will return
     * <code>null</code>.  The date will only be resolved to the level
     * of day, month and year.
     *
     * <P>For ciations before about the year 2000, the date created
     * and date completed were set to the same value.
     *
     * @return The date that this citation was completed.
     */
    public Date dateCompleted() {
        return mDateCompleted;
    }

    /**
     * Returns the latest date on which a change was made to this
     * citation as a result of maintenance.  Note that no information
     * is provided in MEDLINE as to the nature of the change, and only
     * the latest revision date is included.  The date will only be
     * resolved to the level of day, month and year.  The value may be
     * <code>null</code> if this citation has not been revised.
     *
     * <P>Note that every record that existed at the time was revised
     * on 18 December 2000, and many of these have been revised
     * subsequently.
     *