Example usage for org.apache.lucene.index IndexReader maxDoc

List of usage examples for org.apache.lucene.index IndexReader maxDoc

  1. HOME
  2. Java
  3. org.apache.lucene
  4. org.apache.lucene.index.*
  5. IndexReader
  6. maxDoc

Introduction

In this page you can find the example usage for org.apache.lucene.index IndexReader maxDoc.

Prototype

public abstract int maxDoc();

Source Link

Document

Returns one greater than the largest possible document number.

Usage

From source file:de.unihildesheim.iw.lucene.document.FeedbackQuery.java

License:Open Source License

/**
 * Get the maximum number of documents that can be retrieved.
 *
 * @param reader Reader to access the index
 * @param docCount Number of documents that should be retrieved
 * @return Actual number of documents possible to retrieve
 *///  w ww  .j  av  a 2s . c om
static int getMaxDocs(@NotNull final IndexReader reader, final int docCount) {
    final int maxRetDocs; // maximum number of documents that can be returned
    if (docCount == Integer.MAX_VALUE) {
        return reader.maxDoc();
    }
    final int maxIdxDocs = reader.maxDoc();
    if (docCount > maxIdxDocs) {
        maxRetDocs = Math.min(maxIdxDocs, docCount);
        LOG.warn("Requested number of feedback documents ({}) "
                + "is larger than the amount of documents in the index ({}). "
                + "Returning only {} feedback documents at maximum.", docCount, maxIdxDocs, maxRetDocs);
    } else {
        maxRetDocs = docCount;
    }
    return maxRetDocs;
}

From source file:edu.coeia.reports.IndexUtil.java

License:Open Source License

public static List<String> getAllFilePaths(final CaseFacade caseFacade) throws IOException {

    List<String> files = new ArrayList<String>();

    String indexDir = caseFacade.getCaseIndexFolderLocation();
    Directory dir = FSDirectory.open(new File(indexDir));
    IndexReader indexReader = IndexReader.open(dir);

    for (int i = 0; i < indexReader.maxDoc(); i++) {
        Document document = indexReader.document(i);
        if (document != null) {
            Field field = document.getField(IndexingConstant.DOCUMENT_TYPE);
            if (field != null && field.stringValue() != null) {
                String path = field.stringValue();

                if (path.equals(IndexingConstant
                        .fromDocumentTypeToString(IndexingConstant.DOCUMENT_GENERAL_TYPE.FILE))) {
                    String relativePath = document.get(IndexingConstant.FILE_PATH);

                    if (!relativePath.isEmpty()) {
                        String fullpath = caseFacade.getFullPath(relativePath);
                        files.add(fullpath);
                    }/*from www . j a va2  s  . c  o  m*/
                }
            }
        }
    }

    indexReader.close();
    return files;
}

From source file:edu.coeia.reports.IndexUtil.java

License:Open Source License

private static List<String> getAllFilePathsHaveAuther(final CaseFacade caseFacade, final List<String> authers)
        throws IOException {

    List<String> files = new ArrayList<String>();

    String indexDir = caseFacade.getCaseIndexFolderLocation();
    Directory dir = FSDirectory.open(new File(indexDir));
    IndexReader indexReader = IndexReader.open(dir);

    for (int i = 0; i < indexReader.maxDoc(); i++) {
        Document document = indexReader.document(i);
        if (document != null) {
            Field field = document.getField(IndexingConstant.DOCUMENT_TYPE);
            if (field != null && field.stringValue() != null) {
                String path = field.stringValue();

                if (path.equals(IndexingConstant
                        .fromDocumentTypeToString(IndexingConstant.DOCUMENT_GENERAL_TYPE.FILE))) {
                    String relativePath = document.get(IndexingConstant.FILE_PATH);
                    String auther = document.get("Author");

                    if (!relativePath.isEmpty() && auther != null && !auther.trim().isEmpty()
                            && Utilities.isFound(authers, auther)) {
                        String fullpath = caseFacade.getFullPath(relativePath);
                        files.add(fullpath);
                    }//from   ww w  . j  ava2 s .  c  o  m
                }
            }
        }
    }

    indexReader.close();
    return files;
}

From source file:edu.coeia.reports.IndexUtil.java

License:Open Source License

public static List<String> getAllAuthers(final CaseFacade caseFacade) throws IOException {

    List<String> files = new ArrayList<String>();

    String indexDir = caseFacade.getCaseIndexFolderLocation();
    Directory dir = FSDirectory.open(new File(indexDir));
    IndexReader indexReader = IndexReader.open(dir);

    for (int i = 0; i < indexReader.maxDoc(); i++) {
        Document document = indexReader.document(i);
        if (document != null) {
            Field field = document.getField(IndexingConstant.DOCUMENT_TYPE);
            if (field != null && field.stringValue() != null) {
                String path = field.stringValue();

                if (path.equals(IndexingConstant
                        .fromDocumentTypeToString(IndexingConstant.DOCUMENT_GENERAL_TYPE.FILE))) {
                    String relativePath = document.get(IndexingConstant.FILE_PATH);
                    String auther = document.get("Author");

                    if (!relativePath.isEmpty() && auther != null && !auther.trim().isEmpty()) {
                        files.add(auther);
                    }//from www.  j  av  a2s  . co m
                }
            }
        }
    }

    indexReader.close();
    return files;
}

From source file:edu.coeia.tasks.CaseDuplicationTask.java

License:Open Source License

private void fillCaseDuplicationMap() throws Exception {
    String indexDir = this.aCase.getCaseLocation() + File.separator + ApplicationConstants.CASE_INDEX_FOLDER;
    Directory dir = FSDirectory.open(new File(indexDir));
    IndexReader indexReader = IndexReader.open(dir);

    try {//from   w  w  w. java2s  .co  m
        for (int i = 0; i < indexReader.maxDoc(); i++) {
            Document document = indexReader.document(i);
            if (document != null) {
                Field field = document.getField(IndexingConstant.DOCUMENT_HASH);
                if (field != null && field.stringValue() != null) {
                    String documentHash = field.stringValue();
                    this.panel.getCaseDuplicationMap().put(documentHash,
                            document.get(IndexingConstant.DOCUMENT_ID));
                }
            }
        }
    } finally {
        indexReader.close();
    }
}

From source file:edu.coeia.tasks.CommonKeywordsTask.java

License:Open Source License

public Map<String, Integer> getAllTermFreqFromItems() throws IOException {
    Map<String, Integer> map = new HashMap<String, Integer>();

    String indexDir = this.aCase.getCaseLocation() + File.separator + ApplicationConstants.CASE_INDEX_FOLDER;
    Directory dir = FSDirectory.open(new File(indexDir));
    IndexReader indexReader = IndexReader.open(dir);
    TermEnum terms = indexReader.terms();

    int factor = indexReader.maxDoc() / 100;

    while (terms.next()) {
        if (isCancelledTask())
            break;

        Term term = terms.term();//from   www  . j av a  2 s  .  c o m

        if (this.isAllowedFeild(term.field().trim())) {
            String termText = term.text();
            int frequency = indexReader.docFreq(term);

            if (frequency >= factor)
                map.put(termText, frequency);
        }
    }

    System.out.println("map size: " + map.size());
    indexReader.close();
    return map;
}

From source file:edu.coeia.tasks.EmailProcessingTask.java

License:Open Source License

private void getAllMessageInEmailSource(final String value, final String constant) {
    IndexReader indexReader = null;

    try {// w w  w. j  ava  2 s .  c  om
        List<Integer> ids = new ArrayList<Integer>();

        Directory dir = FSDirectory.open(new File(this.panel.getCaseFacade().getCaseIndexFolderLocation()));
        indexReader = IndexReader.open(dir);
        Map<Entry, Integer> messageCounter = new HashMap<Entry, Integer>();

        for (int i = 0; i < indexReader.maxDoc(); i++) {
            if (this.isCancelledTask()) {
                break;
            }

            Document document = indexReader.document(i);

            if (document != null) {
                Field field = document.getField(constant);
                if (field != null && field.stringValue() != null) {
                    String tmp = field.stringValue();

                    if (tmp.endsWith(value)) {
                        EmailItem item = (EmailItem) ItemFactory.newInstance(document, panel.getCaseFacade(),
                                false);

                        String to = "";
                        if (item.getTo() == null || item.getTo().trim().isEmpty())
                            to = value;
                        else
                            to = item.getTo();

                        if (checkingItemType(item)) {
                            Entry entry = new Entry(item.getFrom(), to, item.getTime());

                            Integer indexNo = messageCounter.get(entry);

                            if (indexNo == null) {
                                messageCounter.put(entry, 1);
                            } else {
                                messageCounter.put(entry, indexNo + 1);
                            }

                            ids.add(Integer.valueOf(item.getDocumentId()));
                        }
                    }
                }
            }
        }

        if (this.type == EMAIL_PROCESSING_TYPE.INBOX || this.type == EMAIL_PROCESSING_TYPE.SEND_ITEM)
            addResultToTable(messageCounter);
        else
            addDateResultToTable(messageCounter);

        messageCounter.clear();
        messageCounter = null;
    } catch (IOException ex) {
        ex.printStackTrace();
        Logger.getLogger(EmailProcessingTask.class.getName()).log(Level.SEVERE, null, ex);
    } finally {
        try {
            indexReader.close();
        } catch (IOException ex) {
            Logger.getLogger(EmailProcessingTask.class.getName()).log(Level.SEVERE, null, ex);
        }
    }
}

From source file:edu.coeia.tasks.ExtensionFrequencyTask.java

License:Open Source License

private int getFactor() throws IOException {
    Directory directory = FSDirectory.open(new File(this.caseFacade.getCaseIndexFolderLocation()));

    IndexReader indexReader = IndexReader.open(directory);

    int factor = indexReader.maxDoc() / 200;
    indexReader.close();/* w  ww . j av  a2 s  .c o  m*/

    return factor;
}

From source file:edu.coeia.tasks.MultimediaLoadingTask.java

License:Open Source License

private void loadItems() throws IOException {
    String indexDir = this.caseFacade.getCaseIndexFolderLocation();
    Directory dir = FSDirectory.open(new File(indexDir));
    IndexReader indexReader = IndexReader.open(dir);

    for (int i = 0; i < indexReader.maxDoc(); i++) {
        if (this.isCancelledTask())
            break;

        Document document = indexReader.document(i);

        if (document != null) {
            Field field = document.getField(IndexingConstant.FILE_MIME);

            if (field != null && field.stringValue() != null) {
                String documentExtension = field.stringValue();
                final StringBuilder fullpath = new StringBuilder();

                if (type == MultimediaViewerPanel.TYPE.IMAGE && isImage(documentExtension)) {
                    fullpath.append(this.caseFacade.getFullPath(document.get(IndexingConstant.FILE_PATH)));
                } else if (type == MultimediaViewerPanel.TYPE.AUDIO && isAudio(documentExtension)) {
                    fullpath.append(this.caseFacade.getFullPath(document.get(IndexingConstant.FILE_PATH)));
                } else if (type == MultimediaViewerPanel.TYPE.ARCHIVE && isArchieve(documentExtension)) {
                    fullpath.append(this.caseFacade.getFullPath(document.get(IndexingConstant.FILE_PATH)));
                } else if (type == MultimediaViewerPanel.TYPE.VIDEO && isVideo(documentExtension)) {
                    fullpath.append(this.caseFacade.getFullPath(document.get(IndexingConstant.FILE_PATH)));
                }//w w  w  .j  av a2 s  .  c o  m

                if (!fullpath.toString().isEmpty()) {
                    EventQueue.invokeLater(new Runnable() {
                        @Override
                        public void run() {
                            File file = new File(fullpath.toString());
                            Object[] data = { file.getAbsolutePath(), FileUtil.getExtension(file),
                                    file.lastModified(), file.isHidden(), file.length() };
                            JTableUtil.addRowToJTable(panel.getTable(), data);
                        }
                    });
                }
            }
        }
    }

    indexReader.close();
}

From source file:edu.illinois.cs.cogcomp.wikifier.utils.examples.LuceneExample.java

License:Open Source License

public static void main(String[] args) throws IOException, ParseException {

    String pathToIndexDir = "testIndex";
    createIndex(pathToIndexDir);// ww  w .  j av  a2s.  c o  m
    IndexReader reader = Lucene.reader(pathToIndexDir);
    Map<String, Float> idfs = Lucene.getIdfs(reader, "text");
    for (String k : idfs.keySet()) {
        System.out.println(k + " " + idfs.get(k));
    }
    System.out.println("TFS");
    for (int i = 0; i < reader.maxDoc(); i++) {
        System.out.println(reader.document(i).getField("title").stringValue());
        Map<String, Float> tfs = Lucene.getTfs(reader, "text", i);
        for (String k : tfs.keySet()) {
            System.out.println(k + " " + tfs.get(k));
        }
    }

}