private void initDocument() {
try {
_document = DocumentBuilderFactory.newInstance().newDocumentBuilder().newDocument();
} catch (ParserConfigurationException e) {
e.printStackTrace();
System.exit(1);
}
_rootNode = _document.createElement("body");
_document.appendChild(_rootNode);
}
private String makeStringFromXML() {
Transformer transformer = null;
try {
transformer = TransformerFactory.newInstance().newTransformer();
} catch (TransformerConfigurationException e) {
e.printStackTrace();
System.exit(1);
}
transformer.setOutputProperty(OutputKeys.INDENT, "no");
DOMSource source = new DOMSource(_document);
StringWriter writer = new StringWriter();
StreamResult result = new StreamResult(writer);
try {
transformer.transform(source, result);
} catch (TransformerException e) {
e.printStackTrace();
System.exit(1);
}
return writer.toString();
}
public static void main(String[] args) throws Exception {
IndexReader reader = null;
FSDirectory dir = null;
String field = null;
boolean IncludeTermFreqs = false;
if (args.length == 0 || args.length > 4) {
usage();
System.exit(1);
}
if (args.length > 0) {
dir = FSDirectory.open(new File(args[0]));
}
for (int i = 1; i < args.length; i++) {
if (args[i].equals("-t")) {
IncludeTermFreqs = true;
} else {
try {
numTerms = Integer.parseInt(args[i]);
} catch (NumberFormatException e) {
field = args[i];
}
}
}
String[] fields = field != null ? new String[] {field} : null;
reader = DirectoryReader.open(dir);
TermStats[] terms = getHighFreqTerms(reader, numTerms, fields);
if (!IncludeTermFreqs) {
// default HighFreqTerms behavior
for (int i = 0; i < terms.length; i++) {
System.out.printf(
"%s:%s %,d \n", terms[i].field, terms[i].termtext.utf8ToString(), terms[i].docFreq);
}
} else {
TermStats[] termsWithTF = sortByTotalTermFreq(reader, terms);
for (int i = 0; i < termsWithTF.length; i++) {
System.out.printf(
"%s:%s \t totalTF = %,d \t doc freq = %,d \n",
termsWithTF[i].field,
termsWithTF[i].termtext.utf8ToString(),
termsWithTF[i].totalTermFreq,
termsWithTF[i].docFreq);
}
}
reader.close();
}
public static void main(String[] args) throws IOException {
IndexReader reader = null;
/*
* Opening the index first simplifies the processing of the
* rest of the command line arguments.
*/
for (int i = 0; i < args.length; i++) {
if (("-index".equals(args[i])) && ((i + 1) < args.length)) {
reader = DirectoryReader.open(FSDirectory.open(new File(args[i + 1])));
if (reader == null) {
System.err.println("Error: Can't open index " + args[i + 1]);
System.exit(1);
}
;
break;
}
;
}
;
if (reader == null) {
System.err.println(usage);
System.exit(1);
}
;
/*
* Process the command line arguments sequentially.
*/
for (int i = 0; i < args.length; i++) {
if ("-index".equals(args[i])) {
/*
* Handled in the previous loop, so just skip the argument.
*/
i++;
} else if ("-list-edocid".equals(args[i])) {
System.out.println("-list-edocid:");
if ((i + 1) >= args.length) {
System.out.println(usage);
break;
}
;
Document d = reader.document(Integer.parseInt(args[i + 1]));
System.out.println(
"Internal docid --> External docid: " + args[i + 1] + " --> " + d.get("externalId"));
i += 1;
} else if ("-list-docids".equals(args[i])) {
System.out.println("-list-docids:");
for (int j = 0; j < reader.numDocs(); j++) {
Document d = reader.document(j);
System.out.println("Internal --> external docid: " + j + " --> " + d.get("externalId"));
}
;
} else if ("-list-fields".equals(args[i])) {
Fields fields = MultiFields.getFields(reader);
System.out.print("\nNumber of fields: ");
if (fields == null) System.out.println("0");
else {
System.out.println(fields.size());
Iterator<String> is = fields.iterator();
while (is.hasNext()) {
System.out.println("\t" + is.next());
}
;
}
;
} else if ("-list-postings".equals(args[i])) {
if ((i + 2) >= args.length) {
System.out.println(usage);
break;
}
;
listPostings(reader, args[i + 1], args[i + 2], Integer.MAX_VALUE);
i += 2;
} else if ("-list-postings-sample".equals(args[i])) {
if ((i + 2) >= args.length) {
System.out.println(usage);
break;
}
;
listPostings(reader, args[i + 1], args[i + 2], 5);
i += 2;
} else if ("-list-stats".equals(args[i])) {
System.out.println("Corpus statistics:");
System.out.println("\tnumdocs\t\t" + reader.numDocs());
System.out.println(
"\turl:\t"
+ "\tnumdocs="
+ reader.getDocCount("url")
+ "\tsumTotalTF="
+ reader.getSumTotalTermFreq("url")
+ "\tavglen="
+ reader.getSumTotalTermFreq("url") / (float) reader.getDocCount("url"));
System.out.println(
"\tkeywords:"
+ "\tnumdocs="
+ reader.getDocCount("keywords")
+ "\tsumTotalTF="
+ reader.getSumTotalTermFreq("keywords")
+ "\tavglen="
+ reader.getSumTotalTermFreq("keywords") / (float) reader.getDocCount("keywords"));
System.out.println(
"\ttitle:\t"
+ "\tnumdocs="
+ reader.getDocCount("title")
+ "\tsumTotalTF="
+ reader.getSumTotalTermFreq("title")
+ "\tavglen="
+ reader.getSumTotalTermFreq("title") / (float) reader.getDocCount("title"));
System.out.println(
"\tbody:\t"
+ "\tnumdocs="
+ reader.getDocCount("body")
+ "\tsumTotalTF="
+ reader.getSumTotalTermFreq("body")
+ "\tavglen="
+ reader.getSumTotalTermFreq("body") / (float) reader.getDocCount("body"));
System.out.println(
"\tinlink:\t"
+ "\tnumdocs="
+ reader.getDocCount("inlink")
+ "\tsumTotalTF="
+ reader.getSumTotalTermFreq("inlink")
+ "\tavglen="
+ reader.getSumTotalTermFreq("inlink") / (float) reader.getDocCount("inlink"));
} else if ("-list-terms".equals(args[i])) {
if ((i + 1) >= args.length) {
System.out.println(usage);
break;
}
;
listTermDictionary(reader, args[i + 1]);
i += 1;
} else if ("-list-termvector".equals(args[i])) {
if ((i + 1) >= args.length) {
System.out.println(usage);
break;
}
;
listTermVectors(reader, args[i + 1]);
i += 1;
} else if ("-list-termvector-field".equals(args[i])) {
if ((i + 2) >= args.length) {
System.out.println(usage);
break;
}
;
listTermVectorField(reader, args[i + 1], args[i + 2]);
i += 2;
} else System.err.println("\nWarning: Unknown argument " + args[i] + " ignored.");
}
;
/*
* Close the index and exit gracefully.
*/
reader.close();
}
public static void main(String[] args) throws Exception {
int res = new IndexSorterArquivoWeb().doMain(NutchConfiguration.create(), args);
System.exit(res);
}
public static void main(String[] args) throws Exception {
if (args.length < 2) {
System.err.println("TermDumper [-c|-v value] field <index...>");
System.exit(1);
}
boolean count = false;
String value = null;
boolean all = false;
int i = 0;
for (; i < args.length; i++) {
String arg = args[i];
if ("-h".equals(arg) || "--help".equals(arg)) {
System.err.println("TermDumper [-c|-v value] field <index...>");
System.exit(1);
} else if ("-c".equals(arg) || "--count".equals(arg)) {
count = true;
} else if ("-v".equals(arg) || "--vaue".equals(arg)) {
value = args[++i];
} else if ("-a".equals(arg) || "--all".equals(arg)) {
all = true;
} else {
break;
}
}
String field = args[i++];
java.util.ArrayList<IndexReader> readers =
new java.util.ArrayList<IndexReader>(args.length - 1);
for (; i < args.length; i++) {
String arg = args[i];
try {
IndexReader reader = IndexReader.open(new MMapDirectory(new File(arg)), true);
readers.add(reader);
} catch (IOException ioe) {
System.err.println("Error reading: " + arg);
}
}
for (IndexReader reader : readers) {
TermDocs termDocs = reader.termDocs();
TermEnum termEnum = reader.terms(new Term(field));
try {
do {
Term term = termEnum.term();
if (term == null || !field.equals(term.field())) break;
if (value == null) {
if (count) {
termDocs.seek(termEnum);
int c = 0;
for (; termDocs.next(); c++) ;
System.out.print(c + " ");
}
System.out.println(term.text());
} else if (value.equals(term.text())) {
termDocs.seek(termEnum);
while (termDocs.next()) {
if (all) {
Document d = reader.document(termDocs.doc());
System.out.println(termDocs.doc());
for (Object o : d.getFields()) {
Field f = (Field) o;
System.out.println(f.name() + " " + d.get(f.name()));
}
} else {
System.out.println(
termDocs.doc() + " " + reader.document(termDocs.doc()).get("url"));
}
}
}
} while (termEnum.next());
} finally {
termDocs.close();
termEnum.close();
}
}
}
