【发布时间】：2015-08-13 12:53:45
【问题描述】：
** 我想使用 Java 中的正则表达式从 PDF 文档中搜索字母数字文本（发票编号 F0000004511）。我怎样才能做到这一点？例如PDF第一页是这样的：
销售 - 发票 T.I.N. No. 02020600021 Fax No. +91-1792-232268 Invoice No. F0000004511
在 PDF 的第二页发票中，F0000004512 没有变化，第三页和第四页的编号相同。我需要根据发票编号搜索并拆分 pdf 页面。我正在使用 APACHE LUCENE 3.4.0 来索引和搜索 pdf。以下代码用于索引 pdf**
public class Indexer {
private final String sourceFilePath = "G:/PDFCopy"; //give the location of the source files location here
private final String indexFilePath = "G:/searchEngine"; //give the location where you guys want to create index
private IndexWriter writer = null;
private File indexDirectory = null;
private String fileContent; //temp storer of all the text parsed from doc and pdf
private Indexer() throws FileNotFoundException, CorruptIndexException, IOException {
try {
long start = System.currentTimeMillis();
createIndexWriter();
checkFileValidity();
closeIndexWriter();
long end = System.currentTimeMillis();
System.out.println("Total Document Indexed : " + TotalDocumentsIndexed());
System.out.println("Total time" + (end - start) / (100 * 60));
} catch (Exception e) {
System.out.println("Sorry task cannot be completed");
}
}
private void createIndexWriter() {
try {
indexDirectory = new File(indexFilePath);
if (!indexDirectory.exists()) {
indexDirectory.mkdir();
}
FSDirectory dir = FSDirectory.open(indexDirectory);
StandardAnalyzer analyzer = new StandardAnalyzer(Version.LUCENE_34);
IndexWriterConfig config = new IndexWriterConfig(Version.LUCENE_34, analyzer);
writer = new IndexWriter(dir, config);
} catch (Exception ex) {
System.out.println("Sorry cannot get the index writer");
}
}
private void checkFileValidity() {
File[] filesToIndex = new File[100]; // suppose there are 100 files at max
filesToIndex = new File(sourceFilePath).listFiles();
for (File file : filesToIndex) {
try {
//to check whenther the file is a readable file or not.
if (!file.isDirectory()
&& !file.isHidden()
&& file.exists()
&& file.canRead()
&& file.length() > 0.0
&& file.isFile() ) {
if(file.getName().endsWith(".txt")){
indexTextFiles(file);//if the file text file no need to parse text.
System.out.println("INDEXED FILE " + file.getAbsolutePath() + " :-) ");
}
else if(file.getName().endsWith(".doc") || file.getName().endsWith(".pdf")){
//different methof for indexing doc and pdf file.
StartIndex(file);
}
}
} catch (Exception e) {
System.out.println("Sorry cannot index " + file.getAbsolutePath());
}
}
}
public void StartIndex(File file) throws FileNotFoundException, CorruptIndexException, IOException {
fileContent = null;
try {
Document doc = new Document();
if (file.getName().endsWith(".doc")) {
//call the doc file parser and get the content of doc file in txt format
fileContent = new DocFileParser().DocFileContentParser(file.getAbsolutePath());
}
if (file.getName().endsWith(".pdf")) {
//call the pdf file parser and get the content of pdf file in txt format
fileContent = new PdfFileParser().PdfFileParser(file.getAbsolutePath());
}
doc.add(new Field("content", fileContent,
Field.Store.YES, Field.Index.ANALYZED,
Field.TermVector.WITH_POSITIONS_OFFSETS));
doc.add(new Field("filename", file.getName(),
Field.Store.YES, Field.Index.ANALYZED));
doc.add(new Field("fullpath", file.getAbsolutePath(),
Field.Store.YES, Field.Index.ANALYZED));
if (doc != null) {
writer.addDocument(doc);
}
System.out.println("Indexed" + file.getAbsolutePath());
} catch (Exception e) {
System.out.println("error in indexing" + (file.getAbsolutePath()));
}
}
private void indexTextFiles(File file) throws CorruptIndexException, IOException {
Document doc = new Document();
doc.add(new Field("content", new FileReader(file)));
doc.add(new Field("filename", file.getName(),
Field.Store.YES, Field.Index.ANALYZED));
doc.add(new Field("fullpath", file.getAbsolutePath(),
Field.Store.YES, Field.Index.ANALYZED));
if (doc != null) {
writer.addDocument(doc);
}
}
private int TotalDocumentsIndexed() {
try {
IndexReader reader = IndexReader.open(FSDirectory.open(indexDirectory));
return reader.maxDoc();
} catch (Exception ex) {
System.out.println("Sorry no index found");
}
return 0;
}
private void closeIndexWriter() {
try {
writer.optimize();
writer.close();
} catch (Exception e) {
System.out.println("Indexer Cannot be closed");
}
}
public static void main(String arg[]) {
try {
new Indexer();
} catch (Exception ex) {
System.out.println("Cannot Start :(");
}
}
}
下面是在索引中搜索的代码。在这里，我直接通过正则表达式进行搜索。但是否可以在所有 pdf 中使用正则表达式值进行搜索并阅读发票编号。最后我需要根据发票编号拆分 pdf。 我需要从正则表达式返回发票没有值并拆分 tht pdf。 （来源 pdf 有 60 页，具有唯一且重复的发票编号）
public class Searcher {
public Searcher(String searchString) {
try {
IndexSearcher searcher = new IndexSearcher(FSDirectory.open(
new File("G:/searchEngine")));
Analyzer analyzer1 = new StandardAnalyzer(Version.LUCENE_34);
QueryParser queryParser = new QueryParser(Version.LUCENE_34, "content", analyzer1);
QueryParser queryParserfilename = new QueryParser(Version.LUCENE_34, "fullpath", analyzer1);
Query query = queryParser.parse(searchString);//to search in the content
Query queryfilename = queryParserfilename.parse(searchString);//to search the file name only
TopDocs hits = searcher.search(query, 10000); //for
ScoreDoc[] document = hits.scoreDocs;
System.out.println("Total no of hits for content: " + hits.totalHits);
for (int i = 0; i < document.length; i++) {
Document doc = searcher.doc(document[i].doc);
String filePath = doc.get("fullpath");
System.out.println(filePath);
}
} catch (Exception e) {
}
}
public static void main(String args[])
{
new Searcher("Invoice No.\\s\\w\\d\\d\\d\\d\\d\\d\\d\\d\\d\\d");
}
}
【问题讨论】：
-
好吧，您似乎正在使用 QueryParser 在 Lucene 版本 3.4 中生成查询。我相信 QueryParser 直到 4.0 版才添加正则表达式支持。要使用正则表达式进行搜索，您需要手动构造 RegexQuery。