fileutils.java

Jena推理机

/*
 * (c) Copyright 2002, 2003, 2004, 2005, 2006, 2007 Hewlett-Packard Development Company, LP
 * [See end of file]
 */

package com.hp.hpl.jena.util;

import java.io.*;

import java.net.URL;
import java.nio.charset.Charset ;

import org.apache.commons.logging.LogFactory;

import com.hp.hpl.jena.n3.RelURI;
import com.hp.hpl.jena.shared.JenaException;
import com.hp.hpl.jena.shared.WrappedIOException;
import com.hp.hpl.jena.JenaRuntime ;

public class FileUtils
{
    public static final String langXML          = "RDF/XML" ;
    public static final String langXMLAbbrev    = "RDF/XML-ABBREV" ;
    public static final String langNTriple      = "N-TRIPLE" ;
    public static final String langN3           = "N3" ;
    public static final String langTurtle       = "TURTLE" ;
    // Non-standard
    public static final String langBDB          = "RDF/BDB" ;
    public static final String langSQL          = "RDF/SQL" ;
    
    /** Java name for UTF-8 encoding */
    public static final String encodingUTF8     = "utf-8" ;
    
    static Charset utf8 = null ;
    static {
        try {
            utf8 = Charset.forName(encodingUTF8) ;
        } catch (Throwable ex)
        {
            LogFactory.getLog(FileUtils.class).warn("Failed to get charset for UTF-8") ;
        }
    }
    
    
    /** Create a reader that uses UTF-8 encoding */ 
    
    static public Reader asUTF8(InputStream in) {
        if ( JenaRuntime.runUnder(JenaRuntime.featureNoCharset) )
            return new InputStreamReader(in) ;
        // Not ,utf8 -- GNUClassPath (0.20) apparently fails on passing in a charset
        // but if passed not the decoder or the name of the charset.
        // Reported and fixed.
        return new InputStreamReader(in, utf8.newDecoder());
    }

    /** Create a buffered reader that uses UTF-8 encoding */ 
    
    static public BufferedReader asBufferedUTF8(InputStream in) {
        return new BufferedReader(asUTF8(in)) ;
    }

    /** Create a writer that uses UTF-8 encoding */ 

    static public Writer asUTF8(OutputStream out) {
        if ( JenaRuntime.runUnder(JenaRuntime.featureNoCharset) )
            return new OutputStreamWriter(out) ;
        return new OutputStreamWriter(out, utf8.newEncoder());
    }

    /** Create a print writer that uses UTF-8 encoding */ 

    static public PrintWriter asPrintWriterUTF8(OutputStream out) {
        return new PrintWriter(asUTF8(out)); 
    }
    
    /** Guess the language/type of model data. Updated by Chris, hived off the
     * model-suffix part to FileUtils as part of unifying it with similar code in FileGraph.
     * 
     * <ul>
     * <li> If the URI of the model starts jdbc: it is assumed to be an RDB model</li>
     * <li> If the URI ends ".rdf", it is assumed to be RDF/XML</li>
     * <li> If the URI end .nt, it is assumed to be N-Triples</li>
     * <li> If the URI end .bdb, it is assumed to be BerkeleyDB model [suppressed at present]</li>
     * </ul>
     * @param name    URL to base the guess on
     * @param otherwise Default guess
     * @return String   Guessed syntax - or the default supplied
     */

    public static String guessLang( String name, String otherwise )
    {
        if ( name.startsWith("jdbc:") || name.startsWith("JDBC:") )
            return langSQL ;
        
        String suffix = getFilenameExt( name );
        if (suffix.equals( "n3" ))   return langN3;
        if (suffix.equals( "nt" ))   return langNTriple;
        if (suffix.equals( "ttl" ))  return langTurtle ;
        if (suffix.equals( "rdf" ))  return langXML;
        if (suffix.equals( "owl" ))  return langXML;
        return otherwise; 
    }    
   
   
    /** Guess the language/type of model data
     * 
     * <ul>
     * <li> If the URI of the model starts jdbc: it is assumed to be an RDB model</li>
     * <li> If the URI ends .rdf, it is assumed to be RDF/XML</li>
     * <li> If the URI ends .n3, it is assumed to be N3</li>
     * <li> If the URI ends .nt, it is assumed to be N-Triples</li>
     * <li> If the URI ends .bdb, it is assumed to be BerkeleyDB model</li>
     * </ul>
     * @param urlStr    URL to base the guess on
     * @return String   Guessed syntax - default is RDF/XML
     */

    public static String guessLang(String urlStr)
    {
        return guessLang(urlStr, langXML) ;
    }

    /** Turn a file: URL or file name into a plain file name */
    
    public static String toFilename(String filenameOrURI)
    {
        // Requirements of windows and Linux differ slightly here
        // Windows wants "file:///c:/foo" => "c:/foo"
        // but Linux only wants "file:///foo" => "/foo"
        // Pragmatically, a path of "/c:/foo", or "/foo" works everywhere.
        // but not "//c:/foo" or "///c:/foo" 
        // else IKVM thinks its a network path on Windows.
        
        // If it's a a file: we apply %-decoding.
        // If there is no scheme name, we don't.

        if ( !isFile(filenameOrURI) )
            return null ;
        // No scheme of file:
        String fn = filenameOrURI ;

        if ( ! fn.startsWith("file:") )
            return fn ;
        
        // file:
        // Convert absolute file names
        if ( fn.startsWith("file:///") )
            fn = fn.substring("file://".length()) ;
        else if ( fn.startsWith("file://localhost/") )
            // NB Leaves the leading slash on. 
            fn = fn.substring("file://localhost".length()) ;
        else
            // Just trim off the file:
            fn = fn.substring("file:".length()) ;

        return decodeFileName(fn) ;
    }
    
    public static String decodeFileName(String s)
    {
        if ( s.indexOf('%') < 0 ) 
            return s ;
        int len = s.length();
        StringBuffer sbuff = new StringBuffer(len) ;

        // This is URIRef.decode()? Is that code used?
        // Just decode % escapes.
        // Not http://www.daml.org/2001/03/daml+oil
        for ( int i =0 ; i < len ; i++ )
        {
            char c = s.charAt(i);
            switch (c)
            {
                case '%':
                    int codepoint = Integer.parseInt(s.substring(i+1,i+3),16) ;
                    char ch = (char)codepoint ;
                    sbuff.append(ch) ;
                    i = i+2 ;
                    break ;
                default:
                    sbuff.append(c);
            }
        }
        return sbuff.toString();
    }
    
    
    /** Turn a plain filename into a "file:" URL */
    public static String toURL(String filename)
    {
        if ( filename.startsWith("file:") )
            return filename ;
        filename = encodeFileName(filename) ;
        return RelURI.resolveFileURL(filename) ;
    }
    
    public static String encodeFileName(String s)
    {
        int len = s.length();
        StringBuffer sbuff = new StringBuffer(len) ;

        // Convert a few charcaters that occur in filenames into a safe form.
        for ( int i = 0 ; i < len ; i++ )
        {
            char c = s.charAt(i);
            switch (c)
            {
                case ' ': case '~':
                    sbuff.append('%') ;
                    sbuff.append(Integer.toHexString(c).toUpperCase()) ;
                    break ;
                default:
                    sbuff.append(c);
            }
        }
        return sbuff.toString();
    }
    
    /** Check whether 'name' is possibly a file reference  
     * 
     * @param name
     * @return boolean False if clearly not a filename. 
     */
    public static boolean isFile(String name)
    {
        String scheme = getScheme(name) ;
        
        if ( scheme == null  )
            // No URI scheme - treat as filename
            return true ;
        
        if ( scheme.equals("file") )
            // file: URI scheme
            return true ;
            
        // Windows: "c:" etc
        if ( scheme.length() == 1 )
            // file: URI scheme
            return true ;