/* * DataSetURI.java * * Created on March 31, 2007, 7:54 AM * * To change this template, choose Tools | Template Manager * and open the template in the editor. */ package org.autoplot.datasource; import java.util.logging.Level; import org.das2.util.monitor.ProgressMonitor; import org.das2.util.monitor.NullProgressMonitor; import org.das2.util.filesystem.FileObject; import org.das2.util.filesystem.FileSystem; import ftpfs.FTPBeanFileSystemFactory; import java.io.BufferedReader; import java.io.File; import java.io.FileInputStream; import java.io.FileNotFoundException; import java.io.FileOutputStream; import java.io.FileReader; import java.io.IOException; import java.io.InputStream; import java.io.InterruptedIOException; import java.io.OutputStream; import java.net.HttpURLConnection; import java.net.MalformedURLException; import java.net.URL; import java.net.URI; import java.net.URISyntaxException; import java.net.URLConnection; import java.net.UnknownHostException; import java.nio.channels.Channels; import java.nio.charset.Charset; import java.text.ParseException; import java.util.ArrayList; import java.util.Arrays; import java.util.Collections; import java.util.Comparator; import java.util.LinkedHashMap; import java.util.LinkedList; import java.util.List; import java.util.Map; import java.util.Map.Entry; import java.util.WeakHashMap; import java.util.logging.Logger; import java.util.regex.Matcher; import java.util.regex.Pattern; import java.util.zip.GZIPInputStream; import org.autoplot.wgetfs.WGetFileSystemFactory; import org.das2.datum.DatumRange; import org.das2.datum.DatumRangeUtil; import org.das2.fsm.FileStorageModel; import org.das2.util.DasProgressMonitorInputStream; import org.das2.util.LoggerManager; import org.das2.util.filesystem.FileSystemSettings; import org.das2.util.filesystem.HtmlUtil; import org.das2.util.filesystem.LocalFileSystem; import org.das2.util.filesystem.URIException; import org.das2.util.filesystem.VFSFileSystemFactory; import org.das2.util.filesystem.WebFileSystem; import org.autoplot.aggregator.AggregatingDataSourceFactory; import org.autoplot.aggregator.AggregatingDataSourceFormat; import org.autoplot.datasource.capability.TimeSeriesBrowse; import org.das2.datum.HttpUtil; import org.das2.qds.ops.Ops; import org.das2.util.Base64; import org.das2.util.filesystem.KeyChain; import org.das2.util.monitor.CancelledOperationException; /** * * Works with DataSourceRegistry to translate a URI into a DataSource. Also, * will provide completions. * * @author jbf * */ public class DataSetURI { private static final Object ACTION_WAIT_EXISTS = "WAIT_EXISTS"; private static final Object ACTION_DOWNLOAD = "DOWNLOAD"; private static final Object ACTION_USE_CACHE = "USE_CACHE"; private static final Logger logger = LoggerManager.getLogger("apdss.uri"); static { logger.fine("load class DataSetURI"); DataSourceRegistry registry = DataSourceRegistry.getInstance(); registry.discoverFactories(); registry.discoverRegistryEntries(); } static { FileSystem.registerFileSystemFactory("zip", new zipfs.ZipFileSystemFactory()); FileSystem.registerFileSystemFactory("ftp", new FTPBeanFileSystemFactory()); if ( System.getProperty("AP_CURL")!=null || System.getProperty("AP_WGET")!=null ) { // TODO: this only handles HTTP and HTTPS. FTP should probably be handled as well, but check curl. FileSystem.registerFileSystemFactory("http", new WGetFileSystemFactory() ); FileSystem.registerFileSystemFactory("https", new WGetFileSystemFactory() ); FileSystem.registerFileSystemFactory("ftp", new WGetFileSystemFactory() ); logger.fine("using wget implementation for http,https and ftp because AP_CURL or AP_WGET is set."); } // The following is commented out until the svn version of dasCore.jar is updated FileSystem.registerFileSystemFactory("sftp", new VFSFileSystemFactory()); FileSystem.settings().setPersistence(FileSystemSettings.Persistence.EXPIRES); if (FileSystemSettings.hasAllPermission()) { File apDataHome = new File( AutoplotSettings.settings().resolveProperty( AutoplotSettings.PROP_FSCACHE ) ); FileSystem.settings().setLocalCacheDir(apDataHome); } } static WeakHashMap dsToFactory = new WeakHashMap<>(); /** * returns the explicit extension, or the file extension if found, or null. * The extension will not contain a period. * Inputs include: * @param surl * @return the extension found, without the ".", or null if no period is found in the filename. */ public static String getExt(String surl) { if ( surl==null ) throw new NullPointerException(); String explicitExt = getExplicitExt(surl); if (explicitExt != null) { return explicitExt; } else { URISplit split = URISplit.parse(surl); if (split.file != null) { int i0 = split.file.lastIndexOf('/'); if (i0 == -1) return null; int i1 = split.file.lastIndexOf('.'); if (i1 != -1 && i1 > i0) { return split.file.substring(i1 + 1); } else { return null; } } else { if ( !surl.contains("/") && !surl.contains("\\") ) { // \\ is for Windows. int i= surl.lastIndexOf("."); if ( i>-1 ) { return surl.substring(i+1); } else { return null; } } else { return null; } } } } /** * return the extension prefix of the URI, if specified. * @param surl * @return null or an extension like "tsds" */ public static String getExplicitExt(String surl) { URISplit split = URISplit.parse(surl); if ( split.vapScheme==null ) return null; int i = split.vapScheme.indexOf('+'); if (i != -1) { return split.vapScheme.substring(i + 1); } else { return null; } } /** * get the data source for the URI. * @param uri the URI. * @return the data source from which the data set can be retrieved. * @throws java.lang.Exception when the DataSourceFactory throws an Exception * @throws IllegalArgumentException if the url extension is not supported. */ public static DataSource getDataSource(URI uri) throws Exception { DataSourceFactory factory = getDataSourceFactory(uri, new NullProgressMonitor()); if ( factory==null ) { throw new IllegalArgumentException("unable to resolve URI: "+uri); } DataSource result = factory.getDataSource(uri); dsToFactory.put(result, factory); return result; } /** * get the data source for the URI. * @param suri the URI. * @return the data source from which the data set can be retrieved. * @throws java.lang.Exception when the DataSourceFactory throws an Exception * @throws IllegalArgumentException if the url extension is not supported. */ public static DataSource getDataSource(String suri) throws Exception { return getDataSource( getURIValid(suri) ); } /** * Prefix the URL with the datasource extension if necessary, so that * the URL would resolve to the dataSource. This is to support TimeSeriesBrowse, * and any time a resouce URL must be understood out of context. * * TODO: note ds.getURI() should return the fully-qualified URI, so this is * no longer necessary. * * @param ds the data source. * @return the canonical URI. */ public static String getDataSourceUri(DataSource ds) { DataSourceFactory factory = dsToFactory.get(ds); if (factory instanceof AggregatingDataSourceFactory) { return ds.getURI(); } if (factory == null) { return ds.getURI(); // nothing we can do } else { URISplit split = URISplit.parse(ds.getURI()); String fext; fext = DataSourceRegistry.getInstance().getExtensionFor(factory).substring(1); if (DataSourceRegistry.getInstance().hasSourceByExt(split.ext)) { DataSourceFactory f2 = DataSourceRegistry.getInstance().getSource(split.ext); if (!factory.getClass().isInstance(f2)) { split.vapScheme = "vap+" + fext; } } else { split.vapScheme = "vap+" + fext; } return URISplit.format(split); } } /** * check that the string uri is aggregating by looking for %Y's (etc) in the * file part of the URI. This also looks for: * * @param surl * @return */ public static boolean isAggregating(String surl) { if ( !DataSourceRegistry.getInstance().hasResourceUri(surl) ) { return false; } int iquest = surl.indexOf('?'); if ( iquest>0 ) surl= surl.substring(0,iquest); surl= surl.replaceAll("%25", "%"); int ipercy = surl.lastIndexOf("%Y"); if (ipercy == -1) ipercy = surl.lastIndexOf("$Y"); if (ipercy == -1) ipercy = surl.lastIndexOf("%y"); if (ipercy == -1) ipercy = surl.lastIndexOf("$y"); if (ipercy == -1) ipercy = surl.lastIndexOf("$(o"); if (ipercy == -1) ipercy = surl.lastIndexOf("%{o"); if (ipercy == -1) ipercy = surl.lastIndexOf("$(periodic"); if (ipercy == -1) ipercy = surl.lastIndexOf("%{periodic"); if (ipercy == -1) ipercy = surl.lastIndexOf("$v"); if (ipercy == -1) ipercy = surl.lastIndexOf("$(v"); if (ipercy == -1) ipercy = surl.lastIndexOf("$x"); if (ipercy == -1) ipercy = surl.lastIndexOf('*'); return ipercy != -1; } /** * taken from unaggregate.jy in the servlet. * @param resourceURI resource URI like "file://tmp/data$Y$m$d.dat" * @param timerange a timerange that will be covered by the span. * @return the strings resolved. * @throws org.das2.util.filesystem.FileSystem.FileSystemOfflineException * @throws java.net.UnknownHostException */ public static String[] unaggregate( String resourceURI, DatumRange timerange ) throws FileSystem.FileSystemOfflineException, UnknownHostException, IOException { int i= AggregatingDataSourceFactory.splitIndex( resourceURI ); String root= resourceURI.substring(0,i); // the static part of the name String template= resourceURI.substring(i); // the templated part of the name FileSystem fs= FileSystem.create( root ); FileStorageModel fsm= FileStorageModel.create( fs, template ); String[] names= fsm.getNamesFor( timerange ); List result= new ArrayList<>(); for ( String n: names ) { result.add( root + n ); } return result.toArray( new String[result.size()] ); } /** * returns the URI to be interpreted by the DataSource. This identifies * a file (or database) resource that can be passed to VFS. * @param uri, the URI understood in the context of all datasources. This should contain "vap" or "vap+" for the scheme. * @return the URI for the datasource resource, or null if it is not valid. */ public static URI getResourceURI(URI uri) { URISplit split = URISplit.parse(uri); return split.resourceUri; } /** * returns the URI to be interpreted by the DataSource. For file-based * data sources, this will probably be the filename plus server-side * parameters, and can be converted to a URL. * * Changes: * 20090916: client-side parameters removed from URI. * @param surl, the URI understood in the context of all datasources. This should contain "vap" or "vap+" for the scheme. * @return the URI for the datasource resource, or null if it is not valid. */ public static URI getResourceURI(String surl) { if ( surl.matches( "file\\:[A-Z]\\:\\\\.*") ) { surl= "file://" + surl.substring(5).replace('\\','/'); } if ( surl.matches( "file\\:/[A-Z]\\:\\\\.*") ) { surl= "file://" + surl.substring(5).replace('\\','/'); } URISplit split = URISplit.parse(surl); return split.resourceUri; } /** * returns a downloadable URL from the Autoplot URI, perhaps popping off the * data source specifier. This assumes that the resource is a URL, * and getResourceURI().toURL() should be used to handle all cases. * * @param uri An Autoplot URI. * @return a URL that can be downloaded, or null if it is not found. */ public static URL getWebURL(URI uri) { try { URI uri1= getResourceURI(uri); if ( uri1==null ) return null; URL rurl = uri1.toURL(); String surl = rurl.toString(); return new URL(surl); } catch (MalformedURLException ex) { throw new RuntimeException(ex); } } /** * creates a new URI from the new URI, in the context of the old URI. For * example, if the old URI had parameters and the new is just a file, then * use the old URI but replace the file. * @param context * @param newUri * @return */ static String newUri(String context, String newUri) { URISplit scontext = URISplit.parse(context,0,false); URISplit newURLSplit = URISplit.parse(newUri); if (newURLSplit.file != null && !newURLSplit.file.equals("")) scontext.file = newURLSplit.file; if (newURLSplit.params != null && !newURLSplit.params.equals("")) scontext.params = newURLSplit.params; return URISplit.format(scontext); } /** * create the URI without the timerange. * @param value a uri. e.g. /tmp/foo$Y$m$d.dat?timerange=2014-001 * @return null or the value without the timerange, e.g. /tmp/foo$Y$m$d.dat */ public static String blurTsbUri(String value) { try { DataSourceFactory dsf= getDataSourceFactory( toUri(value), new NullProgressMonitor() ); if ( dsf==null ) return null; // I was getting this because I removed a datasource (HAPI) TimeSeriesBrowse tsb= dsf.getCapability( TimeSeriesBrowse.class ); if (tsb==null ) return null; tsb.setURI(value); return tsb.blurURI(); } catch (URISyntaxException | IOException | IllegalArgumentException | ParseException ex) { return null; } } /** * create the URI without the timerange. * @param value a uri. e.g. /tmp/foo$Y$m$d.dat?timerange=2014-001 * @return null or the value without the timerange, e.g. /tmp/foo$Y$m$d.dat */ public static String blurTsbResolutionUri(String value) { try { DataSourceFactory dsf= getDataSourceFactory( new URI(value), new NullProgressMonitor() ); TimeSeriesBrowse tsb= dsf.getCapability( TimeSeriesBrowse.class ); if (tsb==null ) { logger.fine("Unable to update the URI because factory doesn't provide TSB"); return null; } tsb.setURI(value); tsb.setTimeResolution(null); return tsb.getURI(); } catch (URISyntaxException | IOException | IllegalArgumentException | ParseException ex) { return null; } } /** * return the URI with a new time. * @param value * @param timeRange * @return the URI with a new time. */ public static String resetUriTsbTime( String value, DatumRange timeRange ) { if ( timeRange==null ) return value; try { DataSourceFactory dsf= getDataSourceFactory( new URI(value), new NullProgressMonitor() ); TimeSeriesBrowse tsb= dsf.getCapability( TimeSeriesBrowse.class ); if (tsb==null ) { logger.fine("Unable to update the URI because factory doesn't provide TSB"); return null; } tsb.setURI(value); //TODO: I bet we don't need to do this. tsb.setTimeRange(timeRange); return tsb.getURI(); } catch (URISyntaxException | IOException | IllegalArgumentException | ParseException ex) { return null; } } /** * return a human-readable abbreviation of the URI, limiting to len characters. * @param ssuri * @param len * @return a string of length no more than len characters */ public static String abbreviateForHumanComsumption(String ssuri, int len) { if ( ssuri.length()>len ) { return "..." + ssuri.substring(ssuri.length()-len-3); } else { return ssuri; } } // mark the special case where a resource is actually a folder. public static class NonResourceException extends IllegalArgumentException { public NonResourceException(String msg) { super(msg); } } /** * for now, just hide the URI stuff from clients, let's not mess with factories * @param uri the URI describing the output format and any arguments. * @return the DataSourceFormat that formats a dataset to the format. */ public static DataSourceFormat getDataSourceFormat(URI uri) { int i = uri.getScheme().indexOf('.'); String ext; if ( isAggregating(uri.toString()) ) { DataSourceFormat agg= new AggregatingDataSourceFormat(); return agg; } if (i != -1) { ext = uri.getScheme().substring(0, i); } else { int i2 = uri.getScheme().indexOf('+'); if ( i2!=-1 ) { ext= uri.getScheme().substring(i2+1); } else { URL url = getWebURL(uri); String file = url.getPath(); i = file.lastIndexOf('.'); ext = i == -1 ? "" : file.substring(i); } } return DataSourceRegistry.getInstance().getFormatByExt(ext); } /** * get the datasource factory for the URL. This has the rarely-used * logic that looks up MIME types for HTTP requests. * @param uri the URI of the data source. * @param mon progress monitor * @return the factory that produces the data source. * @throws java.io.IOException * @throws URISyntaxException if the schemeSpecficPart is not itself a URI. * TODO: this should probably throw UnrecognizedDataSourceException */ public static DataSourceFactory getDataSourceFactory( URI uri, ProgressMonitor mon) throws IOException, IllegalArgumentException, URISyntaxException { String suri= DataSetURI.fromUri(uri); //suri= URISplit.makeCanonical(suri); //TODO: I was seeing an error and did this fix when playing with GUI testing. if ( isAggregating( suri ) ) { String eext = DataSetURI.getExplicitExt( suri ); if (eext != null) { DataSourceFactory delegateFactory = DataSourceRegistry.getInstance().getSource(eext); AggregatingDataSourceFactory factory = new AggregatingDataSourceFactory(); factory.setDelegateDataSourceFactory(delegateFactory); return factory; } else { return new AggregatingDataSourceFactory(); } } // The user explicitly specified the source. E.g. vap+cdaweb:... String ext = DataSetURI.getExplicitExt( suri ); if (ext != null && !suri.startsWith("vap+X:") ) { return DataSourceRegistry.getInstance().getSource(ext); } URI resourceUri; // The scheme-specific part of the URI must itself be a URI, typically a URL pointing at the data. URIs are used // to support other protocols like sftp. String resourceSuri = uri.getRawSchemeSpecificPart(); if ( resourceSuri.startsWith("'") ) { throw new IllegalArgumentException("URI starts with single quote"); } resourceUri = new URI(resourceSuri); //bug3055130 okay ext = DataSetURI.getExt(uri.toString()); if (ext == null) ext = ""; DataSourceFactory factory; // see if we can identify it by ext, to avoid the head request. factory = DataSourceRegistry.getInstance().getSource(ext); // rte_1512402504_20121004_002144.xml: actually I think it was the parens on the next expression. if (factory == null && ( resourceUri.getScheme()!=null && ( resourceUri.getScheme().equals("http") || resourceUri.getScheme().equals("https") ) ) ) { // get the mime type URL url = resourceUri.toURL(); mon.setTaskSize(-1); mon.started(); mon.setProgressMessage("doing HEAD request to find dataset type"); try { URLConnection c = url.openConnection(); c= HtmlUtil.checkRedirect(c); c.setConnectTimeout( FileSystem.settings().getConnectTimeoutMs() ); c.setReadTimeout( FileSystem.settings().getReadTimeoutMs() ); String mime = c.getContentType(); if (mime == null) { throw new IOException("failed to connect"); } String cd = c.getHeaderField("Content-Disposition"); // support VxOWare if (cd != null) { Pattern p= Pattern.compile(".*filename=\"?(.+)\"?"); Matcher m= p.matcher(cd); if ( m.matches() ) { String filename = m.group(1); int i0 = filename.lastIndexOf('.'); ext = filename.substring(i0); factory = DataSourceRegistry.getInstance().getSource(ext); } } if ( factory==null ) factory = DataSourceRegistry.getInstance().getSourceByMime(mime); if ( c instanceof HttpURLConnection ) { ((HttpURLConnection)c).disconnect(); } } finally { mon.finished(); } } // maybe it was actually a directory if (factory == null) { if (ext.equals("") || ext.equals("X") ) { throw new NonResourceException("resource has no extension or mime type"); } else { factory = DataSourceRegistry.getInstance().getSource(ext); } } if (factory == null) { throw new IllegalArgumentException("Unsupported extension: " + ext); } return factory; } /** * get the InputStream from the path part of the URI. The stream must be closed by the client. * * @param url URL like http://autoplot.org/data/autoplot.dat * @param mon monitor that will monitor the stream as it is transmitted. * @return the InputStream, which must be closed by the client. TODO: check usages... * @throws IOException */ public static InputStream getInputStream(URL url, ProgressMonitor mon) throws IOException { URISplit split = URISplit.parse(url.toString()); try { URI spath = getWebURL( DataSetURI.toUri(split.path)).toURI(); FileSystem fs = FileSystem.create(spath); FileObject fo = fs.getFileObject(split.file.substring(split.path.length())); if (!fo.isLocal()) { logger.log(Level.FINE, "getInputStream(URL): downloading file {0} from {1}", new Object[] { fo.getNameExt(), url.toString() } ); } return fo.getInputStream(mon); } catch (URISyntaxException ex) { throw new IOException("URI Syntax Exception: " + ex.getMessage()); } } /** * get the InputStream from the path part of the URI. The stream must be closed by the client. * * @param uri URI like vap+dat:http://autoplot.org/data/autoplot.dat * @param mon monitor that will monitor the stream as it is transmitted. * @return the InputStream, which must be closed by the client. TODO: check usages... * @throws IOException */ public static InputStream getInputStream(URI uri, ProgressMonitor mon) throws IOException { logger.entering("DataSetURI", "getInputStream", uri ); URISplit split = URISplit.parse( uri ); FileSystem fs; fs = FileSystem.create( DataSetURI.toUri(split.path) ); String filename = split.file.substring(split.path.length()); if (fs instanceof LocalFileSystem) filename = DataSourceUtil.unescape(filename); FileObject fo = fs.getFileObject(filename); if (!fo.isLocal()) { logger.log(Level.FINE, "getInputStream(URI): downloading file {0} from {1}{2}", new Object[] { fo.getNameExt(), fs.getRootURI(), filename } ); } InputStream result= fo.getInputStream(mon); logger.exiting("DataSetURI", "getInputStream" ); return result; } /** * canonical method for converting string from the wild into a URI-safe string. * This contains the code that converts a colloquial string URI into a * properly formed URI. * For example: * space is converted to "%20" * %Y is converted to $Y * This does not add file: or vap:. Pluses are only changed in the params part. * @param suri * @throws IllegalArgumentException if the URI cannot be made safe. * @return */ public static URI toUri( String suri ) { try { if ( !URISplit.isUriEncoded(suri) ) { suri = suri.replaceAll("%([^0-9])", "%25$1"); //suri = suri.replaceAll("#", "%23" ); //surl = surl.replaceAll("%", "%25" ); // see above //suri = suri.replaceAll("&", "%26" ); //surl = surl.replaceAll("/", "%2F" ); //surl = surl.replaceAll(":", "%3A" ); //suri = suri.replaceAll(";", "%3B" ); suri = suri.replaceAll("<", "%3C"); suri = suri.replaceAll(">", "%3E"); //suri = suri.replaceAll("\\?", "%3F" ); suri = suri.replaceAll(" ", "%20"); suri = suri.replaceAll("\\[", "%5B"); // Windows appends these in temporary download rte_1495358356 suri = suri.replaceAll("\\]", "%5D"); suri = suri.replaceAll("\\^", "%5E"); } if ( suri.startsWith("\\") ) { return new File(suri).toURI(); } return new URI(suri); //bug 3055130 okay } catch (URISyntaxException ex) { throw new IllegalArgumentException(ex); } } /** * canonical method for converting URI to human-readable string, containing * spaces and other illegal characters. Note pluses in the query part * are interpreted as spaces. * See also URISplit.uriDecode,etc. * @param uri * @return */ public static String fromUri( URI uri ) { //String surl= uri.getScheme() + ":" + uri.getSchemeSpecificPart(); // This is how we should do this! String surl= uri.toString(); int i= surl.indexOf('?'); String query= i==-1 ? "" : surl.substring(i); if ( i!=-1 ) { //if ( query.contains("+") && !query.contains("%20") ) { // query = query.replaceAll("\\+", " " ); //} return URISplit.uriDecode(surl.substring(0,i)) + query; } else { return URISplit.uriDecode(surl); } } /** * canonical method for converting from File to human-readable string. This * simply tacks on "file://" to the filename. This was introduced so that there * is one canonical way to do this. * @param file * @return */ public static String fromFile( File file ) { String s= file.getAbsolutePath().replaceAll("\\\\","/"); if ( s.length()>0 && s.charAt(0)!='/' ) { return "file:///"+s; // Windows C:/tmp/myfile.vap } else { return "file://"+s; } } /** * Legacy behavior was to convert pluses into spaces in URIs. This caused problems * distinguishing spaces from pluses, so we dropped this as the default behavior. * If data sources are to support legacy URIs, then they should use this routine * to mimic the behavior. * This checks if the URI already contains spaces, and will not convert if there * are already spaces. * @param ssheet * @return */ public static String maybePlusToSpace(String ssheet) { if ( ssheet.contains(" ") ) return ssheet; return ssheet.replaceAll("\\+"," "); } /** * check that the file has length>0 and throw EmptyFileException if it does. This * code is the standard way this should be done. * @param file * @throws EmptyFileException */ public static void checkLength( File file ) throws EmptyFileException { if ( file.length()==0 ) { throw new EmptyFileException(file); } } /** * Check that the file does not contain html content that was intended for human consumption. This hopes to solve * the problem where an html login screen is presented in hotels, etc. This is looking for "html" or "doc" at the * beginning of the file, see DataSourceUtil.isHtmlStream. * * @param tfile the local copy of the file. * @param source the source of the file. * @throws HtmlResponseIOException * @throws FileNotFoundException */ private static void checkNonHtml( File tfile, URL source ) throws HtmlResponseIOException, FileNotFoundException { FileInputStream fi= null; HtmlResponseIOException ex2=null; try { fi = new FileInputStream(tfile); byte[] magic = new byte[5]; int bytes = fi.read(magic); if ( bytes==5 ) { String ss= new String(magic,"UTF-8"); if ( DataSourceUtil.isHtmlStream(ss) ) { ex2= new HtmlResponseIOException( "file appears to be html: "+tfile, source ); } } } catch (IOException ex) { logger.log(Level.SEVERE, ex.getMessage(), ex); } finally { try { if ( fi!=null ) fi.close(); } catch (IOException ex) { logger.log(Level.SEVERE, ex.getMessage(), ex); } } if ( ex2!=null ) throw ex2; } /** * return a file reference for the url. This is initially to fix the problem * for Windows where new URL( "file://c:/myfile.dat" ).getPath() -> "/myfile.dat". * This will use a temporary local file in some cases, such as when the URL * has parameters, which prevent use with the FileSystem model. * * TODO: why is there both getFile(url,mon) and getFile( suri, allowHtml, mon )??? * * @param url the URL of the file. * @param mon progress monitor * @return the File * @throws java.io.IOException * @see #getFile(java.lang.String, boolean, org.das2.util.monitor.ProgressMonitor) */ public static File getFile(URL url, ProgressMonitor mon) throws IOException { URISplit split = URISplit.parse(url.toString()); try { if ( split.path==null || split.path.length()==0 ) { throw new IllegalArgumentException("expected file but didn't find one, check URI for question mark"); } FileSystem fs = FileSystem.create( getWebURL( toUri(split.path) ).toURI() ); String filename = split.file.substring(split.path.length()); if (fs instanceof LocalFileSystem) filename = DataSourceUtil.unescape(filename); FileObject fo = fs.getFileObject(filename); if (!fo.isLocal()) { logger.log(Level.FINE, "getFile: downloading file {0} from {1}", new Object[] { fo.getNameExt(), url.toString() } ); } else { logger.log(Level.FINE, "using local copy of {0}", fo.getNameExt()); } File tfile; if ( fo.exists() ) { tfile = fo.getFile(mon); //TODO: there's a bug here: where we rename the file after unzipping it, but we don't check to see if the .gz is newer. checkNonHtml( tfile, url ); } else { FileObject foz= fs.getFileObject(filename+".gz"); // repeat the .gz logic that FileStorageModel.java has. if ( foz.exists() ) { File fz= foz.getFile(mon); checkNonHtml( fz, url ); File tfile1= new File( fz.getPath().substring(0, fz.getPath().length() - 3) + ".temp" ); tfile= new File( fz.getPath().substring(0, fz.getPath().length() - 3 ) ); org.das2.util.filesystem.FileSystemUtil.gunzip( fz, tfile1); if ( tfile.exists() ) { if ( ! tfile.delete() ) { throw new IllegalArgumentException("unable to delete "+tfile ); } } // it shouldn't, but to be safe... if ( ! tfile1.renameTo(tfile) ) { throw new IllegalArgumentException("unable to rename "+tfile1 + " to "+ tfile ); } } else { //repeated code lurking, line 698 if ( split.path.endsWith("/tmp/") ) { // try to download the file directly. return downloadResourceAsTempFile(url, mon); } throw new FileNotFoundException("File not found: "+url ); } } return tfile; } catch (URISyntaxException ex) { throw new IOException("URI Syntax Exception: " + ex.getMessage()); } } /** * check if the URI can be handled as a URL. For example, * http://autoplot.org/data/foo.dat is true, and * sftp://jbf@klunk.physics.uiowa.edu/home/jbf/data.nobackup/gnc_A_July_06_2012.mat.hdf5 is false. * @param uri * @return true if the URI can be used as a URL. */ private static boolean isUrl( URI uri ) { String s= uri.getScheme(); return s.equals("http") || s.equals("https") || s.equals("ftp") || s.equals("file"); } /** * provide standard logic for identifying the cache location for a file. The file * will not be downloaded, but clients can check to see if such resource has already been loaded. * There is one case where this might be slow, and that's when a zip file must be downloaded to get * the location. * @param suri the uri like http://autoplot.org/data/autoplot.dat * @return the cache file, like /home/jbf/autoplot_data/fscache/http/autoplot.org/data/autoplot.dat */ public static File getCacheFilename( URI suri ) { URISplit split = URISplit.parse( suri ); if ( split.scheme.equals("http") || split.scheme.equals("https") || split.scheme.equals("ftp") || split.scheme.equals("sftp") ) { try { URI root= new URI( split.file ); // from WebFileSystem. File local = FileSystem.settings().getLocalCacheDir(); logger.log( Level.FINE, "WFS localRoot={0}", local); String s = root.getScheme() + "/" + root.getHost() + "/" + root.getPath(); //TODO: check getPath local = new File(local, s); return local; } catch (URISyntaxException ex) { Logger.getLogger(DataSetURI.class.getName()).log(Level.SEVERE, null, ex); } } return null; } /** * return a file reference for the url. The file must have a nice name * on the server, and cannot be the result of a server query with parameters. (Use * downloadResourceAsTempFile for this). * This is initially to fix the problem * for Windows where new URL( "file://c:/myfile.dat" ).getPath() -> "/myfile.dat". * @param suri the URL. * @param allowHtml skip html test that tests for html content. * @param mon progress monitor * @return a local copy of the file. * @throws java.io.IOException * @see #getFile(java.net.URL, org.das2.util.monitor.ProgressMonitor) * @see FileSystemUtil#doDownload(java.lang.String, org.das2.util.monitor.ProgressMonitor) */ public static File getFile( String suri, boolean allowHtml, ProgressMonitor mon) throws IOException { URISplit split = URISplit.parse( suri ); URL url= isUrl( split.resourceUri ) ? split.resourceUri.toURL() : null; try { FileSystem fs = FileSystem.create(toUri(split.path),mon.getSubtaskMonitor("create filesystem")); // mon because of ZipFileSystem String filename = split.file.substring(split.path.length()); FileObject fo = fs.getFileObject(filename); File tfile; if ( fo.exists() ) { tfile = fo.getFile(mon); //TODO: there's a bug here: where we rename the file after unzipping it, but we don't check to see if the .gz is newer. if ( !allowHtml && tfile.exists() && url!=null ) checkNonHtml( tfile, url ); } else { synchronized ( DataSetURI.class ) { // all this needs review, because often Apache servers will also unpack files. FileObject foz= fs.getFileObject(filename+".gz"); // repeat the .gz logic that FileStorageModel.java has. if ( foz.exists() ) { logger.log(Level.FINE, "getting file from compressed version: {0}", foz); File fz= foz.getFile(mon); if ( !allowHtml && fz.exists() && url!=null ) checkNonHtml( fz, url ); File tfile1= new File( fz.getPath().substring(0, fz.getPath().length() - 3) + ".temp" ); tfile= new File( fz.getPath().substring(0, fz.getPath().length() - 3 ) ); if ( !tfile.exists() ) { // another thread already unpacked it. org.das2.util.filesystem.FileSystemUtil.gunzip( fz, tfile1); if ( tfile.exists() ) { if ( ! tfile.delete() ) { throw new IllegalArgumentException("unable to delete "+tfile ); } } // it shouldn't, but to be safe... if ( ! tfile1.renameTo(tfile) ) { if ( !tfile.exists() ) { throw new IllegalArgumentException("unable to rename "+tfile1 + " to "+ tfile ); } } } else { logger.log(Level.FINE, "another thread appears to have already prepared {0}", tfile); } } else { // there is repeated code lurking. See line 616. if ( split.path.endsWith("/tmp/") && url!=null ) { // try to download the file directly. return downloadResourceAsTempFile( url, mon); } if ( fs instanceof WebFileSystem && ((WebFileSystem)fs).isOffline() ) { throw new FileNotFoundException( "File not found in cache of offline filesystem: "+ split.resourceUri +"\n(Offline because of \""+ ((WebFileSystem)fs).getOfflineMessage() + "\")" ); } else { if ( fo.exists() ) { throw new IOException( "Unknown I/O Exception occurred" ); } else { throw new FileNotFoundException("File not found: "+ split.resourceUri ); } } } } } return tfile; } catch ( URIException ex ) { throw new IOException(ex); } catch ( IllegalArgumentException ex ) { logger.log( Level.SEVERE, ex.getMessage(), ex ); if ( ex.getMessage().startsWith("root does not exist") ) { // kludgy bugfix 3053225: why can't FS throw IOException throw new FileNotFoundException(ex.getMessage()); } else if ( ex.getMessage().startsWith("local root does not exist") ) { throw new FileNotFoundException(ex.getMessage()); } else if ( ex.getMessage().contains("unable to create") ) { IOException ex2= new IOException(ex); throw ex2; } else if ( ex.getMessage().contains("unable to delete") ) { IOException ex2= new IOException(ex); throw ex2; } else if ( ex.getMessage().contains("root must contain user name") ) { IOException ex2= new IOException(ex); throw ex2; } else { IOException ex2= new IOException("Unsupported protocol: "+suri,ex); throw ex2;//TODO: we should probably never hide the original message. } } } /** * retrieve the file specified in the URI, possibly using the VFS library to * download the resource to a local cache. The URI should be a downloadable * file, and not the vap scheme URI. * @param uri resource to download, such as "sftp://user@host/file.dat." * @param mon * @return * @throws IOException * @see FileSystemUtil#doDownload(java.lang.String, org.das2.util.monitor.ProgressMonitor) */ public static File getFile( URI uri, ProgressMonitor mon) throws IOException { String suri= fromUri( uri ); return getFile(suri,false,mon); } /** * This loads the URL to a local temporary file. If the temp file * is already downloaded and less than 10 seconds old, it will be used. * * @see DataSetURI.downloadResourceAsTempFile. * @param url the address to download. * @param mon a progress monitor. * @return a File in the FileSystemCache. The file will have question marks and ampersands removed. * @throws IOException */ public static File downloadResourceAsTempFile( URL url, ProgressMonitor mon ) throws IOException { return downloadResourceAsTempFile( url, -1, mon ); } /** * This was introduced when we needed access to a URL with arguments. This allows us * to download the file in a script and then read the file. It will put the * file in the .../fscache/temp/ directory, and the parameters are encoded in the name. * Note this cannot be used to download HTML content, because checkNonHtml is * called. We may introduce "downloadHtmlResourceAsTempFile" or similar if * it's needed. * * This will almost always download, very little caching is done. We allow subsequent * calls within 10 seconds to use the same file (by default). The timeoutSeconds parameter * can be used to set this to any limit. Files older than one day are deleted. * * This is not deleted if the file is already local. Do not delete this file * yourself, it should be deleted when the process exits. * * @param url the address to download. * @param timeoutSeconds if positive, the number of seconds to allow use of a downloaded resource. If -1, then the default ten seconds is used. 12 hours is the longest allowed interval. * @param mon a progress monitor, or null. * @return a File in the FileSystemCache. The file will have question marks and ampersands removed. * @throws IOException * @see HtmlUtil#getInputStream(java.net.URL) which is not used but serves a similar function. */ public static File downloadResourceAsTempFile( URL url, int timeoutSeconds, ProgressMonitor mon ) throws IOException { if ( timeoutSeconds==-1 ) timeoutSeconds= 10; if ( timeoutSeconds>43200 ) { throw new IllegalArgumentException("timeoutSeconds is greater than 12 hours."); } if ( mon==null ) mon= new NullProgressMonitor(); String userInfo; try { userInfo = KeyChain.getDefault().getUserInfo(url); } catch ( CancelledOperationException ex ) { userInfo= null; } URISplit split = URISplit.parse( url.toString() ); // get the folder to put the file. if ( split.file.startsWith("file:/") ) { if ( split.params!=null && split.params.length()>0 ) { throw new IllegalArgumentException("local file URLs cannot have arguments"); } try { return new File(new URL(split.file).toURI()); } catch (URISyntaxException ex) { throw new IllegalArgumentException(ex); } } File local= FileSystem.settings().getLocalCacheDir(); //FileSystem fs = FileSystem.create( toUri(split.path) ); int is; if ( split.path.contains("@") ) { is= split.path.indexOf("@")+1; } else { is= split.scheme.length()+3; } String id= split.scheme + "/" + split.path.substring(is); // fs.getLocalRoot().toString().substring(FileSystem.settings().getLocalCacheDir().toString().length()); final long tnow= System.currentTimeMillis(); // since we cannot deleteOnExit, we now delete any file in temp that is over a day old. synchronized (DataSetURI.class) { File localCache1= new File( local, "temp" ); FileSystemUtil.deleteFilesInTree( localCache1, new FileSystemUtil.Check() { @Override public boolean check(File f) { return tnow - f.lastModified() > 86400000; } } ); } File localCache= new File( local, "temp" ); localCache= new File( localCache, id ); if ( !localCache.exists() ) { if ( !localCache.mkdirs() ) { throw new IOException("unable to make directory: "+localCache); } } String filename = split.file.substring( split.path.length()); if ( split.params!=null && split.params.length()>0 ) { String safe= split.params; safe= safe.replaceAll("\\+","_"); // 2011 safeName uses "plus" for this safe= safe.replaceAll("-","."); // 2011 safeName uses "_" for this, but it should be different than "+" safe= Ops.safeName(safe); // create a Java identifier from this, that will be safe. filename= filename.replaceAll("@","_")+"@"+safe.replaceAll("@","_"); // mimic wget on Windows } else { filename= filename.replaceAll("@","_")+"@"; } if ( filename.length() > 50 ) { // https://sourceforge.net/support/tracker.php?aid=3509357 String[] ss= filename.split("@",-2); String base= ss[0]; if ( base.length()>50 ) base= base.substring(0,50); String args= ss[1]; if ( args.length()>0 ) args= String.format( "%09x", args.hashCode() ); filename= base + String.format( "%09x", ss[0].hashCode() ) + "@" + args; } filename = new File( localCache, filename ).toString(); Object action; File result= new File( filename ); // final name File tempfile= new File(filename + ".temp"); logger.log( Level.FINEST, "downloadResourceAsTempFile:\n sURL: {0}\n file: {1}", new Object[]{url, tempfile}); synchronized (DataSetURI.class) { // clean up after old dead processes that left file behind. I had hudson test that would block test035 from working. //TODO: this is a quick-n-dirty fix. We should be able to check if there's another thread or process loading, or //look for movement in the file... Note no process can keep a file open for more than 3600sec. long tlimit= Math.min( timeoutSeconds, 3600 ) ; if ( tempfile.exists() ) { logger.log(Level.FINEST, "tlimit= {0}", tlimit); logger.log(Level.FINEST, "(tnow-newf.lastModified())/1000 {0}", (tnow - tempfile.lastModified())); if ( ( tnow-tempfile.lastModified() ) / 1000 > tlimit ) { // clean up old files if ( !tempfile.delete() ) { logger.log(Level.FINEST, "old temp file could not be deleted"); } else { logger.log(Level.FINEST, "old temp file was deleted"); } } } if ( result.exists() ) { // clean up old files logger.log(Level.FINEST, "tlimit= {0}", tlimit); logger.log(Level.FINEST, "(tnow-result.lastModified())/1000 = {0}", (tnow - result.lastModified())); if ( ( tnow-result.lastModified() ) / 1000 > tlimit ) { if ( !result.delete() ) { logger.log(Level.FINEST, "old file could not be deleted"); } else { logger.log(Level.FINEST, "old file was deleted"); } } } //TODO: check expires tag and delete after this time. if ( result.exists() && ( tnow-result.lastModified() ) / 1000 < timeoutSeconds && !tempfile.exists() ) { logger.log(Level.FINE, "using young temp file {0}", result); action= ACTION_USE_CACHE; } else if ( tempfile.exists() ) { //TODO: check for old temp file logger.log(Level.FINE, "waiting for other thread to load temp resource {0}", tempfile); action= ACTION_WAIT_EXISTS; } else { File newName= result; while ( newName.exists() ) { String[] ss= filename.split("@",-2); if ( ss.length==2 ) { filename= ss[0] + "@" + ss[1] + "@0"; } else { int i= Integer.parseInt(ss[2]); filename= ss[0] + "@" + ss[1] + "@" + ( i+1 ); } newName= new File( filename ); } if ( !newName.equals(result) ) { // DANGER: I think there may be a bug here where another thread has handed off a file reference, but it has not been opened. if ( !result.renameTo(newName) ) { // move old files out of the way. This is surely going to cause problems on Windows... logger.log(Level.INFO, "unable to move old file out of the way. Using alternate name {0}", newName); result= newName; tempfile= new File( filename + ".temp" ); } } logger.log(Level.FINE, "this thread will downloading temp resource {0}", tempfile); action= ACTION_DOWNLOAD; try (OutputStream out = new FileOutputStream(result) ) { // touch the file out.write( "DataSetURI.downloadResourceAsTempFile: This placeholding temporary file should not be used.\n".getBytes() ); // I bet we see this message again! } // I bet we see this message again! OutputStream outf= new FileOutputStream(tempfile); outf.close(); } } if ( action==ACTION_USE_CACHE ) { logger.log(Level.FINEST,"downloadResourceAsTempFile-> use cache"); return result; } else if (action==ACTION_WAIT_EXISTS ) { long t0= System.currentTimeMillis(); logger.log(Level.FINEST, "downloadResourceAsTempFile-> waitExists"); mon.setProgressMessage("waiting for resource"); mon.started(); try { long l0= tempfile.length(); long tlength= System.currentTimeMillis(); while ( tempfile.exists() ) { try { Thread.sleep(300); if ( System.currentTimeMillis()-t0 > 60000 ) { logger.log(Level.FINE, "waiting for other process to finish loading %s...{0}", tempfile); } if ( tempfile.length()!=l0 ) { l0= tempfile.length(); tlength= System.currentTimeMillis(); } else { if ( System.currentTimeMillis()-tlength > 3 * FileSystem.settings().getConnectTimeoutMs() ) { logger.log(Level.WARNING, "timeout waiting for lengthening of file {0} which another thread is loading", tempfile); throw new IOException("timeout waiting for lengthening of file "+tempfile+" which another thread is loading"); } } if ( mon.isCancelled() ) { throw new InterruptedIOException("cancel pressed"); } } catch (InterruptedException ex) { logger.log(Level.SEVERE, ex.getMessage(), ex); } } } finally { mon.finished(); } return result; } else { boolean fail= true; InputStream in=null; try { Logger loggerUrl= org.das2.util.LoggerManager.getLogger( "das2.url" ); mon.setProgressMessage("downloading "+url); mon.started(); logger.log(Level.FINEST,"downloadResourceAsTempFile-> transfer"); logger.log(Level.FINE, "reading URL {0}", url); loggerUrl.log(Level.FINE,"GET to get data {0}", url); URLConnection urlc= url.openConnection(); urlc.setRequestProperty("Accept-Encoding", "gzip"); // RFE urlc.setConnectTimeout( FileSystem.settings().getConnectTimeoutMs() ); // Reiner describes hang at LANL urlc.setReadTimeout( FileSystem.settings().getReadTimeoutMs() ); urlc.setAllowUserInteraction(false); if ( userInfo != null) { String encode = Base64.getEncoder().encodeToString( userInfo.getBytes()); urlc.setRequestProperty("Authorization", "Basic " + encode); } urlc= HttpUtil.checkRedirect(urlc); in= urlc.getInputStream(); Map> headers = urlc.getHeaderFields(); List contentEncodings=headers.get("Content-Encoding"); boolean hasGzipHeader=false; if (contentEncodings!=null) { for (String header:contentEncodings) { if (header.equalsIgnoreCase("gzip")) { hasGzipHeader=true; break; } } } long contentLength= -1; List contentLengths= headers.get("Content-Length"); if ( contentLengths!=null && contentLengths.size()>0 ) { contentLength= Long.parseLong( contentLengths.get(0) ); } if ( hasGzipHeader ) { logger.fine("temp file is compressed"); in= new GZIPInputStream(in); } else { logger.fine("temp file is not compressed"); } ProgressMonitor loadMonitor= mon.getSubtaskMonitor("loading"); if ( contentLength>-1 ) loadMonitor.setTaskSize(contentLength); in= new DasProgressMonitorInputStream( in, loadMonitor ); if ( urlc instanceof HttpURLConnection ) { final HttpURLConnection hurlc= (HttpURLConnection) urlc; ((DasProgressMonitorInputStream)in).addRunWhenClosedRunnable( new Runnable() { @Override public void run() { hurlc.disconnect(); } }); } if ( contentLength>-1 ) { ((DasProgressMonitorInputStream)in).setStreamLength(contentLength); } OutputStream out= new FileOutputStream( tempfile ); DataSourceUtil.transfer( Channels.newChannel(in), Channels.newChannel(out) ); fail= false; logger.log(Level.FINE,"downloadResourceAsTempFile-> transfer was successful"); } catch ( IOException ex ) { throw new IOException(url.toString(),ex); } finally { if ( fail ) { // clean up if there was an exception if ( !tempfile.delete() ) { logger.log(Level.WARNING, "failed to delete after exception: {0}", tempfile); } if ( !result.delete() ) { logger.log(Level.WARNING, "failed to delete after exception: {0}", result); } } mon.finished(); if ( in!=null ) in.close(); // This will throw the InterruptedIOException } } //checkNonHtml( tempfile, url ); // until 9/22/2011 we didn't check this... synchronized ( DataSetURI.class ) { if ( ! result.delete() ) { // on Windows, rename cannot clobber throw new IllegalArgumentException("unable to delete "+result + " to make way for "+ tempfile ); } if ( ! tempfile.renameTo( result ) ) { throw new IllegalArgumentException("unable to rename "+tempfile + " to "+ result ); } } return result; } /** * retrieve the file specified in the URI, possibly using the VFS library to * download the resource to a local cache. The URI should be a downloadable * file, and not the vap scheme URI. * @param uri resource to download, such as "sftp://user@host/file.dat." * @param mon * @return * @throws IOException */ public static File getFile( String uri, ProgressMonitor mon) throws IOException { return getFile( uri, false, mon ); } /** * get the file, allowing it to have "<html>" in the front. Normally this is not * allowed because of http://sourceforge.net/tracker/?func=detail&aid=3379717&group_id=199733&atid=970682 * @param url * @param mon * @return * @throws IOException */ public static File getHtmlFile( URL url, ProgressMonitor mon ) throws IOException { return getFile( url.toString(), true, mon ); } /** * get a URI from the string which is believed to be valid. This was introduced * because a number of codes called getURI without checking for null, which could be * returned when the URI could not be parsed ("This is not a uri"). Codes that * didn't check would produce a null pointer exception, and now they will produce * a more accurate error. * @param surl * @return * @throws URISyntaxException */ public static URI getURIValid( String surl ) throws URISyntaxException { URI result= getURI( surl ); if ( result==null ) { throw new IllegalArgumentException("URI cannot be formed from \""+surl+"\""); } else { return result; } } /** * canonical method for getting the Autoplot URI. If no protocol is specified, then file:// is * used. Note URIs may contain prefix like vap+bin:http://www.cdf.org/data.cdf. The * result will start with an Autoplot scheme like "vap:" or "vap+cdf:" * * Note 20111117: "vap+cdaweb:" -> URI( "vap+cdaweb:file:///" that's why this works to toUri doesn't. * @param surl the string from the user that should be representable as a URI. * @return the URI or null if it's clearly not a URI. * @throws java.net.URISyntaxException * */ public static URI getURI(String surl) throws URISyntaxException { URISplit split = URISplit.maybeAddFile(surl,0); if ( split==null ) return null; surl = split.surl; if (surl.endsWith("://")) { surl += "/"; // what strange case is this? } //boolean isAlreadyEscaped = split.surl.contains("%25") || split.surl.contains("%20") || split.surl.contains("+"); // TODO: cheesy //if (!isAlreadyEscaped) { surl = surl.replaceAll("%([^0-9])", "%25$1"); surl = surl.replaceAll("<", "%3C"); surl = surl.replaceAll(">", "%3E"); surl = surl.replaceAll(" ", "%20"); // drop the spaces are pluses in filenames. surl = surl.replaceAll("\\^", "%5E"); surl = surl.replaceAll("\\\\", "%5C"); surl = surl.replaceAll("\\|", "%7C"); //} if (split.vapScheme != null) { if ( split.vapScheme.contains(" ") ) { split.vapScheme= split.vapScheme.replace(" ","+"); } surl = split.vapScheme + ":" + surl; } surl = URISplit.format(URISplit.parse(surl)); // add "vap:" if it's not there if ( !( surl.startsWith("vap") ) ) { URISplit split2= URISplit.parse(surl); String vapScheme= URISplit.implicitVapScheme(split2); if ( vapScheme.contains("&") ) { throw new IllegalArgumentException("Address contains ampersand in what looks like a filename: "+surl); } if ( vapScheme.equals("") ) { //one last hangout for generic "vap" type. Better hope they don't use it... Call it vap+X so it's easier to debug. vapScheme="vap+X"; } surl= vapScheme + ":" + surl; } URI result = new URI(surl); //bug 3055130 okay return result; } /** * canonical method for getting the URL. These will always be web-downloadable * URLs. * @param surl the string from the user that should be representable as a URI. * @return null or the URL if available. * @throws java.net.MalformedURLException */ public static URL getURL(String surl) throws MalformedURLException { try { URI uri = getURIValid(surl); return getWebURL(uri); } catch (URISyntaxException ex) { throw new MalformedURLException(ex.getMessage()); } } /** * Represents a single suggestion in a completion list. For example, * http://autoplot.org/data/versioning/data_<C> would result in completions for each file in the folder, * and the aggregation, and each is represented by one completion result. The completion result contains * the human-readable label and a documentation string, as well as the replacement text. */ public static class CompletionResult { /** * the suggestion to replace the completeable string. */ public String completion; /** * human-readable documentation. */ public String doc; /** * human-readable label. */ public String label; /** * the string that will be replaced, or null. */ public String completable; /** * If true then the service-provider code believes this completion should be usable after the completion. */ public boolean maybePlot; protected CompletionResult(String completion, String doc) { this(completion, null, doc, null, false); } protected CompletionResult(String completion, String doc, String completable, boolean maybePlot) { this(completion, null, doc, completable, maybePlot ); } /** * @param completion * @param label the presentation string * @param doc a descriptive string, for example used in a tooltip * @param completable the string that is being completed. (Not really used, but see DataSetUrlCompletionItem) * @param maybePlot true indicates accepting the completion should result in a valid URI. * @see org.das2.jythoncompletion.DataSetUrlCompletionItem */ protected CompletionResult(String completion, String label, String doc, String completable, boolean maybePlot) { this.completion = completion; this.completable = completable; this.label = label != null ? label : ( completion!= null ? completion : completable ); this.doc = doc; this.maybePlot = maybePlot; } public static final CompletionResult SEPARATOR= new CompletionResult("====", "Used to request a separator"); } /** * this is never used in the application code. It must be left over from an earlier system. * This is used in Test005, some scripts, and IDL codes, so don't delete it! * @param surl * @param carotpos * @param mon * @return * @throws Exception */ public static List getCompletions(final String surl, final int carotpos, ProgressMonitor mon) throws Exception { if ( carotpos==0 || ( !surl.substring(0,carotpos).contains(":") && ( carotpos<4 && surl.substring(0, carotpos).equals( "vap".substring(0,carotpos ) ) || ( surl.length()>3 && surl.substring(0, 3).equals( "vap" ) ) ) ) ) { return getTypesCompletions( surl, carotpos, mon ); } URISplit split = URISplit.parse(surl, carotpos, true); if ( ( split.vapScheme!=null && split.file == null ) || ( split.file!=null && split.resourceUriCarotPos > split.file.length()) && DataSourceRegistry.getInstance().hasSourceByExt(DataSetURI.getExt(surl))) { return getFactoryCompletions(URISplit.format(split), split.formatCarotPos, mon); } else { if ( split.vapScheme==null && split.scheme==null ) { String[] types= new String[] { "ftp://", "http://", "https://", "file:/", "sftp://" }; List result= new ArrayList<>(); String completable= surl.substring(0, carotpos); for (String type : types) { if (type.length() >= carotpos && type.startsWith(completable)) { result.add(new CompletionResult(type, null, type, completable, false)); } } return result; } else { int firstSlashAfterHost = split.authority == null ? 0 : split.authority.length(); if (split.resourceUriCarotPos <= firstSlashAfterHost) { return getHostCompletions(URISplit.format(split), split.formatCarotPos, mon); } else { return getFileSystemCompletions(URISplit.format(split), split.formatCarotPos, true, true, null, mon); } } } } /** * get completions for hosts, by looking in the user's cache area. * @param surl * @param carotpos * @param mon * @return possibly immutable list. * @throws IOException */ public static List getHostCompletions(final String surl, final int carotpos, ProgressMonitor mon) throws IOException { URISplit split = URISplit.parse(surl.substring(0, carotpos)); String prefix; String surlDir; if (split.path == null) { prefix = ""; surlDir = ""; } else { prefix = split.file.substring(split.path.length()); surlDir = split.path; } mon.setLabel("getting list of cache hosts"); String[] s; if ( split.scheme==null ) { List completions = new ArrayList<>(); s= new String[] { "ftp://", "http://", "https://", "file:///", "sftp://", }; for (String item : s) { completions.add(new DataSetURI.CompletionResult(item + surl + "/", item + surl + "/")); } return completions; } File cacheF = new File(FileSystem.settings().getLocalCacheDir(), split.scheme); if (!cacheF.exists()) return Collections.emptyList(); s = cacheF.list(); if ( s==null ) return Collections.emptyList(); boolean foldCase = true; if (foldCase) { prefix = prefix.toLowerCase(); } List completions = new ArrayList<>(s.length); for (String item : s) { String scomp = foldCase ? item.toLowerCase() : item; if (scomp.startsWith(prefix)) { StringBuilder result1 = new StringBuilder(item); result1.append( "/" ); // drill down single entries, since often the root doesn't provide a list. String[] s2 = new File(cacheF, result1.toString()).list(); if ( s2==null ) { // does not exist s2= new String[0]; } while (s2.length == 1 && new File(cacheF, result1 + "/" + s2[0]).isDirectory()) { result1.append( s2[0] ).append( "/" ); s2 = new File(cacheF, result1.toString() ).list(); if ( s2==null ) { // does not exist s2= new String[0]; } } completions.add(new DataSetURI.CompletionResult(surlDir + result1.toString(), result1.toString(), null, surl.substring(0, carotpos), true)); } } // check for single completion that is just a folder name with /. if (completions.size() == 1) { if ((completions.get(0).completion).equals(surlDir + prefix + "/")) { // maybe we should do something special. } } Collections.sort(completions,new Comparator() { @Override public int compare(CompletionResult o1, CompletionResult o2) { //String[] s1= o1.completion.split("\\."); //String[] s2= o2.completion.split("\\."); //TODO: reverse so these are edu.uiowa.physics, then sort... return o1.completion.compareTo(o2.completion); } }); return completions; } /** * get completions within the filesystem that appear to form aggregations. Note this does not appear * to be used, having no Java references. * @param surl * @param carotpos * @param mon * @return * @throws IOException * @throws URISyntaxException */ public static List getFileSystemAggCompletions(final String surl, final int carotpos, ProgressMonitor mon) throws IOException, URISyntaxException { URISplit split = URISplit.parse(surl.substring(0, carotpos),carotpos,false); String surlDir = URISplit.uriDecode(split.path); mon.setLabel("getting remote listing"); FileSystem fs; String[] s; fs = FileSystem.create( DataSetURI.toUri(surlDir) ); s = fs.listDirectory("/"); Arrays.sort(s); List completions = new ArrayList<>(5); String[] s2= new String[s.length]; for ( int i=0; i0 ) { //String sagg= DataSourceUtil.makeAggregation( s2[0], s2 ); List files= new LinkedList( Arrays.asList(s2) ); List saggs= DataSourceUtil.findAggregations( files, true ); for ( String sagg: saggs ) { sagg= URISplit.removeParam( sagg, "timerange" ); completions.add( new DataSetURI.CompletionResult( sagg, "Use aggregation", "", true ) ); } } return completions; } /** * get completions within the filesystem for the directory listing of files. * @param surl * @param carotpos * @param inclAgg include aggregations it sees. These are a guess. * @param inclFiles include files as well as aggregations. * @param acceptPattern if non-null, files and aggregations much match this. * @param mon * @return * @throws IOException * @throws URISyntaxException */ public static List getFileSystemCompletions(final String surl, final int carotpos, boolean inclAgg, boolean inclFiles, String acceptPattern, ProgressMonitor mon) throws IOException, URISyntaxException { return getFileSystemCompletions( surl, carotpos, inclAgg, inclFiles ? null : new ArrayList(), acceptPattern, mon ); } /** * gets completions based on cached folders. This supports use when the * filesystem is offline, when parents are not web filesystems, and presents * used folders separately. * @param surl http://sarahandjeremy.net/ * @param carotpos the position of the carot. Presently everything after the carot is ignored. * @param inclAgg include aggregations it sees. These are a guess. * @param inclFiles include files as well as aggregations. * @param acceptPattern if non-null, files and aggregations much match this. * @param mon * @return possibly immutable list. * @throws IOException * @throws URISyntaxException */ public static List getFileSystemCacheCompletions(final String surl, final int carotpos, boolean inclAgg, boolean inclFiles, String acceptPattern, ProgressMonitor mon) throws IOException, URISyntaxException { URISplit split = URISplit.parse(surl.substring(0, carotpos)); String prefix; String surlDir; if (split.path == null) { prefix = ""; surlDir = ""; } else { prefix = split.file.substring(split.path.length()); surlDir = split.path; } mon.setLabel("getting list of cached folders"); String[] s; if ( split.scheme==null ) { throw new IllegalArgumentException("need scheme and hostname"); } File cacheF = new File(FileSystem.settings().getLocalCacheDir(), split.scheme + '/' + split.path.substring(split.scheme.length()+3) ); if (!cacheF.exists()) return Collections.emptyList(); s = cacheF.list(); // s could be a zip file. Guard against null. There should be a local listing method in FileSystem. if ( s==null ) return Collections.emptyList(); boolean foldCase = true; if (foldCase) { prefix = prefix.toLowerCase(); } List completions = new ArrayList<>(s.length); for (String item : s) { String scomp = foldCase ? item.toLowerCase() : item; if (scomp.startsWith(prefix) && !scomp.endsWith(".listing")) { File ff = new File(cacheF, item); if ( ! ff.isDirectory() ) continue; String[] ss= ff.list(); if ( ss==null || ss.length==0 ) continue; StringBuilder result1 = new StringBuilder(item); result1.append( "/" ); // drill down single entries, since often the root doesn't provide a list. String[] s2 = new File(cacheF, result1.toString()).list(); if ( s2==null ) { // does not exist s2= new String[0]; } while (s2.length == 1 && new File(cacheF, result1 + "/" + s2[0]).isDirectory()) { result1.append(s2[0]).append("/"); s2 = new File(cacheF, result1.toString()).list(); if ( s2==null ) { // does not exist s2= new String[0]; } } completions.add(new DataSetURI.CompletionResult(surlDir + result1.toString(), result1.toString(), null, surl.substring(0, carotpos), true)); } } // check for single completion that is just a folder name with /. if (completions.size() == 1) { if ((completions.get(0).completion).equals(surlDir + prefix + "/")) { // maybe we should do something special. } } return completions; } /** * Get the completions from the FileSystem, including aggregation suggestions. * @param surl the URI, e.g. file:/home/jbf/pngwalk/product_$Y$m.png * @param carotpos the position of the editor carot (cursor) where the completions * @param inclAgg include aggregations it sees. These are a guess. * @param inclFiles if null, list files, but is non-null, then only include files in the list of regex. * @param acceptPattern if non-null, files and aggregations much match this. * @param mon * @return list of results. * @throws IOException * @throws URISyntaxException */ public static List getFileSystemCompletions(final String surl, final int carotpos, boolean inclAgg, List inclFiles, String acceptPattern, ProgressMonitor mon) throws IOException, URISyntaxException { URISplit split = URISplit.parse(surl.substring(0, carotpos),carotpos,false); String prefix = URISplit.uriDecode(split.file.substring(split.path.length())); String surlDir = URISplit.uriDecode(split.path); mon.setLabel("getting remote listing"); FileSystem fs; String[] s; if ( surlDir.equals("file:" ) || surlDir.equals("file://" ) ) { //TODO: could go ahead and list CompletionResult t0; if ( split.vapScheme!=null ) { t0= new CompletionResult( split.vapScheme + ":" + "file:///","need three slashes"); } else { t0= new CompletionResult("file:///","need three slashes"); } List completions= Collections.singletonList(t0); return completions; } boolean onlyAgg= false; String prefixPrefix= ""; if ( surlDir.contains("$Y") ) { // $Y must be first for now. This will be generalized after verified DatumRange timeRange= null; try { URISplit split1= URISplit.parse(surl); Map parms= URISplit.parseParams(split1.params); String stimeRange= parms.get(URISplit.PARAM_TIME_RANGE); if ( stimeRange!=null ) { timeRange= DatumRangeUtil.parseTimeRange(stimeRange); } } catch ( ParseException e ) { logger.log(Level.WARNING, "parse exception: {0}", e); } int ip= surlDir.indexOf("$Y"); String s1= surlDir.substring(0,ip); String s2= surlDir.substring(ip,surlDir.length()-1); FileSystem fsp= FileSystem.create( DataSetURI.toUri(s1), mon ); FileStorageModel fsm= FileStorageModel.create( fsp, s2 ); fs= fsp; // careful, we only use this for case insensitive check List ss= new ArrayList(); String [] ss2= fsm.getNamesFor(timeRange); int nn= Math.min( 2, ss2.length ); for ( int i=0; i