package org.autoplot.datasource;

import java.net.URI;
import java.net.URISyntaxException;
import java.text.ParseException;
import java.util.ArrayList;
import java.util.Collections;
import java.util.LinkedHashMap;
import java.util.List;
import java.util.Map;
import java.util.Map.Entry;
import java.util.logging.Level;
import java.util.logging.Logger;
import java.util.regex.Matcher;
import java.util.regex.Pattern;
import org.das2.datum.DatumRange;
import org.das2.datum.DatumRangeUtil;
import org.das2.util.LoggerManager;

/**
 * Class for containing the elemental parts of a URI, and utility
 * routines for working with URIs.
 *
 * We need a working definition of well-formed and colloquial URIs:
 *<blockquote><pre>
 * = well-formed URIs =
 *   <vapScheme>:<fileResource>?<params>
 *   <vapScheme>:[<identifier>?]<params>
 *   <vapScheme>:<params>
 *   * they are valid URIs: they contain no spaces, etc.
 * == params ==
 *   ampersand-delimited (&) list of name=value pairs, or just value.
 *   vap+cdaweb:ds=ac_k0_epm&H_lo&timerange=2010-01
 * = colloquial URIs =
 *   * these are Strings that can be converted into URIs.
 *   * spaces in file names are converted into %20. 
 *   * spaces in parameter lists are converted into pluses.
 *   * pluses in parameter lists are converted into %2B.
 *   * note that if there are pluses but the URI is valid, then pluses may be left alone. 
 * </pre></blockquote>
 * 
 * This routine knows nothing about the data source that will interpret the
 * URI, so this needs to be established.
 * 
 * @author jbf
 */
public class URISplit {

    private static final Logger logger= LoggerManager.getLogger( LogNames.APDSS_URI );
    
    /**
     * The following are suggestions for parameter names to encourage consistency between implementations.
     * See http://autoplot.org/developer.URI_syntax
     * For example, if your URI accepts a time range like "Nov 2011", then use the timerange=Nov+2011 in your URI. If you
     * do, then for example DefaultTimeSeriesBrowseEditor will work for you.
     */

    /**
     * time range subset.
     */
    public static final String PARAM_TIME_RANGE= "timerange";

    public static final String PARAM_TIME_RESOLUTION="resolution";

    /**
     * subset of rank 2 data.  For example, columns of excel workbook or ascii table.
     * rank2=[3,5] or rank2=Bx-Bz
     */
    public static final String PARAM_RANK2="rank2";

    /**
     * used for the number of records to read.
     */
    public static final String PARAM_REC_COUNT="recCount";

    /**
     * first positional parameter, typically interpreted the same as PARAM_ID
     */
    public static final String PARAM_ARG_0="arg_0";

    /**
     * typically the dataset id.
     */
    public static final String PARAM_ID="id";

    /**
     * some datasources support periodic checks to see if data sources have updated, such as:
     *   AggregatingDataSource
     *   AbstractDataSources (most of those based on files)
     */
    public static final String PARAM_FILE_POLL_UPDATES= "filePollUpdates";

    /**
     * make the URI canonical, with the vap+&lt;ext&gt;: prefix. 
     * This will also now sort the parameters, when this can be done.
     * @param suri, such as "/tmp/x.cdf"
     * @return "vap+cdf:file:///tmp/x.cdf"
     */
    public static String makeCanonical(String suri) {
        logger.log(Level.FINEST, "makeCanonical {0}", suri);
        if ( suri==null ) return null;
        URISplit split= URISplit.parse(suri);
        if ( !DataSourceRegistry.getInstance().hasParamOrder(suri) ) {
            Map<String,String> paramsLoose= URISplit.parseParams(split.params);
            LinkedHashMap<String,String> params= new LinkedHashMap<>();
            String arg_0= paramsLoose.remove("arg_0");
            if ( arg_0!=null ) params.put( "arg_0", arg_0 );
            List<String> keys= new ArrayList<>( paramsLoose.keySet() );
            Collections.sort(keys);
            if ( keys.remove("start_time") ) keys.add("start_time"); //
            if ( keys.remove("end_time" ) ) keys.add("end_time");
            for ( String k: keys ) {
                params.put( k, paramsLoose.get(k) );
            }
            split.params= URISplit.formatParams(params);
            if ( split.params.length()==0 ) {
                split.params=null;
            }
        }
        suri= URISplit.format(split); // make canonical
        if ( !suri.startsWith("vap+") && split.ext!=null && split.ext.length()>1 ) {
            suri= "vap+"+split.ext.substring(1)+":"+suri;
        }
        logger.log(Level.FINEST, "makeCanonical results in {0}", suri);
        return suri;
    }

    /**
     * make the URI colloquial, e.g. removing "vap+cdf:" from "vap+cdf:file:///tmp/x.cdf"
     * URIs that do not have a resource URI are left alone.
     * @param suri a URI 
     * @return the URI, more colloquial and readable.
     */
    public static String makeColloquial(String suri) {
        logger.log(Level.FINEST, "makeColloquial {0}", suri);
        if ( suri==null ) return null;
        suri= suri.trim();
        if ( suri.equals("") ) return "";
        URISplit split= URISplit.parse(suri);
        if ( split.vapScheme!=null ) {
            if ( split.ext!=null && split.ext.length()>1 && split.vapScheme.equals("vap+"+split.ext.substring(1) ) ) {
                split.vapScheme= null;
            }
        }
        String result= URISplit.format(split);
        if ( result.endsWith("file:///") && suri.endsWith(":") ) { // kludge around "file:/// that is added to "vap+cdaweb:"
            logger.log(Level.FINEST, "makeColloquial results in {0}", suri);
            return suri;
        }
        logger.log(Level.FINEST, "makeColloquial results in {0}", result);
        return result;
    }

    /**
     * ensure that the reference, which may be relative, absolute.
     * NOTE this is only implemented for unix filenames. TODO: Windows.
     * For example:<ul>
     * <li>/tmp/,foo.dat -> /home/t/foo.dat
     * <li>/tmp/,/home/jbf/foo.dat -> /home/jbf/foo.dat
     * </ul>
     * @param path the absolute directory.
     * @param suri the URI, which may be relative to path.
     * @return the absolute path
     */
    public static String makeAbsolute( String path, String suri ) {
        int i= suri.indexOf(':');
        if ( i==-1 ) { // it's a file.
            boolean isAbsolute= suri.startsWith("/");
            if ( !isAbsolute ) {
                String pwd= path;
                if ( pwd.endsWith("/.") ) pwd= pwd.substring(0,pwd.length()-2);
                if ( !pwd.endsWith("/")) {
                    pwd= pwd + "/"; //TODO: Windows...
                }
                suri= pwd + suri;
            }
        }
        return suri;
    }

    /**
     * scheme for Autoplot, if provided.  e.g.  vap+cdf.  If not provided,
     * then "vap:" is implicit.
     */
    public String vapScheme;
    /**
     * scheme for resource, e.g. jdbc.mysql
     */
    public String scheme;
    /**
     * the complete, modified surl.   file:///home/jbf/mydata.qds
     * this is the resource name, and doesn't contain the vapScheme.
     */
    public String surl;
    /**
     * the resource that is handled by the DataSource.  This may be null if surl doesn't form a valid uri.
     * 
     */
    public URI resourceUri;
    /**
     * the resource uri up to the authority, e.g.  jbdc:mysql://192.168.0.203:3306
     */
    public String authority;
    /**
     * the resource uri including the path part.
     */
    public String path;
    /**
     * contains the resource string up to the query part.
     */
    public String file;
    public String ext;
    /**
     * contains the parameters part, a ampersand-delimited set of parameters. For example, column=field2&rank2.
     */
    public String params;

    /**
     * additional processes to be applied to the URI.  For example, slice0(0) means slice the dataset at this point.
     */
    public String filters;

    /**
     * position of the caret after modifications to the surl are made.  This
     * is with respect to surl, the URI for the datasource, without the "vap" scheme.
     */
    public int resourceUriCarotPos;
    /**
     * position of the caret after modifications to the surl are made.  This
     * is with respect to formatted URI, which probably includes the explicit "vap:" scheme.
     */
    public int formatCarotPos;

    static List<String> otherSchemes= Collections.emptyList();

    /**
     * allow parsing of script:, bookmarks:, pngwalk:, etc
     * @param otherSchemes
     */
    public static void setOtherSchemes( List<String> otherSchemes ) {
        URISplit.otherSchemes= otherSchemes;
    }

    /**
     * add "file:/" to a resource string that appears to reference the local filesystem.
     * return the parsed string, or null if the string doesn't appear to be from a file.
     * @param surl
     * @param caretPos
     * @return null or the URISplit
     */
    public static URISplit maybeAddFile(String surl, int caretPos) {
        URISplit result = new URISplit();

        if (surl.length() == 0) {
            surl = "file:///";
            caretPos = surl.length();
            result.surl = surl;
            result.vapScheme = null;
            result.resourceUriCarotPos = caretPos;
            result.formatCarotPos = caretPos;
        }

        String scheme;  // identify a scheme, if any.  This might be vap+foo:, or http:
        int i0 = surl.indexOf(':');
        if (i0 == -1) {
            scheme = "";
        } else if (i0 == 1) { // one letter scheme is assumed to be windows drive letter.
            scheme = "";
        } else {
            if ( surl.substring(0,i0).contains("/") ) {
                scheme = "";
            } else {
                scheme = surl.substring(0, i0);
            }
        }

        if ( scheme.startsWith("vap") || otherSchemes.contains(scheme) ) {
            String resourcePart = surl.substring(i0 + 1);
            if ( !scheme.equals("vap") ) { // legacy URIs would often have informationless "vap:" prefix.  We remove this now.
                result.vapScheme = scheme;
            }
            if (scheme.equals("vap+internal")) { // leave the resourcePart alone. TODO: jdbc and other non-file URIs.
                result.surl= resourcePart;
            } else {
                URISplit resourceSplit = maybeAddFile(resourcePart, caretPos - (i0 + 1));  //TODO: jdbc and vap+inline
                if ( resourceSplit==null ) {
                    result.surl= resourcePart;
                    result.file= "";
                    result.formatCarotPos= caretPos;
                } else {
                    result.surl = resourceSplit.surl;
                    result.formatCarotPos = (caretPos > i0) ? resourceSplit.resourceUriCarotPos + (i0 + 1) : caretPos;
                    result.resourceUriCarotPos = result.formatCarotPos - (scheme.length() + 1); // with respect to resource part.
                }
            }

        } else {
            result.surl = surl;
            result.resourceUriCarotPos = caretPos;
        }

        if (scheme.equals("")) {
            boolean isFile= true;
            int iquery= surl.indexOf('?');
            if ( iquery==-1 ) {
                int ieq= surl.indexOf('=');
                //kludge in support for "ripples(30,30")
                int ch0= surl.length()>0 ? surl.charAt(0) : (char)0;
                int ch1= surl.length()>1 ? surl.charAt(1) : (char)0;
                boolean notSlashStart= ch0!='/' && ch0!='\\' && ch1!='/' && ch1!='\\' && ch1!=':';
                if ( notSlashStart || ( ieq>-1 && !(surl.charAt(0)=='/') ) ) {
                    isFile= false;
                }
            }

            if ( !isFile ) {
                return null;

            } else {
                if ( surl.startsWith("~/" ) ) { // finally add this to unix.
                    surl= System.getProperty("user.home") + surl.substring(1);
                    result.resourceUriCarotPos += System.getProperty("user.home").length()-1;
                }
                result.surl = "file://";
                result.scheme= "file";
                result.resourceUriCarotPos += 7;
                if ((surl.charAt(0) == '/')) {
                    result.surl += surl;
                } else {
                    result.surl += ('/' + surl); // Windows c:
                    result.resourceUriCarotPos += 1;
                }
                int iq= result.surl.indexOf('?');
                if ( iq==-1 ) iq= result.surl.length();

                result.surl = result.surl.replaceAll("\\\\", "/"); //TODO: what if \ in query part?

                int spaceCount= charCount( result.surl, ' ', 0, result.surl.length() );
                result.surl = replaceAll( result.surl, " ", "%20", 0, iq );

                result.formatCarotPos+= spaceCount*2; //account for inserted characters.
                result.resourceUriCarotPos+= spaceCount*2;
            }
            
        }

        return result;
    }

    private static int charCount( String src, char find, int start, int end ) {
        int count=0;
        for ( int i=start; i<end; i++ ) {
            if ( src.charAt(i)==find ) count++;
        }
        return count;
    }

    /**
     * replace all characters in the given range.
     * @param src
     * @param regex
     * @param replacement
     * @param start
     * @param end
     * @return
     */
    private static String replaceAll( String src, String regex, String replacement, int start, int end ) {
        String prefix= src.substring(0,start);
        String middle= src.substring(start,end);
        String suffix= src.substring(end);
        return prefix + middle.replaceAll(regex, replacement) + suffix;
    }

    /**
     * added to avoid widespread use of parse(uri.toString).  This way its all being done with same code,
     * and keep the URI abstraction.
     * @param uri
     * @return
     */
    public static URISplit parse( URI uri ) {
        return parse( DataSetURI.fromUri(uri), 0, true );
    }

    /**
     * split the url string into components.  This does not try to identify
     * the vap scheme, since that might require interaction with the server to
     * get mime type.  This inserts the scheme "file://" when the scheme is 
     * absent.
     * The string http://www.example.com/data/myfile.nc?myVariable is split into:<ul>
     *   <li>scheme, http
     *   <li>authority, http://www.example.com
     *   <li>path, the directory with http://www.example.com/data/
     *   <li>file, the file, http://www.example.com/data/myfile.nc
     *   <li>ext, the extenion, .nc
     *   <li>params, myVariable or null
     * </ul>
     * @param suri the uri to be parsed
     * @return the components.
     */
    public static URISplit parse(String suri) {
        return parse(suri, 0, true);
    }

    /**
     * return the vap scheme in split.vapScheme or the one inferred by the 
     * extension.  Returns an empty string (not "vap") if one cannot be inferred.
     * e.g:
     *    /home/jbf/myfile.jyds --> vap+jyds
     *    vap+txt:/home/jbf/myfile.csv --> vap+txt
     * This was introduced as part of the effort to get rid of extraneous "vap:"s
     * that would be added to URIs.
     *
     * @param split
     * @return the vap scheme or empty string.
     */
    public static String implicitVapScheme( URISplit split ) {
        if ( split.vapScheme!=null ) return split.vapScheme;
        if ( split.ext!=null && split.ext.length()>1 ) return "vap+"+split.ext.substring(1);
        return "";
    }

    /**
     * convenient method to remove a parameter (or parameters) from the list of parameters
     * @param surl any URI or web address
     * @param parm the name to remove
     * @return the URI with the parameter removed, and the question mark removed when no parameters remain.
     */
    public static String removeParam( String surl, String ... parm ) {
        URISplit split= URISplit.parse(surl);
        Map <String,String> params= URISplit.parseParams( split.params );
        for ( String p: parm ) {
            params.remove(p);
        }
        split.params= URISplit.formatParams(params);
        if ( params.isEmpty() ) split.params=null;
        if ( split.vapScheme!=null && !surl.startsWith(split.vapScheme) ) split.vapScheme=null;
        return URISplit.format(split);
    }


    /**
     * convenient method for adding or replacing a parameter to the URI.
     * @param surl any URI or web address
     * @param name the parameter name to add
     * @param value the parameter value to add
     * @return the uri with the question mark and parameter added.
     */
    public static String putParam( String surl, String name, String value ) {
        URISplit split= URISplit.parse(surl);
        Map <String,String> params= URISplit.parseParams( split.params );
        params.put( name, value );
        split.params= URISplit.formatParams(params);
        if ( split.vapScheme!=null && !surl.startsWith(split.vapScheme) ) split.vapScheme=null;
        return format(split);
    }

    /**
     * convenient method for getting a parameter in the URI.  
     * @param surl
     * @param name parameter name.
     * @param deft default value if the parameter is not found.
     * @return
     */
    public static String getParam( String surl, String name, String deft ) {
        URISplit split= URISplit.parse(surl);
        Map <String,String> params= URISplit.parseParams( split.params );
        String val= params.get( name );
        if ( val==null ) val= deft;
        return val;
    }
    
    /**
     * returns group 1 if there was a match, null otherwise.
     * @param s
     * @param regex
     * @return
     */
    private static String magikPop(String s, String regex) {
        Pattern p = Pattern.compile(regex);
        Matcher m = p.matcher(s);
        if (m.matches()) {
            return m.group(1);
        } else {
            return null;
        }
    }
    
    /**
     * only split on the delimiter when we are not within the exclude delimiters.  For example,
     * <code>
     * x=getDataSet("http://autoplot.org/data/autoplot.cdf?Magnitude&noDep=T")&y=getDataSet('http://autoplot.org/data/autoplot.cdf?BGSEc&slice1=2')&sqrt(x)
     * </code>
     * @param s the string to split.
     * @param delim the delimiter to split on, for example the ampersand (&).
     * @param exclude1 for example the single quote (')
     * @param exclude2 for example the double quote (")  Note URIs don't support these anyway.
     * @return the split.
     */
    public static String[] guardedSplit( String s, char delim, char exclude1, char exclude2 ) {    
        if ( delim=='_') throw new IllegalArgumentException("_ not allowed for delim");
        StringBuilder scopyb= new StringBuilder(s.length());
        char inExclude= (char)0;
        
        for ( int i=0; i<s.length(); i++ ) {
            char c= s.charAt(i);
            if ( inExclude==0 ) {
                if ( c==exclude1 || c==exclude2 ) inExclude= c;
            } else {
                if ( c==inExclude ) inExclude= 0;
            }
            if ( inExclude>(char)0 ) c='_';
            scopyb.append(c);            
        }
        String[] ss= scopyb.toString().split(""+delim);
        
        int i1= 0;
        for ( int i=0; i<ss.length; i++ ) {
            int i2= i1+ss[i].length();
            ss[i]= s.substring(i1,i2);
            i1= i2+1;
        } 
        return ss;
    }
    

    /**
     * interpret the scheme part to vapScheme and scheme.  If the resource URI is
     * valid, then this will be set as well.  result.surl may be modified.
     * @param result
     */
    private static void parseScheme(URISplit result, boolean normalize) throws URISyntaxException {
        String surl = result.surl;
        int h = surl.indexOf(':'); // "c:" should be "file:///c:"

        String scheme = surl.substring(0, h);
        

        if (scheme.startsWith("vap")) {
            result.vapScheme = scheme;
            result.formatCarotPos = result.resourceUriCarotPos + scheme.length() + 1;
            result.surl = surl.substring(h + 1);
            result.scheme = magikPop(result.surl, "([a-zA-Z\\+]+)\\:.*");
            int iq= result.surl.indexOf('?');
            if ( iq==-1 ) iq= result.surl.length();
            try {
                result.resourceUri = new URI(uriEncode(result.surl.substring(0,iq)));
                result.scheme = result.resourceUri.getScheme();
            } catch (URISyntaxException ex) {
                // do nothing, this field may be null.
            }
        } else {
            if ( scheme.length()<20 && scheme.matches("[a-z.]+") && scheme.contains(".")) { // handle legacy schemes like cdf.http://...
                logger.log( Level.WARNING, "URI scheme contains .: {0} converting from vap.xxx to vap+xxx", surl);
                int j = scheme.indexOf('.');
                result.vapScheme = "vap+" + scheme.substring(0, j);
                result.surl = result.surl.substring(j + 1);
                if (result.resourceUriCarotPos > j) result.resourceUriCarotPos -= (j + 1);
                result.formatCarotPos = result.resourceUriCarotPos + result.vapScheme.length() + 1;
                result.scheme = magikPop(result.surl, "([a-zA-Z\\+]+)\\:.*");
                int iq= result.surl.indexOf('?');
                if ( iq==-1 ) iq= result.surl.length();
                try {
                    result.resourceUri = new URI(uriEncode(result.surl.substring(0,iq)));
                    result.scheme = result.resourceUri.getScheme();
                } catch (URISyntaxException ex) {
                    // do nothing, this field may be null.
                }
            } else {
                if (result.vapScheme == null && normalize ) {
                    result.formatCarotPos = result.resourceUriCarotPos;
                }
                result.surl = surl;
                result.scheme = magikPop(result.surl, "([a-zA-Z\\+]+)\\:.*");
                int iq= result.surl.indexOf('?');
                if ( iq==-1 ) iq= surl.length();
                try {
                    result.resourceUri = new URI(uriEncode(result.surl.substring(0,iq)));
                    result.scheme = result.resourceUri.getScheme();
                } catch (URISyntaxException ex) {
                    // do nothing, this field may be null.
                }
            }
        }
    }

    /**
     * split the UI string into components, keeping track of the caret position
     * when characters are inserted.  This does not try to identify
     * the vap scheme, since that might require interaction with the server to
     * get mime type.  This inserts the scheme "file://" when the scheme is 
     * absent.
     * For example, the string http://www.example.com/data/myfile.nc?myVariable is split into:<ul>
     *   <li>vapScheme, vap+nc
     *   <li>scheme, http
     *   <li>authority, http://www.example.com
     *   <li>path, the directory with http://www.example.com/data/
     *   <li>file, the file, http://www.example.com/data/myfile.nc
     *   <li>ext, the extension, .nc or null.
     *   <li>params, myVariable or null.
     *   <li>filters, the fragment of the URI following hash character.
     * </ul>
     * @param surl  the string to parse
     * @param caretPos the position of the caret, the relative position will be preserved through normalization in formatCaretPos
     * @param normalize normalize the surl by adding implicit "vap", etc.
     * @throws IllegalArgumentException 
     * @return the decomposed uri.
     */
    public static URISplit parse( String surl, int caretPos, boolean normalize) {

        logger.log( Level.FINE, "URISplit.parse(\"{0}\",{1},{2})", new Object[]{ surl, caretPos, normalize });

        if ( surl.startsWith("file:/") && surl.endsWith(":") && surl.length()<11 && surl.charAt(surl.length()-3)=='/' ) { // kludge for file:///c:<CARET> on Windows.
            if ( caretPos==surl.length() ) caretPos++;
            surl= surl+"/";
        }

//        // finally, kludge for Unix ~.  TODO: Get this working some time...
//        if ( surl.startsWith("~") ) {
//            surl= System.getProperty("user.home") + surl.substring(1);
//            caretPos += ( System.getProperty("user.home").length() -1 );
//        }

        if ( surl.startsWith("http://autoplot.org/autoplot.jnlp?") ) {
            String[] popFront= new String[] { "http://autoplot.org/autoplot.jnlp?version=devel&", "http://autoplot.org/autoplot.jnlp?"  };
            for ( String s: popFront ) {
                if ( surl.startsWith(s) ) {
                    surl= surl.substring(s.length());
                    caretPos= ( caretPos<s.length() ? 0 : caretPos-s.length() );
                }
            }         
        }

        URISplit result = maybeAddFile(surl, caretPos);

        if ( result==null ) {
            result= new URISplit();
            result.surl= surl;
            result.vapScheme= null;
            result.formatCarotPos= caretPos;
            return result;
        }
        if ( result.vapScheme!=null && "vap+internal".equals(result.vapScheme) ) result.file=""; // non-files will get "" for the file, and this should too.
        try {
            if ( result.vapScheme==null || result.file==null ) {
                parseScheme(result, normalize);
            }
        } catch (URISyntaxException ex) {
            result.surl = uriEncode(result.surl); //TODO: move resourceUriCarotPos
            try {
                parseScheme(result, normalize);
            } catch (URISyntaxException ex1) {
                throw new RuntimeException(ex1);
            }
        }

        int i;

        String rsurl = result.surl;

        int iquery;
        // check for just one ?
        iquery = rsurl.indexOf('?');
        int ieq= rsurl.indexOf('=');

        String file;

        file = result.resourceUri == null ? null : result.resourceUri.getPath();
        if (file == null) {
            if (iquery == -1) {
                if ( ieq==-1 ) {
                    file= rsurl;
                } else {
                    file= null;
                }
            } else {
                file = rsurl.substring(0, iquery);
            }
        }

        String ext = null;
        if (file != null) {
            i = file.lastIndexOf('.');
            if (i>-1 ) {
                int i2= file.indexOf('/',i);
                if ( i2==-1 ) {
                    ext= file.substring(i);
                } else {
                    ext= "";
                }
            } else {
                ext =  "";
            }
            if ( ext.length()>0 && ext.contains("&") ) { // catch errors when & is used instead of ?: .../WAV_2011227_SRV_V17.PKT&type=B
                throw new IllegalArgumentException("The extension of a filename cannot contain ampersand (&): "+ surl);
            }
        }

        String params = null;
        int fileEnd=-1;

            //int ipipe= file.indexOf("|");
            //if ( ipipe>-1 ) {
            //    result.process= file.substring(ipipe);
            //    file= file.substring(0,ipipe);
            //} else {
            //    result.process= "";
            //}
        
        int ihash;
        if ( iquery==-1 ) {
            ihash= rsurl.indexOf('#');
        } else {
            ihash= rsurl.indexOf('#',iquery);
        }
        if ( ihash==-1 ) ihash= rsurl.length();
        
        if (file != null && iquery != -1) {
            fileEnd = iquery;
            params = rsurl.substring(iquery + 1,ihash);
        } else {
            if ( ieq>-1 && ( file==null || file.contains("=") && !( file.contains("(") || file.contains("{") ) ) ) { //TODO: this surely needs more attention.
                // file:///home/jbf/fun/camE_spot5/2012/05/$(d,Y=2012,m=04)/$H$M$S.jpg
                iquery = 0;
                if ( rsurl.startsWith("file:///") ) { // old code used to insert file://, so we check for it here in case of old URIs.
                    params= rsurl.substring(8,ihash);
                } else {
                    params= rsurl.substring(0,ihash);
                }
            } else {
                iquery = rsurl.length();
                fileEnd = rsurl.length();
            }
        }
        //if ( params!=null && params.length()==0 ) {  https://sourceforge.net/p/autoplot/bugs/1913/
        //    params=null;
        //}
        
        if ( ihash<rsurl.length() ) {
            result.filters= "|" + rsurl.substring(ihash+1).replaceAll("#","|");
        } else {
            result.filters= null;
        }

        if ( rsurl.substring(0,iquery ).contains("\\") ){
            rsurl= rsurl.substring(0,iquery).replaceAll("\\\\","/") + rsurl.substring(iquery);
        }
        if (result.scheme != null) {
            int iauth = result.scheme.length() + 1;
            while (iauth < rsurl.length() && rsurl.charAt(iauth) == '/') {
                iauth++;
            }
            iauth = rsurl.indexOf('/', iauth);
            if (iauth == -1) iauth = rsurl.length();
            if ( rsurl.charAt(iauth-1)==':' && rsurl.charAt(iauth-3)==':'  ) {
                    iauth= iauth-2;
            }
            result.authority = rsurl.substring(0, iauth);
        }

        if ( ext!=null && ext.length()==0 ) ext=null;
        
        if (file != null) {
            i = rsurl.lastIndexOf('/', iquery);
            if (i == -1) {
                result.path = rsurl.substring(0, iquery);
                result.file = rsurl.substring(0, iquery);
                result.ext = ext;
            } else {
                String surlDir = rsurl.substring(0, i);
                result.path = surlDir + "/";
                result.file = rsurl.substring(0, fileEnd);
                result.ext = ext;
            }
        }
        result.params = params;

        if ( "".equals(result.file) ) result.file=null;
        return result;


    }

    private static int indexOf(String s, char ch, char ignoreBegin, char ignoreEnd) {
        int i = s.indexOf(ch);
        int i0 = s.indexOf(ignoreBegin);
        int i1 = s.indexOf(ignoreEnd);
        if (i != -1 && i0 < i && i < i1) {
            i = -1;
        }
        return i;
    }

    /**
     * Split the parameters (if any) into name,value pairs. URLEncoded parameters are decoded, but the string may be decoded 
     * already.  Items without equals (=) are inserted as "arg_N"=name.
     * @param params null or String containing the list of ampersand-delimited parameters.
     * @return the map, which will be empty when there are no params.
     */
    public static LinkedHashMap<String, String> parseParams(String params) {
        LinkedHashMap<String, String> result = new LinkedHashMap<>();
        if (params == null) {
            return result;
        }
        if (params.trim().equals("")) {
            return result;
        }

        params = URISplit.uriDecode(params);
//        if ( params.contains("+") && params.contains(" ") ) {  // this may be a problem.  We know spaces are not encoded as pluses.
//            System.err.println("params appear to be decoded already"); // logger okay
//        } else {
//            if ( params.contains("+") && !params.contains("%20") ) { // legacy
//               params = params.replaceAll("+", " " );
//            }
//            params = URISplit.uriDecode(params);
//            //params = params.replaceAll("\\+", " "); // in the parameters, plus (+) is the same as space ( ).
//        }

        String[] ss = params.split("&");

        int argc = 0;

        for (String s : ss) {
            int j = indexOf(s, '=', '(', ')');
            String name, value;
            if (j == -1) {
                name = s;
                value = "";
                name = name.replaceAll("%3D", "=" ); // https://sourceforge.net/tracker/?func=detail&aid=3049295&group_id=199733&atid=970682
                result.put("arg_" + (argc++), name);
            } else {
                name = s.substring(0, j).trim();
                value = s.substring(j + 1);
                if ( name.equals( URISplit.PARAM_TIME_RANGE ) ) {
                    value= value.replaceAll("\\+", " ");
                }
                value = value.replaceAll("%3D", "=" ); // https://sourceforge.net/tracker/?func=detail&aid=3049295&group_id=199733&atid=970682
                value = value.replaceAll("%26", "&");
                result.put(name, value);
            }
        }
        return result;
    }

    /**
     * spaces and other URI syntax elements are URL-encoded.  
     * Note some calls of this routine should check for an empty string result
     * and then set split.params=null instead of "", to avoid the extraneous
     * question mark.
     * 
     * @param parms
     * @return "" or the parameters delimited by ampersands.
     */
    public static String formatParams(Map<String,String> parms) {
        StringBuilder result = new StringBuilder("");
        for ( Entry<String,String> e: parms.entrySet() ) {
            String key = (String) e.getKey();
            if (key.startsWith("arg_")) {
                if (!e.getValue().equals("")) {
                    result.append("&").append(e.getValue());
                }
            } else {
                String value = (String) e.getValue();
                if (value != null) {
                    if ( key.equals( URISplit.PARAM_TIME_RANGE ) ) {
                        value= value.replaceAll("\\s+","+");
                    }
                    value= value.replaceAll("&", "%26");
                    result.append("&").append(key).append("=").append(value);
                } else {
                    result.append("&").append(key);
                }
            }
        }
        return (result.length() == 0) ? "" : result.substring(1);
    }

    /**
     * format the URI using vapScheme, file and params.  
     * If file is missing but params is present, then return params:
     *   vap+cdaweb:ds=myds
     * If file is present, then format with file and params:
     *   vap+cdf:file://tmp/my.cdf?myVar
     * Else, just use the surl that is in there already. 
     * Note if split.params is non-null, it will be appended with a question mark, even if empty.
     * @param split
     * @return formatted URI.
     */
    public static String format(URISplit split) {
        String result = "";
        if ( split.vapScheme!=null && split.vapScheme.length()>0 && !split.vapScheme.equals("vap") ) result= result + split.vapScheme + ":";
        if ( split.file==null && split.params!=null ) {
            result= result + split.params;
        } else if ( split.file!=null ) {
            result= result + split.file;
            if (split.params != null ) { //&& split.params.length()>0 ) {  This is needed for completions.
                result += "?" + split.params;
            }
        } else if ( split.surl!=null ) {
            result+= split.surl;
        }
        return result;
    }
    
    /**
     * convenience method for creating URIs.  
     * @param vapScheme null or the data source scheme, such as "vap+das2server" or "vap+cdaweb"
     * @param resourceUri null or the resource uri, such as "http://www-pw.physics.uiowa.edu/das/das2Server"
     * @param args null or a map of arguments, including "arg_0" for a positional argument.  
     * @return the URI.  If vapScheme is null, then the URI will be implicit.
     * @see org.autoplot.jythonsupport#uri
     */
    public static String format( String vapScheme, String resourceUri, Map<String,Object> args ) {
        Map<String,String> largs;
        if ( args!=null ) {
            largs= new LinkedHashMap(); //
            for ( Entry<String,Object> e: args.entrySet() ) {
                if ( e.getValue()==null ) {
                    largs.put( e.getKey(), "" );
                } else {
                    largs.put( e.getKey(), String.valueOf(e.getValue()) );
                }
            }
        } else {
            largs= null;
        }
        if ( resourceUri==null ) {
            if ( vapScheme==null ) {
                throw new IllegalArgumentException("vapScheme must be specified when resourceUri is null");
            } 
            if ( largs!=null ) {
                return vapScheme + formatParams(largs);
            } else {
                return vapScheme;
            }
        } else {
            URISplit split= URISplit.parse(resourceUri);
            if ( vapScheme!=null ) {
                split.vapScheme= vapScheme;
            }
            if ( largs!=null ) {
                split.params= formatParams(largs);
            }
            return URISplit.format(split);
        }
    }
    
    /**
     * We need a standard way to detect if a string has already been URL encoded.
     * The problem is we want valid URIs that are also readable, so just using
     * simple encode/decode logic is not practical.
     *
     * This means:
     * - no spaces
     * - contains %[0-9][0-9]
     * @param surl
     * @return
     */
    public static boolean isUriEncoded( String surl ) {
        boolean result= false;
        // check for illegal characters.
        if ( surl.contains(" ") ) result= false;
        // check for encoded characters.
        if ( Pattern.compile("%[0-9A-F][0-9A-F]").matcher(surl).find() ) result= true;
        return result;
    }

    /**
     * convert " " to "%20", etc, by looking for and encoding illegal characters.
     * We can't just aggressively convert...
     * @param surl 
     * @return
     */
    public static String uriEncode(String surl) {
        if ( isUriEncoded(surl) ) return surl;
        surl = surl.replaceAll("%([^0-9])", "%25$1");  //%Y, %j, etc
        surl = surl.replaceAll("\\%24", "\\$"); // What's this--seems backward.  We like $'s in URIs...

        surl = surl.replaceAll(" ", "%20" );
        //surl = surl.replaceAll("#", "%23" );
        //surl = surl.replaceAll("%", "%25" ); // see above
        //surl = surl.replaceAll("&", "%26" );
        //surl = surl.replaceAll("\\+", "%2B" );
        //surl = surl.replaceAll("/", "%2F" );
        //surl = surl.replaceAll(":", "%3A" );
        //surl = surl.replaceAll(";", "%3B" );
        surl = surl.replaceAll("<", "%3C");
        surl = surl.replaceAll(">", "%3E");
        //surl = surl.replaceAll("\\?", "%3F" );
        surl = surl.replaceAll("\\[", "%5B"); // Windows appends these in temporary downloadf rte_1495358356
        surl = surl.replaceAll("\\]", "%5D");
        surl = surl.replaceAll("\\^", "%5E");

        return surl;
    }

    /**
     * convert "+" to " ", etc, by using URLDecoder and catching the UnsupportedEncodingException that will never occur.
     * We have to be careful for elements like %Y than are
     * not to be decoded.
     * TODO: we need to use standard escape/unescape code, possibly changing %Y to $Y beforehand.
     * @param s
     * @return
     */
    public static String uriDecode(String s) {
        if ( !isUriEncoded(s) ) return s;
        String surl= s;
//        if ( surl.contains("+") && !surl.contains("%20") ) { // legacy
//            surl = surl.replaceAll("+", " " );
//        }
        surl = surl.replaceAll("%20", " " );
        //surl = surl.replaceAll("%23", "#" );
        surl = surl.replaceAll("%25", "%" );
        //surl = surl.replaceAll("%26", "&" );
        surl = surl.replaceAll("%2B", "+" );
        //surl = surl.replaceAll("%2F", "/" );
        //surl = surl.replaceAll("%3A", ":" );
        //surl = surl.replaceAll("%3B", ";" );
        surl = surl.replaceAll("%3C", "<" );
        surl = surl.replaceAll("%3E", ">" );
        //surl = surl.replaceAll("%3F", "?" );
        surl = surl.replaceAll("%5B", "\\[" ); // Windows appends these in temporary downloadf rte_1495358356
        surl = surl.replaceAll("%5D", "\\]" );
        surl = surl.replaceAll("%5E", "^" );

        return surl;
    }

    /**
     * Helper method to get the timerange from the URI
     * @param uri
     * @return the DatumRange if "timerange=" is found, or null if not.
     * @throws ParseException
     */
    public static DatumRange parseTimeRange( String uri ) throws ParseException {
        URISplit split= URISplit.parse(uri);
        Map<String,String> params= URISplit.parseParams(split.params);
        String str= params.get( URISplit.PARAM_TIME_RANGE );
        if ( str!=null ) {
            DatumRange timerange= DatumRangeUtil.parseTimeRange( str );
            return timerange;
        } else {
            return null;
        }
    }

    @Override
    public String toString() {
        return "\nvapScheme: " + vapScheme + "\nscheme: " + scheme + "\nresourceUri: " + resourceUri + "\npath: " + path + "\nfile: " + file + "\next: " + ext + "\nparams: " + params + "\nfilters: "+filters + "\nsurl: " + surl + "\ncaretPos: " + resourceUriCarotPos + "\nformatCarotPos: " + formatCarotPos;
    }
}