jena-core/src/main/java/org/apache/jena/util/SplitIRI.java

/*
 * Licensed to the Apache Software Foundation (ASF) under one
 * or more contributor license agreements.  See the NOTICE file
 * distributed with this work for additional information
 * regarding copyright ownership.  The ASF licenses this file
 * to you under the Apache License, Version 2.0 (the
 * "License"); you may not use this file except in compliance
 * with the License.  You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

package org.apache.jena.util;

import java.util.Objects;

import org.apache.jena.graph.Node;
//Has copies of    import org.apache.jena.riot.system.RiotChars;

/**
 * Code to split an URI or IRI into prefix and local part.
 * This code does <em>not</em> consider any prefix mapping.
 * The split is based on finding the last {@code /} or {@code #}
 * character.
 * <p>
 * Historically, 'prefix' is referred to as 'namespace'
 * reflecting RDF/XML history.
 * <p>
 * For display, use {@link #localname} and {@link #namespace}.
 * These pragmatic follow Turtle (e.g. localname allows a plain trailing dot)
 * but does not escape any characters.
 *
 * These operations are for display (c.f. {@link Object#toString} and do not guarantee a round-trip;
 * namespace+localname may not be the  exact input IRI.
 * A URI is split never split before the last {@code /} or {@code #}, if present.
 * See {@link #splitpoint} for more details.
 * <p>
 * The functions
 * {@link #namespaceTTL} and {@link #localnameTTL}
 * provide a strict Turtle split, if possible;
 * the local name is escaped if necessary.
 * {@link #namespaceTTL} can be used to build a set of prefix mappings.
 * <p>
 * The functions {@link #namespaceXML} and {@link #localnameXML}
 * apply the rules for XML qnames.
 * <p>
 * This code forms the machinery behind {@link Node#getLocalName}
 * {@link Node#getNameSpace} for URI Nodes following the XML rules.
 */
public class SplitIRI
{
    /** Return the 'namespace' (prefix) for a URI string.
     * Use with {@link #localname}.
     * Return the input string if there is no splitpoint.
     */
    public static String namespace(String string) {
        Objects.requireNonNull(string, "string argument is null");
        int i = splitpoint(string);
        if ( i < 0 )
            return string;
        return string.substring(0, i);
    }

    /** Calculate a localname - do not escape PN_LOCAL_ESC.
     * This is not guaranteed to be legal Turtle.
     * Use with {@link #namespace}
     * Return an empty string if there is no split point.
     */
    public static String localname(String string) {
        Objects.requireNonNull(string, "string argument is null");
        int i = splitpoint(string);
        if ( i < 0 )
            return "";
        return string.substring(i);
    }

    /**
     * Return the 'namespace' (prefix) for a URI string,
     * legal for Turtle and goes with {@link #localnameTTL}.
     * This operation does not guaratee that the argument has a legal localname.
     */
    public static String namespaceTTL(String string) {
        int i = splitpoint(string);
        if ( i < 0 )
            return string;
        String ns = string.substring(0, i);
        return ns;
    }

    /**
     * Calculate a localname - enforce legal Turtle
     * escapes for localnames (rule PN_LOCAL_ESC),
     * A final '.' is escaped.
     * Return "" for "no split".
     * Use with {@link #namespaceTTL}
     */
    public static String localnameTTL(String string) {
        String x = localname(string);
        if ( x.isEmpty())
            return x;
        // This will escape the final DOT but leave internal dots alone.
        return escape_PN_LOCAL_ESC(x);
    }

    private static String escape_PN_LOCAL_ESC(String x) {
        // Assume that escapes are rare so scan once to make sure there
        // is work to do then scan again doing the work.
        //'\' ('_' | '~' | '.' | '-' | '!' | '$' | '&' | "'" | '(' | ')' | '*' | '+' | ',' | ';' | '=' | '/' | '?' | '#' | '@' | '%')

        int N = x.length();
        boolean escchar = false;
        for ( int i = 0; i < N; i++ ) {
            char ch = x.charAt(i);
            if ( needsEscape(ch, (i==N-1)) ) {
                escchar = true;
                break;
            }
        }
        if ( ! escchar )
            return x;
        StringBuilder sb = new StringBuilder(N+10);
        for ( int i = 0; i < N; i++ ) {
            char ch = x.charAt(i);
            // DOT only needs escaping at the end
            if ( needsEscape(ch, (i==N-1) )  )
                sb.append('\\');
            sb.append(ch);
        }
        return sb.toString();
    }

    /**
     * Scan for chars needing escape.
     * Return the index of the first.
     * Return -1 for no escape needed.
     *
     */
    private static int needsEscape(String x) {
        int N = x.length();
        for ( int i = 0; i < N; i++ ) {
            char ch = x.charAt(i);
            if ( needsEscape(ch, (i==N-1)) ) {
                return i;
            }
        }
        return -1;
    }

    private static boolean needsEscape(char ch, boolean finalChar) {
        if ( ch == '.' )
            // Only needed at the end.
            return finalChar;
        return isPN_LOCAL_ESC(ch);
    }

    // @formatter:off
    /* From the RDF 1.1 Turtle specification:
        [136s]  PrefixedName    ::=     PNAME_LN | PNAME_NS
        Productions for terminals

        [163s]  PN_CHARS_BASE   ::=     [A-Z] | [a-z] | [#x00C0-#x00D6] | [#x00D8-#x00F6] | [#x00F8-#x02FF] | [#x0370-#x037D] | [#x037F-#x1FFF] | [#x200C-#x200D] | [#x2070-#x218F] | [#x2C00-#x2FEF] | [#x3001-#xD7FF] | [#xF900-#xFDCF] | [#xFDF0-#xFFFD] | [#x10000-#xEFFFF]
        [164s]  PN_CHARS_U  ::=     PN_CHARS_BASE | '_'
        [166s]  PN_CHARS    ::=     PN_CHARS_U | '-' | [0-9] | #x00B7 | [#x0300-#x036F] | [#x203F-#x2040]
        [167s]  PN_PREFIX   ::=     PN_CHARS_BASE ((PN_CHARS | '.')* PN_CHARS)?

        [168s]  PN_LOCAL    ::=     (PN_CHARS_U | ':' | [0-9] | PLX) ((PN_CHARS | '.' | ':' | PLX)* (PN_CHARS | ':' | PLX))?
        [169s]  PLX     ::=     PERCENT | PN_LOCAL_ESC
        [170s]  PERCENT     ::=     '%' HEX HEX
        [171s]  HEX     ::=     [0-9] | [A-F] | [a-f]
        [172s]  PN_LOCAL_ESC    ::=     '\' ('_' | '~' | '.' | '-' | '!' | '$' | '&' | "'" | '(' | ')' | '*' | '+' | ',' | ';' | '=' | '/' | '?' | '#' | '@' | '%')
*/
    // @formatter:on

    /** Find the URI split point, return the index into the string that is the
     *  first character of a legal Turtle local name.
     * <p>
     * This is a pragmatic choice, not just finding the maximal point.
     * For example, with escaping '/' can be included but that means
     * {@code http://example/path/abc} could split to give {@code http://example/}
     * and {@code path/abc} .
     * <p>
     * Split URN's after ':'.
     * <p>
     * This function does not enforce the Turtle rule that the final character can not be '.'.
     *
     * @param uri URI string
     * @return The split point, or -1 for "not found".
     */
    public static int splitpoint(String uri) {
        boolean isURN = uri.startsWith("urn:");
        // Fast track.  Still need to check validity of the prefix part.
        int idx1 = uri.lastIndexOf('#');
        // Not so simple - \/ in local names
        int idx2 = isURN ? uri.lastIndexOf(':') : uri.lastIndexOf('/');

        // If absolute.
        int idx3 = uri.indexOf(':');

        // Note: local names can't end in "." in Turtle.
        // This is handled by escape_PN_LOCAL_ESC which will escape it as "\."

        // Cases
        //   "abc#def"
        //   "/abc"
        //   "/"
        //   "/path/path#frag
        //   "/path/path#abc/def" :: / in fragment, split is at the "#".

        int limit;
        if ( idx1 >= 0 && idx2 < 0 ) {
            // No path "/" (or ":" if a URN)
            limit = idx1;
        } else if ( idx1 < 0 && idx2 >= 0 ) {
            // No fragment
            limit = idx2;
        } else if ( idx1 >= 0 && idx2 >= 0 ) {
            // Fragment and path. Use fragment.
            // If "/" is in the fragment, it is not the split point.
            limit = idx1;
        } else {
            limit = -1;
        }

        // At least idx3, the case of no "/" and no "#" in an absolute IRI
        if ( idx3 >= 0 )
            limit = Math.max(limit, idx3);

        // Limit is our guess.
        // Now search end of URI to this guess checking the characters found.

        int splitPoint = -1;
        // Work backwards, checking for
        // ((PN_CHARS | '.' | ':' | PLX)*
        for ( int i = uri.length()-1; i > limit; i-- ) {
            char ch = uri.charAt(i);
            if ( /*RiotChars.*/isPNChars_U_N(ch) || /*RiotChars.*/isPN_LOCAL_ESC(ch) || ch == ':' || ch == '-' || ch == '.' )
                continue;
            splitPoint = i+1;
            break;
        }
        // limit was at the end.  No split point (we could escape the limit point)
        if ( splitPoint == -1 )
            splitPoint = limit+1;
        // No split point.
        if ( splitPoint >= uri.length() )
            return -1;

        // Check the first character of the local name.
        // All characters are legal localname name characters but may not satisfy the additional
        // first character rule.  Move forward to first legal first character.
        int ch = uri.charAt(splitPoint);
        while ( ch == '.' || ch == '-' ) {
            splitPoint++;
            if ( splitPoint >= uri.length() )
                return -1;
            ch = uri.charAt(splitPoint);
        }

        // Checking the final '.' is done when checking for escapes.
        return splitPoint;
    }

    private static boolean checkhex(String uri, int i) {
        return /*RiotChars.*/isHexChar(uri.charAt(i));
    }

    /**
     * Split point, according to XML 1.0 qname rules.
     * This is the longest XML 1.0 NCName at the end of the uri.
     * Return a split at the end of the string if there is no match
     * (e.g. the URI string ends in '/' or '#').
     */
    public static int splitXML(String string) { return splitXML11(string); }

    /**
     * splitXML with local stricter XML 1.0 rules (only single Java characters)
     * @deprecated Retained as a record of previous split handling.
     */
    @Deprecated(forRemoval = true)
    public static int splitXML10(String string) { return splitNamespaceXML10(string); }

    /**
     * Split point, according to XML 1.1 qname rules.
     * This is the longest XML 1.1 NCName at the end of the uri.
     * Return a split at the end of the string if there is no match
     * (e.g. the URI string ends in '/' or '#').
     */
    private static int splitXML11(String string) { return splitNamespaceXML11(string); }

    /**
     * Namespace, according to XML 1.0 qname rules.
     * Use with {@link #localnameXML}.
     */
    public static String namespaceXML(String string) {
        int i = splitXML(string);
        return string.substring(0, i);
    }

    /** Localname, according to XML 1.0 qname rules. */
    public static String localnameXML(String string) {
        int i = splitXML(string);
        return string.substring(i);
    }

    // Extracted from RiotChars
    // When/if RIOT becomes accessible to this code, then refactor

    private static boolean /*RiotChars.*/isPN_LOCAL_ESC(char ch) {
        switch (ch) {
            case '\\': case '_':  case '~': case '.': case '-': case '!': case '$':
            case '&':  case '\'': case '(': case ')': case '*': case '+': case ',':
            case ';':  case '=':  case '/': case '?': case '#': case '@': case '%':
                return true;
            default:
                return false;
        }
    }

    /** ASCII 0-9 */
    private static boolean isDigit(int ch) {
        return range(ch, '0', '9');
    }

    private static boolean isPNCharsBase(int ch) {
        // PN_CHARS_BASE ::= [A-Z] | [a-z] | [#x00C0-#x00D6] | [#x00D8-#x00F6] | [#x00F8-#x02FF] |
        //                   [#x0370-#x037D] | [#x037F-#x1FFF] | [#x200C-#x200D] | [#x2070-#x218F] |
        //                   [#x2C00-#x2FEF] | [#x3001-#xD7FF] | [#xF900-#xFDCF] | [#xFDF0-#xFFFD] |
        //                   [#x10000-#xEFFFF]
        return
            r(ch, 'a', 'z') || r(ch, 'A', 'Z') || r(ch, 0x00C0, 0x00D6) || r(ch, 0x00D8, 0x00F6) || r(ch, 0x00F8, 0x02FF) ||
            r(ch, 0x0370, 0x037D) || r(ch, 0x037F, 0x1FFF) || r(ch, 0x200C, 0x200D) || r(ch, 0x2070, 0x218F) ||
            r(ch, 0x2C00, 0x2FEF) || r(ch, 0x3001, 0xD7FF) ||
            // Surrogate pairs
            r(ch, 0xD800, 0xDFFF) ||
            r(ch, 0xF900, 0xFDCF) || r(ch, 0xFDF0, 0xFFFD) ||
            r(ch, 0x10000, 0xEFFFF); // Outside the basic plane.
    }

    private static boolean isPNChars_U(int ch) {
        //PN_CHARS_BASE | '_'
        return isPNCharsBase(ch) || ( ch == '_' );
    }

    private static boolean isPNChars_U_N(int ch) {
        // PN_CHARS_U | [0-9]
        return isPNCharsBase(ch) || ( ch == '_' ) || isDigit(ch);
    }

    private static boolean isPNChars(int ch) {
        // PN_CHARS ::=  PN_CHARS_U | '-' | [0-9] | #x00B7 | [#x0300-#x036F] | [#x203F-#x2040]
        return isPNChars_U(ch) || isDigit(ch) || ( ch == '-' ) || ch == 0x00B7 || r(ch, 0x300, 0x036F) || r(ch, 0x203F, 0x2040);
    }

    /** Hexadecimal character */
    private static boolean isHexChar(int ch) {
        return range(ch, '0', '9') || range(ch, 'a', 'f') || range(ch, 'A', 'F');
    }

    private static int valHexChar(int ch) {
        if ( range(ch, '0', '9') )
            return ch - '0';
        if ( range(ch, 'a', 'f') )
            return ch - 'a' + 10;
        if ( range(ch, 'A', 'F') )
            return ch - 'A' + 10;
        return -1;
    }

    private static boolean r(int ch, int a, int b) { return ( ch >= a && ch <= b ); }

    private static boolean range(int ch, char a, char b) {
        return (ch >= a && ch <= b);
    }

    // -------- --------
    // Splitting for XML.
    //   An additional rule is that the local part can not start with a digit.
    //   In Turtle, it can.

    // XML Namespaces:
    // A qname name is NCName ':' NCName
    // NCName             ::=      NCNameStartChar NCNameChar*
    // NCNameChar         ::=      NameChar - ':'
    // NCNameStartChar    ::=      Letter | '_'
    //
    // XML 1.0
    // NameStartChar      ::= ":" | [A-Z] | "_" | [a-z] | [#xC0-#xD6] |
    //                        [#xD8-#xF6] | [#xF8-#x2FF] |
    //                        [#x370-#x37D] | [#x37F-#x1FFF] |
    //                        [#x200C-#x200D] | [#x2070-#x218F] |
    //                        [#x2C00-#x2FEF] | [#x3001-#xD7FF] |
    //                        [#xF900-#xFDCF] | [#xFDF0-#xFFFD] | [#x10000-#xEFFFF]
    // NameChar           ::= NameStartChar | "-" | "." | [0-9] | #xB7 |
    //                        [#x0300-#x036F] | [#x203F-#x2040]
    // Name               ::= NameStartChar (NameChar)*

    // XMLChar does not support [#x10000-#xEFFFF]
    // XML11Char does support [#x10000-#xEFFFF]


    /**
     * Given an absolute URI, determine the split point between the namespace
     * part and the localname part. If there is no valid localname part then the
     * length of the string is returned. The algorithm tries to find the longest
     * NCName at the end of the uri, not immediately preceded by the first
     * colon in the string.
     * <p>
     * This operation follows XML QName rules which are more complicated than
     * needed for Turtle and TriG.   For example, QName can't start with a digit.
     *
     * @param uri
     * @return the index of the first character of the localname
     * @see SplitIRI
     */
    private static int splitNamespaceXML11(String uri) {
        // Updated for XML 1.1
        // The main difference is that XML11Char class works for character beyond 0xFFFF.
        // This is not a restriction in XML 1.0 itself.

        char ch;
        int lg = uri.length();
        if (lg == 0)
            return 0;
        int i = lg-1;
        for (; i >= 1; i--) {
            ch = uri.charAt(i);
            if ( !XML11Char.isXML11NCName(ch) ) // ****
                break;
        }

        int j = i + 1;

        if ( j >= lg )
            return lg;

        // Check we haven't split up a %-encoding.
        if ( j >= 2 && uri.charAt(j-2) == '%' )
            j = j+1;
        if ( j >= 1 && uri.charAt(j-1) == '%' ) {
            j = j+2;
            if ( j > lg )
                // JENA-1941: Protect against overshoot in the case of "%x"
                // at end of a (bad) URI.
                return lg;
        }

        // Have found the leftmost NCNameChar from the
        // end of the URI string.
        // Now scan forward for an NCNameStartChar
        // The split must start with NCNameStart.
        for (; j < lg; j++) {
            ch = uri.charAt(j);
            if (XML11Char.isXML11NCNameStart(ch)) // ****
            {
                // "mailto:" is special.
                // split "mailto:me" as "mailto:m" and "e" !
                // Keep part after mailto: with at least one character.
                if ( j == 7 && uri.startsWith("mailto:"))
                    // Don't split at "mailto:"
                    continue;
                else
                    break;
            }
        }
        return j;
    }

    private static int splitNamespaceXML10(String uri) {
        // The original code.
        char ch;
        int lg = uri.length();
        if (lg == 0)
            return 0;
        int i = lg-1;
        for (; i >= 1; i--) {
            ch = uri.charAt(i);
            if ( !XMLChar.isNCName(ch) )
                break;
        }

        int j = i + 1;

        if ( j >= lg )
            return lg;

        // Check we haven't split up a %-encoding.
        if ( j >= 2 && uri.charAt(j-2) == '%' )
            j = j+1;
        if ( j >= 1 && uri.charAt(j-1) == '%' ) {
            j = j+2;
            if ( j > lg )
                // JENA-1941: Protect against overshoot in the case of "%x"
                // at end of a (bad) URI.
                return lg;
        }

        // Have found the leftmost NCNameChar from the
        // end of the URI string.
        // Now scan forward for an NCNameStartChar
        // The split must start with NCNameStart.
        for (; j < lg; j++) {
            ch = uri.charAt(j);
            if (XMLChar.isNCNameStart(ch))
            {
                // "mailto:" is special.
                // split "mailto:me" as "mailto:m" and "e" !
                // Keep part after mailto: with at least one character.
                if ( j == 7 && uri.startsWith("mailto:"))
                    // Don't split at "mailto:"
                    continue;
                else
                    break;
            }
        }
        return j;
    }

}