001/*
002 * Licensed to the Apache Software Foundation (ASF) under one or more
003 * contributor license agreements.  See the NOTICE file distributed with
004 * this work for additional information regarding copyright ownership.
005 * The ASF licenses this file to You under the Apache License, Version 2.0
006 * (the "License"); you may not use this file except in compliance with
007 * the License.  You may obtain a copy of the License at
008 *
009 *     http://www.apache.org/licenses/LICENSE-2.0
010 *
011 * Unless required by applicable law or agreed to in writing, software
012 * distributed under the License is distributed on an "AS IS" BASIS,
013 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
014 * See the License for the specific language governing permissions and
015 * limitations under the License.
016 */
017package org.apache.commons.configuration2.tree.xpath;
018
019import java.util.ArrayList;
020import java.util.Collections;
021import java.util.LinkedList;
022import java.util.List;
023import java.util.StringTokenizer;
024
025import org.apache.commons.configuration2.tree.ExpressionEngine;
026import org.apache.commons.configuration2.tree.NodeAddData;
027import org.apache.commons.configuration2.tree.NodeHandler;
028import org.apache.commons.configuration2.tree.QueryResult;
029import org.apache.commons.jxpath.JXPathContext;
030import org.apache.commons.jxpath.ri.JXPathContextReferenceImpl;
031import org.apache.commons.lang3.StringUtils;
032
033/**
034 * <p>
035 * A specialized implementation of the {@code ExpressionEngine} interface that
036 * is able to evaluate XPATH expressions.
037 * </p>
038 * <p>
039 * This class makes use of <a href="http://commons.apache.org/jxpath/"> Commons
040 * JXPath</a> for handling XPath expressions and mapping them to the nodes of a
041 * hierarchical configuration. This makes the rich and powerful XPATH syntax
042 * available for accessing properties from a configuration object.
043 * </p>
044 * <p>
045 * For selecting properties arbitrary XPATH expressions can be used, which
046 * select single or multiple configuration nodes. The associated
047 * {@code Configuration} instance will directly pass the specified property keys
048 * into this engine. If a key is not syntactically correct, an exception will be
049 * thrown.
050 * </p>
051 * <p>
052 * For adding new properties, this expression engine uses a specific syntax: the
053 * &quot;key&quot; of a new property must consist of two parts that are
054 * separated by whitespace:
055 * </p>
056 * <ol>
057 * <li>An XPATH expression selecting a single node, to which the new element(s)
058 * are to be added. This can be an arbitrary complex expression, but it must
059 * select exactly one node, otherwise an exception will be thrown.</li>
060 * <li>The name of the new element(s) to be added below this parent node. Here
061 * either a single node name or a complete path of nodes (separated by the
062 * &quot;/&quot; character or &quot;@&quot; for an attribute) can be specified.</li>
063 * </ol>
064 * <p>
065 * Some examples for valid keys that can be passed into the configuration's
066 * {@code addProperty()} method follow:
067 * </p>
068 *
069 * <pre>
070 * &quot;/tables/table[1] type&quot;
071 * </pre>
072 *
073 * <p>
074 * This will add a new {@code type} node as a child of the first {@code table}
075 * element.
076 * </p>
077 *
078 * <pre>
079 * &quot;/tables/table[1] @type&quot;
080 * </pre>
081 *
082 * <p>
083 * Similar to the example above, but this time a new attribute named
084 * {@code type} will be added to the first {@code table} element.
085 * </p>
086 *
087 * <pre>
088 * &quot;/tables table/fields/field/name&quot;
089 * </pre>
090 *
091 * <p>
092 * This example shows how a complex path can be added. Parent node is the
093 * {@code tables} element. Here a new branch consisting of the nodes
094 * {@code table}, {@code fields}, {@code field}, and {@code name} will be added.
095 * </p>
096 *
097 * <pre>
098 * &quot;/tables table/fields/field@type&quot;
099 * </pre>
100 *
101 * <p>
102 * This is similar to the last example, but in this case a complex path ending
103 * with an attribute is defined.
104 * </p>
105 * <p>
106 * <strong>Note:</strong> This extended syntax for adding properties only works
107 * with the {@code addProperty()} method. {@code setProperty()} does not support
108 * creating new nodes this way.
109 * </p>
110 * <p>
111 * From version 1.7 on, it is possible to use regular keys in calls to
112 * {@code addProperty()} (i.e. keys that do not have to contain a whitespace as
113 * delimiter). In this case the key is evaluated, and the biggest part pointing
114 * to an existing node is determined. The remaining part is then added as new
115 * path. As an example consider the key
116 * </p>
117 *
118 * <pre>
119 * &quot;tables/table[last()]/fields/field/name&quot;
120 * </pre>
121 *
122 * <p>
123 * If the key does not point to an existing node, the engine will check the
124 * paths {@code "tables/table[last()]/fields/field"},
125 * {@code "tables/table[last()]/fields"}, {@code "tables/table[last()]"}, and so
126 * on, until a key is found which points to a node. Let's assume that the last
127 * key listed above can be resolved in this way. Then from this key the
128 * following key is derived: {@code "tables/table[last()] fields/field/name"} by
129 * appending the remaining part after a whitespace. This key can now be
130 * processed using the original algorithm. Keys of this form can also be used
131 * with the {@code setProperty()} method. However, it is still recommended to
132 * use the old format because it makes explicit at which position new nodes
133 * should be added. For keys without a whitespace delimiter there may be
134 * ambiguities.
135 * </p>
136 *
137 * @since 1.3
138 * @version $Id: XPathExpressionEngine.java 1790899 2017-04-10 21:56:46Z ggregory $
139 */
140public class XPathExpressionEngine implements ExpressionEngine
141{
142    /** Constant for the path delimiter. */
143    static final String PATH_DELIMITER = "/";
144
145    /** Constant for the attribute delimiter. */
146    static final String ATTR_DELIMITER = "@";
147
148    /** Constant for the delimiters for splitting node paths. */
149    private static final String NODE_PATH_DELIMITERS = PATH_DELIMITER
150            + ATTR_DELIMITER;
151
152    /**
153     * Constant for a space which is used as delimiter in keys for adding
154     * properties.
155     */
156    private static final String SPACE = " ";
157
158    /** Constant for a default size of a key buffer. */
159    private static final int BUF_SIZE = 128;
160
161    /** Constant for the start of an index expression. */
162    private static final char START_INDEX = '[';
163
164    /** Constant for the end of an index expression. */
165    private static final char END_INDEX = ']';
166
167    /** The internally used context factory. */
168    private final XPathContextFactory contextFactory;
169
170    /**
171     * Creates a new instance of {@code XPathExpressionEngine} with default
172     * settings.
173     */
174    public XPathExpressionEngine()
175    {
176        this(new XPathContextFactory());
177    }
178
179    /**
180     * Creates a new instance of {@code XPathExpressionEngine} and sets the
181     * context factory. This constructor is mainly used for testing purposes.
182     *
183     * @param factory the {@code XPathContextFactory}
184     */
185    XPathExpressionEngine(XPathContextFactory factory)
186    {
187        contextFactory = factory;
188    }
189
190    /**
191     * {@inheritDoc} This implementation interprets the passed in key as an XPATH
192     * expression.
193     */
194    @Override
195    public <T> List<QueryResult<T>> query(T root, String key,
196            NodeHandler<T> handler)
197    {
198        if (StringUtils.isEmpty(key))
199        {
200            QueryResult<T> result = createResult(root);
201            return Collections.singletonList(result);
202        }
203        else
204        {
205            JXPathContext context = createContext(root, handler);
206            List<?> results = context.selectNodes(key);
207            if (results == null)
208            {
209                results = Collections.emptyList();
210            }
211            return convertResults(results);
212        }
213    }
214
215    /**
216     * {@inheritDoc} This implementation creates an XPATH expression that
217     * selects the given node (under the assumption that the passed in parent
218     * key is valid). As the {@code nodeKey()} implementation of
219     * {@link org.apache.commons.configuration2.tree.DefaultExpressionEngine
220     * DefaultExpressionEngine} this method does not return indices for nodes.
221     * So all child nodes of a given parent with the same name have the same
222     * key.
223     */
224    @Override
225    public <T> String nodeKey(T node, String parentKey, NodeHandler<T> handler)
226    {
227        if (parentKey == null)
228        {
229            // name of the root node
230            return StringUtils.EMPTY;
231        }
232        else if (handler.nodeName(node) == null)
233        {
234            // paranoia check for undefined node names
235            return parentKey;
236        }
237
238        else
239        {
240            StringBuilder buf =
241                    new StringBuilder(parentKey.length()
242                            + handler.nodeName(node).length()
243                            + PATH_DELIMITER.length());
244            if (parentKey.length() > 0)
245            {
246                buf.append(parentKey);
247                buf.append(PATH_DELIMITER);
248            }
249            buf.append(handler.nodeName(node));
250            return buf.toString();
251        }
252    }
253
254    @Override
255    public String attributeKey(String parentKey, String attributeName)
256    {
257        StringBuilder buf =
258                new StringBuilder(StringUtils.length(parentKey)
259                        + StringUtils.length(attributeName)
260                        + PATH_DELIMITER.length() + ATTR_DELIMITER.length());
261        if (StringUtils.isNotEmpty(parentKey))
262        {
263            buf.append(parentKey).append(PATH_DELIMITER);
264        }
265        buf.append(ATTR_DELIMITER).append(attributeName);
266        return buf.toString();
267    }
268
269    /**
270     * {@inheritDoc} This implementation works similar to {@code nodeKey()}, but
271     * always adds an index expression to the resulting key.
272     */
273    @Override
274    public <T> String canonicalKey(T node, String parentKey,
275            NodeHandler<T> handler)
276    {
277        T parent = handler.getParent(node);
278        if (parent == null)
279        {
280            // this is the root node
281            return StringUtils.defaultString(parentKey);
282        }
283
284        StringBuilder buf = new StringBuilder(BUF_SIZE);
285        if (StringUtils.isNotEmpty(parentKey))
286        {
287            buf.append(parentKey).append(PATH_DELIMITER);
288        }
289        buf.append(handler.nodeName(node));
290        buf.append(START_INDEX);
291        buf.append(determineIndex(parent, node, handler));
292        buf.append(END_INDEX);
293        return buf.toString();
294    }
295
296    /**
297     * {@inheritDoc} The expected format of the passed in key is explained in
298     * the class comment.
299     */
300    @Override
301    public <T> NodeAddData<T> prepareAdd(T root, String key,
302            NodeHandler<T> handler)
303    {
304        if (key == null)
305        {
306            throw new IllegalArgumentException(
307                    "prepareAdd: key must not be null!");
308        }
309
310        String addKey = key;
311        int index = findKeySeparator(addKey);
312        if (index < 0)
313        {
314            addKey = generateKeyForAdd(root, addKey, handler);
315            index = findKeySeparator(addKey);
316        }
317        else if (index >= addKey.length() - 1)
318        {
319            invalidPath(addKey, " new node path must not be empty.");
320        }
321
322        List<QueryResult<T>> nodes =
323                query(root, addKey.substring(0, index).trim(), handler);
324        if (nodes.size() != 1)
325        {
326            throw new IllegalArgumentException("prepareAdd: key '" + key
327                    + "' must select exactly one target node!");
328        }
329
330        return createNodeAddData(addKey.substring(index).trim(), nodes.get(0));
331    }
332
333    /**
334     * Creates the {@code JXPathContext} to be used for executing a query. This
335     * method delegates to the context factory.
336     *
337     * @param root the configuration root node
338     * @param handler the node handler
339     * @return the new context
340     */
341    private <T> JXPathContext createContext(T root, NodeHandler<T> handler)
342    {
343        return getContextFactory().createContext(root, handler);
344    }
345
346    /**
347     * Creates a {@code NodeAddData} object as a result of a
348     * {@code prepareAdd()} operation. This method interprets the passed in path
349     * of the new node.
350     *
351     * @param path the path of the new node
352     * @param parentNodeResult the parent node
353     * @param <T> the type of the nodes involved
354     */
355    <T> NodeAddData<T> createNodeAddData(String path,
356            QueryResult<T> parentNodeResult)
357    {
358        if (parentNodeResult.isAttributeResult())
359        {
360            invalidPath(path, " cannot add properties to an attribute.");
361        }
362        List<String> pathNodes = new LinkedList<>();
363        String lastComponent = null;
364        boolean attr = false;
365        boolean first = true;
366
367        StringTokenizer tok =
368                new StringTokenizer(path, NODE_PATH_DELIMITERS, true);
369        while (tok.hasMoreTokens())
370        {
371            String token = tok.nextToken();
372            if (PATH_DELIMITER.equals(token))
373            {
374                if (attr)
375                {
376                    invalidPath(path, " contains an attribute"
377                            + " delimiter at a disallowed position.");
378                }
379                if (lastComponent == null)
380                {
381                    invalidPath(path,
382                            " contains a '/' at a disallowed position.");
383                }
384                pathNodes.add(lastComponent);
385                lastComponent = null;
386            }
387
388            else if (ATTR_DELIMITER.equals(token))
389            {
390                if (attr)
391                {
392                    invalidPath(path,
393                            " contains multiple attribute delimiters.");
394                }
395                if (lastComponent == null && !first)
396                {
397                    invalidPath(path,
398                            " contains an attribute delimiter at a disallowed position.");
399                }
400                if (lastComponent != null)
401                {
402                    pathNodes.add(lastComponent);
403                }
404                attr = true;
405                lastComponent = null;
406            }
407
408            else
409            {
410                lastComponent = token;
411            }
412            first = false;
413        }
414
415        if (lastComponent == null)
416        {
417            invalidPath(path, "contains no components.");
418        }
419
420        return new NodeAddData<>(parentNodeResult.getNode(), lastComponent,
421                attr, pathNodes);
422    }
423
424    /**
425     * Returns the {@code XPathContextFactory} used by this instance.
426     *
427     * @return the {@code XPathContextFactory}
428     */
429    XPathContextFactory getContextFactory()
430    {
431        return contextFactory;
432    }
433
434    /**
435     * Tries to generate a key for adding a property. This method is called if a
436     * key was used for adding properties which does not contain a space
437     * character. It splits the key at its single components and searches for
438     * the last existing component. Then a key compatible key for adding
439     * properties is generated.
440     *
441     * @param root the root node of the configuration
442     * @param key the key in question
443     * @param handler the node handler
444     * @return the key to be used for adding the property
445     */
446    private <T> String generateKeyForAdd(T root, String key,
447            NodeHandler<T> handler)
448    {
449        int pos = key.lastIndexOf(PATH_DELIMITER, key.length());
450
451        while (pos >= 0)
452        {
453            String keyExisting = key.substring(0, pos);
454            if (!query(root, keyExisting, handler).isEmpty())
455            {
456                StringBuilder buf = new StringBuilder(key.length() + 1);
457                buf.append(keyExisting).append(SPACE);
458                buf.append(key.substring(pos + 1));
459                return buf.toString();
460            }
461            pos = key.lastIndexOf(PATH_DELIMITER, pos - 1);
462        }
463
464        return SPACE + key;
465    }
466
467    /**
468     * Determines the index of the given child node in the node list of its
469     * parent.
470     *
471     * @param parent the parent node
472     * @param child the child node
473     * @param handler the node handler
474     * @param <T> the type of the nodes involved
475     * @return the index of this child node
476     */
477    private static <T> int determineIndex(T parent, T child,
478            NodeHandler<T> handler)
479    {
480        return handler.getChildren(parent, handler.nodeName(child)).indexOf(
481                child) + 1;
482    }
483
484    /**
485     * Helper method for throwing an exception about an invalid path.
486     *
487     * @param path the invalid path
488     * @param msg the exception message
489     */
490    private static void invalidPath(String path, String msg)
491    {
492        throw new IllegalArgumentException("Invalid node path: \"" + path
493                + "\" " + msg);
494    }
495
496    /**
497     * Determines the position of the separator in a key for adding new
498     * properties. If no delimiter is found, result is -1.
499     *
500     * @param key the key
501     * @return the position of the delimiter
502     */
503    private static int findKeySeparator(String key)
504    {
505        int index = key.length() - 1;
506        while (index >= 0 && !Character.isWhitespace(key.charAt(index)))
507        {
508            index--;
509        }
510        return index;
511    }
512
513    /**
514     * Converts the objects returned as query result from the JXPathContext to
515     * query result objects.
516     *
517     * @param results the list with results from the context
518     * @param <T> the type of results to be produced
519     * @return the result list
520     */
521    private static <T> List<QueryResult<T>> convertResults(List<?> results)
522    {
523        List<QueryResult<T>> queryResults =
524                new ArrayList<>(results.size());
525        for (Object res : results)
526        {
527            QueryResult<T> queryResult = createResult(res);
528            queryResults.add(queryResult);
529        }
530        return queryResults;
531    }
532
533    /**
534     * Creates a {@code QueryResult} object from the given result object of a
535     * query. Because of the node pointers involved result objects can only be
536     * of two types:
537     * <ul>
538     * <li>nodes of type T</li>
539     * <li>attribute results already wrapped in {@code QueryResult} objects</li>
540     * </ul>
541     * This method performs a corresponding cast. Warnings can be suppressed
542     * because of the implementation of the query functionality.
543     *
544     * @param resObj the query result object
545     * @param <T> the type of the result to be produced
546     * @return the {@code QueryResult}
547     */
548    @SuppressWarnings("unchecked")
549    private static <T> QueryResult<T> createResult(Object resObj)
550    {
551        if (resObj instanceof QueryResult)
552        {
553            return (QueryResult<T>) resObj;
554        }
555        else
556        {
557            return QueryResult.createNodeResult((T) resObj);
558        }
559    }
560
561    // static initializer: registers the configuration node pointer factory
562    static
563    {
564        JXPathContextReferenceImpl
565                .addNodePointerFactory(new ConfigurationNodePointerFactory());
566    }
567}