DefaultExpressionEngine.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.commons.configuration2.tree;

import java.util.Collection;
import java.util.LinkedList;
import java.util.List;

import org.apache.commons.lang3.StringUtils;

/**
 * <p>
 * A default implementation of the {@code ExpressionEngine} interface providing the &quot;native&quot; expression
 * language for hierarchical configurations.
 * </p>
 * <p>
 * This class implements a rather simple expression language for navigating through a hierarchy of configuration nodes.
 * It supports the following operations:
 * </p>
 * <ul>
 * <li>Navigating from a node to one of its children using the child node delimiter, which is by the default a dot
 * (&quot;.&quot;).</li>
 * <li>Navigating from a node to one of its attributes using the attribute node delimiter, which by default follows the
 * XPATH like syntax {@code [@&lt;attributeName&gt;]}.</li>
 * <li>If there are multiple child or attribute nodes with the same name, a specific node can be selected using a
 * numerical index. By default indices are written in parenthesis.</li>
 * </ul>
 * <p>
 * As an example consider the following XML document:
 * </p>
 *
 * <pre>
 *  &lt;database&gt;
 *    &lt;tables&gt;
 *      &lt;table type=&quot;system&quot;&gt;
 *        &lt;name&gt;users&lt;/name&gt;
 *        &lt;fields&gt;
 *          &lt;field&gt;
 *            &lt;name&gt;lid&lt;/name&gt;
 *            &lt;type&gt;long&lt;/name&gt;
 *          &lt;/field&gt;
 *          &lt;field&gt;
 *            &lt;name&gt;usrName&lt;/name&gt;
 *            &lt;type&gt;java.lang.String&lt;/type&gt;
 *          &lt;/field&gt;
 *         ...
 *        &lt;/fields&gt;
 *      &lt;/table&gt;
 *      &lt;table&gt;
 *        &lt;name&gt;documents&lt;/name&gt;
 *        &lt;fields&gt;
 *          &lt;field&gt;
 *            &lt;name&gt;docid&lt;/name&gt;
 *            &lt;type&gt;long&lt;/type&gt;
 *          &lt;/field&gt;
 *          ...
 *        &lt;/fields&gt;
 *      &lt;/table&gt;
 *      ...
 *    &lt;/tables&gt;
 *  &lt;/database&gt;
 * </pre>
 *
 * <p>
 * If this document is parsed and stored in a hierarchical configuration object, for instance the key
 * {@code tables.table(0).name} can be used to find out the name of the first table. In opposite
 * {@code tables.table.name} would return a collection with the names of all available tables. Similarly the key
 * {@code tables.table(1).fields.field.name} returns a collection with the names of all fields of the second table. If
 * another index is added after the {@code field} element, a single field can be accessed:
 * {@code tables.table(1).fields.field(0).name}. The key {@code tables.table(0)[@type]} would select the type attribute
 * of the first table.
 * </p>
 * <p>
 * This example works with the default values for delimiters and index markers. It is also possible to set custom values
 * for these properties so that you can adapt a {@code DefaultExpressionEngine} to your personal needs.
 * </p>
 * <p>
 * The concrete symbols used by an instance are determined by a {@link DefaultExpressionEngineSymbols} object passed to
 * the constructor. By providing a custom symbols object the syntax for querying properties in a hierarchical
 * configuration can be altered.
 * </p>
 * <p>
 * Instances of this class are thread-safe and can be shared between multiple hierarchical configuration objects.
 * </p>
 *
 * @since 1.3
 */
public class DefaultExpressionEngine implements ExpressionEngine {
    /**
     * A default instance of this class that is used as expression engine for hierarchical configurations per default.
     */
    public static final DefaultExpressionEngine INSTANCE = new DefaultExpressionEngine(DefaultExpressionEngineSymbols.DEFAULT_SYMBOLS);

    /** The symbols used by this instance. */
    private final DefaultExpressionEngineSymbols symbols;

    /** The matcher for node names. */
    private final NodeMatcher<String> nameMatcher;

    /**
     * Creates a new instance of {@code DefaultExpressionEngine} and initializes its symbols.
     *
     * @param syms the object with the symbols (must not be <strong>null</strong>)
     * @throws IllegalArgumentException if the symbols are <strong>null</strong>
     */
    public DefaultExpressionEngine(final DefaultExpressionEngineSymbols syms) {
        this(syms, null);
    }

    /**
     * Creates a new instance of {@code DefaultExpressionEngine} and initializes its symbols and the matcher for comparing
     * node names. The passed in matcher is always used when the names of nodes have to be matched against parts of
     * configuration keys.
     *
     * @param syms the object with the symbols (must not be <strong>null</strong>)
     * @param nodeNameMatcher the matcher for node names; can be <strong>null</strong>, then a default matcher is used
     * @throws IllegalArgumentException if the symbols are <strong>null</strong>
     */
    public DefaultExpressionEngine(final DefaultExpressionEngineSymbols syms, final NodeMatcher<String> nodeNameMatcher) {
        if (syms == null) {
            throw new IllegalArgumentException("Symbols must not be null!");
        }

        symbols = syms;
        nameMatcher = nodeNameMatcher != null ? nodeNameMatcher : NodeNameMatchers.EQUALS;
    }

    @Override
    public String attributeKey(final String parentKey, final String attributeName) {
        final DefaultConfigurationKey key = new DefaultConfigurationKey(this, parentKey);
        key.appendAttribute(attributeName);
        return key.toString();
    }

    /**
     * {@inheritDoc} This implementation works similar to {@code nodeKey()}; however, each key returned by this method has
     * an index (except for the root node). The parent key is prepended to the name of the current node in any case and
     * without further checks. If it is <strong>null</strong>, only the name of the current node with its index is returned.
     */
    @Override
    public <T> String canonicalKey(final T node, final String parentKey, final NodeHandler<T> handler) {
        final String nodeName = handler.nodeName(node);
        final T parent = handler.getParent(node);
        final DefaultConfigurationKey key = new DefaultConfigurationKey(this, parentKey);
        key.append(StringUtils.defaultString(nodeName));

        if (parent != null) {
            // this is not the root key
            key.appendIndex(determineIndex(node, parent, nodeName, handler));
        }
        return key.toString();
    }

    /**
     * Determines the index of the given node based on its parent node.
     *
     * @param node the current node
     * @param parent the parent node
     * @param nodeName the name of the current node
     * @param handler the node handler
     * @param <T> the type of the nodes to be dealt with
     * @return the index of this node
     */
    private <T> int determineIndex(final T node, final T parent, final String nodeName, final NodeHandler<T> handler) {
        return findChildNodesByName(handler, parent, nodeName).indexOf(node);
    }

    /**
     * Returns a list with all child nodes of the given parent node which match the specified node name. The match is done
     * using the current node name matcher.
     *
     * @param handler the {@code NodeHandler}
     * @param parent the parent node
     * @param nodeName the name of the current node
     * @param <T> the type of the nodes to be dealt with
     * @return a list with all matching child nodes
     */
    private <T> List<T> findChildNodesByName(final NodeHandler<T> handler, final T parent, final String nodeName) {
        return handler.getMatchingChildren(parent, nameMatcher, nodeName);
    }

    /**
     * Finds the last existing node for an add operation. This method traverses the node tree along the specified key. The
     * last existing node on this path is returned.
     *
     * @param <T> the type of the nodes to be dealt with
     * @param keyIt the key iterator
     * @param node the current node
     * @param handler the node handler
     * @return the last existing node on the given path
     */
    protected <T> T findLastPathNode(final DefaultConfigurationKey.KeyIterator keyIt, final T node, final NodeHandler<T> handler) {
        final String keyPart = keyIt.nextKey(false);

        if (keyIt.hasNext()) {
            if (!keyIt.isPropertyKey()) {
                // Attribute keys can only appear as last elements of the path
                throw new IllegalArgumentException("Invalid path for add operation: " + "Attribute key in the middle!");
            }
            final int idx = keyIt.hasIndex() ? keyIt.getIndex() : handler.getMatchingChildrenCount(node, nameMatcher, keyPart) - 1;
            if (idx < 0 || idx >= handler.getMatchingChildrenCount(node, nameMatcher, keyPart)) {
                return node;
            }
            return findLastPathNode(keyIt, findChildNodesByName(handler, node, keyPart).get(idx), handler);
        }
        return node;
    }

    /**
     * Recursive helper method for evaluating a key. This method processes all facets of a configuration key, traverses the
     * tree of properties and fetches the results of all matching properties.
     *
     * @param <T> the type of nodes to be dealt with
     * @param keyPart the configuration key iterator
     * @param node the current node
     * @param results here the found results are stored
     * @param handler the node handler
     */
    protected <T> void findNodesForKey(final DefaultConfigurationKey.KeyIterator keyPart, final T node, final Collection<QueryResult<T>> results,
        final NodeHandler<T> handler) {
        if (!keyPart.hasNext()) {
            results.add(QueryResult.createNodeResult(node));
        } else {
            final String key = keyPart.nextKey(false);
            if (keyPart.isPropertyKey()) {
                processSubNodes(keyPart, findChildNodesByName(handler, node, key), results, handler);
            }
            if (keyPart.isAttribute() && !keyPart.hasNext() && handler.getAttributeValue(node, key) != null) {
                results.add(QueryResult.createAttributeResult(node, key));
            }
        }
    }

    /**
     * Gets the {@code DefaultExpressionEngineSymbols} object associated with this instance.
     *
     * @return the {@code DefaultExpressionEngineSymbols} used by this engine
     * @since 2.0
     */
    public DefaultExpressionEngineSymbols getSymbols() {
        return symbols;
    }

    /**
     * {@inheritDoc} This implementation takes the given parent key, adds a property delimiter, and then adds the node's
     * name. The name of the root node is a blank string. Note that no indices are returned.
     */
    @Override
    public <T> String nodeKey(final T node, final String parentKey, final NodeHandler<T> handler) {
        if (parentKey == null) {
            // this is the root node
            return StringUtils.EMPTY;
        }
        final DefaultConfigurationKey key = new DefaultConfigurationKey(this, parentKey);
        key.append(handler.nodeName(node), true);
        return key.toString();
    }

    /**
     * <p>
     * Prepares Adding the property with the specified key.
     * </p>
     * <p>
     * To be able to deal with the structure supported by hierarchical configuration implementations the passed in key is of
     * importance, especially the indices it might contain. The following example should clarify this: Suppose the current
     * node structure looks like the following:
     * </p>
     *
     * <pre>
     *  tables
     *     +-- table
     *             +-- name = user
     *             +-- fields
     *                     +-- field
     *                             +-- name = uid
     *                     +-- field
     *                             +-- name = firstName
     *                     ...
     *     +-- table
     *             +-- name = documents
     *             +-- fields
     *                    ...
     * </pre>
     * <p>
     * In this example a database structure is defined, for example all fields of the first table could be accessed using the key
     * {@code tables.table(0).fields.field.name}. If now properties are to be added, it must be exactly specified at which
     * position in the hierarchy the new property is to be inserted. So to add a new field name to a table it is not enough
     * to say just
     * </p>
     *
     * <pre>
     * config.addProperty(&quot;tables.table.fields.field.name&quot;, &quot;newField&quot;);
     * </pre>
     * <p>
     * The statement given above contains some ambiguity. For instance it is not clear, to which table the new field should
     * be added. If this method finds such an ambiguity, it is resolved by following the last valid path. Here this would be
     * the last table. The same is true for the {@code field}; because there are multiple fields and no explicit index is
     * provided, a new {@code name} property would be added to the last field - which is probably not what was desired.
     * </p>
     * <p>
     * To make things clear explicit indices should be provided whenever possible. In the example above the exact table
     * could be specified by providing an index for the {@code table} element as in {@code tables.table(1).fields}. By
     * specifying an index it can also be expressed that at a given position in the configuration tree a new branch should
     * be added. In the example above we did not want to add an additional {@code name} element to the last field of the
     * table, but we want a complete new {@code field} element. This can be achieved by specifying an invalid index (like
     * -1) after the element where a new branch should be created. Given this our example would run:
     * </p>
     *
     * <pre>
     * config.addProperty(&quot;tables.table(1).fields.field(-1).name&quot;, &quot;newField&quot;);
     * </pre>
     * <p>
     * With this notation it is possible to add new branches everywhere. We could for instance create a new {@code table}
     * element by specifying
     * </p>
     *
     * <pre>
     * config.addProperty(&quot;tables.table(-1).fields.field.name&quot;, &quot;newField2&quot;);
     * </pre>
     * <p>
     * (Note that because after the {@code table} element a new branch is created indices in following elements are not
     * relevant; the branch is new so there cannot be any ambiguities.)
     * </p>
     *
     * @param <T> the type of the nodes to be dealt with
     * @param root the root node of the nodes hierarchy
     * @param key the key of the new property
     * @param handler the node handler
     * @return a data object with information needed for the add operation
     */
    @Override
    public <T> NodeAddData<T> prepareAdd(final T root, final String key, final NodeHandler<T> handler) {
        final DefaultConfigurationKey.KeyIterator it = new DefaultConfigurationKey(this, key).iterator();
        if (!it.hasNext()) {
            throw new IllegalArgumentException("Key for add operation must be defined!");
        }

        final T parent = findLastPathNode(it, root, handler);
        final List<String> pathNodes = new LinkedList<>();

        while (it.hasNext()) {
            if (!it.isPropertyKey()) {
                throw new IllegalArgumentException("Invalid key for add operation: " + key + " (Attribute key in the middle.)");
            }
            pathNodes.add(it.currentKey());
            it.next();
        }

        return new NodeAddData<>(parent, it.currentKey(), !it.isPropertyKey(), pathNodes);
    }

    /**
     * Called by {@code findNodesForKey()} to process the sub nodes of the current node depending on the type of the current
     * key part (children, attributes, or both).
     *
     * @param <T> the type of the nodes to be dealt with
     * @param keyPart the key part
     * @param subNodes a list with the sub nodes to process
     * @param nodes the target collection
     * @param handler the node handler
     */
    private <T> void processSubNodes(final DefaultConfigurationKey.KeyIterator keyPart, final List<T> subNodes, final Collection<QueryResult<T>> nodes,
        final NodeHandler<T> handler) {
        if (keyPart.hasIndex()) {
            if (keyPart.getIndex() >= 0 && keyPart.getIndex() < subNodes.size()) {
                findNodesForKey((DefaultConfigurationKey.KeyIterator) keyPart.clone(), subNodes.get(keyPart.getIndex()), nodes, handler);
            }
        } else {
            subNodes.forEach(node -> findNodesForKey((DefaultConfigurationKey.KeyIterator) keyPart.clone(), node, nodes, handler));
        }
    }

    /**
     * {@inheritDoc} This method supports the syntax as described in the class comment.
     */
    @Override
    public <T> List<QueryResult<T>> query(final T root, final String key, final NodeHandler<T> handler) {
        final List<QueryResult<T>> results = new LinkedList<>();
        findNodesForKey(new DefaultConfigurationKey(this, key).iterator(), root, results, handler);
        return results;
    }
}