2 * Copyright 2005 the original author or authors.
4 * Licensed under the Apache License, Version 2.0 (the "License");
5 * you may not use this file except in compliance with the License.
6 * You may obtain a copy of the License at
8 * http://www.apache.org/licenses/LICENSE-2.0
10 * Unless required by applicable law or agreed to in writing, software
11 * distributed under the License is distributed on an "AS IS" BASIS,
12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 * See the License for the specific language governing permissions and
14 * limitations under the License.
17 package org.wamblee.crawler.impl;
19 import java.io.InputStream;
20 import java.util.ArrayList;
21 import java.util.Iterator;
22 import java.util.List;
24 import org.apache.commons.httpclient.NameValuePair;
25 import org.dom4j.Document;
26 import org.dom4j.DocumentException;
27 import org.dom4j.Element;
28 import org.dom4j.io.SAXReader;
29 import org.wamblee.crawler.Configuration;
30 import org.wamblee.crawler.GetPageRequest;
31 import org.wamblee.crawler.PageRequest;
32 import org.wamblee.crawler.PostPageRequest;
33 import org.wamblee.xml.XslTransformer;
36 * Parsing of the configuration from an XML file.
38 public class ConfigurationParser {
40 private static final String ELEM_URL = "url";
42 private static final String ELEM_TYPE = "type";
44 private static final String ELEM_PATTERN = "pattern";
46 private static final String ELEM_METHOD = "method";
48 private static final String ELEM_XSLT = "xslt";
50 private static final String ELEM_PARAM = "param";
52 private static final String AT_NAME = "name";
54 private static final String AT_VALUE = "value";
56 private static final String METHOD_POST = "post";
58 private static final String METHOD_GET = "get";
60 private static final int MAX_TRIES = 3;
62 private static final int MAX_DELAY = 100;
64 private XslTransformer _transformer;
67 * Constructs the configuration parser.
69 public ConfigurationParser(XslTransformer aTransformer) {
70 _transformer = aTransformer;
74 * Parses the configuration from an input stream.
75 * @param aStream Input file.
76 * @return Configuration.
78 public Configuration parse(InputStream aStream) {
80 SAXReader reader = new SAXReader();
81 Document document = reader.read(aStream);
83 Element root = document.getRootElement();
84 List<UrlConfig> urlConfigs = parseUrlConfigs(root);
85 List<PageTypeConfig> pageTypeConfigs = parsePageTypeConfigs(root);
86 return new ConfigurationImpl(urlConfigs, pageTypeConfigs);
87 } catch (DocumentException e) {
88 throw new RuntimeException("Problem parsing config file", e);
93 * Parses the URL-based configuration.
94 * @param aRoot Root of the configuration file document.
95 * @return List of URL-based configurations.
97 private List<UrlConfig> parseUrlConfigs(Element aRoot) {
98 List<UrlConfig> configs = new ArrayList<UrlConfig>();
99 for (Iterator i = aRoot.elementIterator(ELEM_URL); i.hasNext();) {
100 Element url = (Element) i.next();
101 UrlConfig config = parseUrlConfig(url);
108 * Parses the page type based configurations.
109 * @param aRoot Root of the configuration file document.
110 * @return LIst of page type based configurations.
112 private List<PageTypeConfig> parsePageTypeConfigs(Element aRoot) {
113 List<PageTypeConfig> configs = new ArrayList<PageTypeConfig>();
114 for (Iterator i = aRoot.elementIterator(ELEM_TYPE); i.hasNext();) {
115 Element url = (Element) i.next();
116 PageTypeConfig config = parsePageTypeConfig(url);
123 * Parses a URL-based configuration.
124 * @param aUrlElem Configuration element.
125 * @return Configuration.
127 private UrlConfig parseUrlConfig(Element aUrlElem) {
128 String pattern = aUrlElem.elementText(ELEM_PATTERN);
129 PageRequest request = parseRequestConfig(aUrlElem);
130 return new UrlConfig(pattern, request);
134 * Parses a page type based configuration.
135 * @param aTypeElem Configuration element.
136 * @return Configuration.
138 private PageTypeConfig parsePageTypeConfig(Element aTypeElem) {
139 String pattern = aTypeElem.elementText(ELEM_PATTERN);
140 PageRequest request = parseRequestConfig(aTypeElem);
141 return new PageTypeConfig(pattern, request);
145 * Parses a request configuration describing how to execute requests.
146 * @param aElem Configuration element.
147 * @return Page request.
149 private PageRequest parseRequestConfig(Element aElem) {
150 String method = aElem.elementText(ELEM_METHOD);
151 String xslt = aElem.elementText(ELEM_XSLT);
152 List<NameValuePair> params = new ArrayList<NameValuePair>();
153 for (Iterator i = aElem.elementIterator(ELEM_PARAM); i.hasNext();) {
154 Element paramElem = (Element) i.next();
155 NameValuePair param = parseParameter(paramElem);
159 NameValuePair[] paramsArray = params.toArray(new NameValuePair[0]);
161 if (METHOD_POST.equals(method)) {
162 request = new PostPageRequest(MAX_TRIES, MAX_DELAY, paramsArray,
164 } else if (METHOD_GET.equals(method) || method == null) {
165 request = new GetPageRequest(MAX_TRIES, MAX_DELAY, paramsArray,
168 throw new RuntimeException("Unknown request method '" + method
169 + "'. Only " + METHOD_GET + " and " + METHOD_POST
176 * Parses a parameter definition.
177 * @param aParam Parameter.
178 * @return Name value pair describing a parameter.
180 private NameValuePair parseParameter(Element aParam) {
181 String name = aParam.attributeValue(AT_NAME);
182 String value = aParam.attributeValue(AT_VALUE);
183 return new NameValuePair(name, value);