wamblee.org Git - utils/blob - crawler/kiss/src/org/wamblee/crawler/kiss/main/ProgramConfigurationParser.java

   1 /*
   2  * Copyright 2005 the original author or authors.
   3  *
   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
   7  *
   8  *      http://www.apache.org/licenses/LICENSE-2.0
   9  *
  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.
  15  */
  16
  17 package org.wamblee.crawler.kiss.main;
  18
  19 import java.io.InputStream;
  20 import java.util.ArrayList;
  21 import java.util.Iterator;
  22 import java.util.List;
  23
  24 import org.dom4j.Attribute;
  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.conditions.AndCondition;
  30 import org.wamblee.conditions.Condition;
  31 import org.wamblee.conditions.PropertyRegexCondition;
  32 import org.wamblee.crawler.kiss.guide.Program;
  33 import org.wamblee.xml.XslTransformer;
  34
  35 /**
  36  * Parse the configuration of desired programs.
  37  */
  38 class ProgramConfigurationParser {
  39     private static final int DEFAULT_SMTP_PORT = 25;
  40
  41     private static final int DEFAULT_PRIORITY = 1;
  42
  43     // Formatting configuration.
  44     private static final String ELEM_FORMAT = "format";
  45
  46     private static final String ELEM_TEXT = "text";
  47
  48     private static final String ELEM_HTML = "html";
  49
  50     // Configuration of interesting programs.
  51
  52     private static final String ELEM_PROGRAM = "program";
  53
  54     private static final String ELEM_PRIORITY = "priority";
  55
  56     private static final String ELEM_PATTERN = "match";
  57
  58     private static final String ELEM_ACTION = "action";
  59
  60     private static final String ELEM_CATEGORY = "category";
  61
  62     private static final String ACTION_NOTIFY = "notify";
  63
  64     private List<ProgramFilter> _filters;
  65
  66     private XslTransformer _transformer;
  67
  68     ProgramConfigurationParser(XslTransformer aTransformer) {
  69         _filters = null;
  70         _transformer = aTransformer;
  71     }
  72
  73     /**
  74      * Parses the condition used to match the desired programs.
  75      *
  76      * @param aStream
  77      *            Input stream to parse from.
  78      * @return Condition.
  79      */
  80     void parse(InputStream aStream) {
  81         List<ProgramFilter> filters = new ArrayList<ProgramFilter>();
  82         try {
  83             SAXReader reader = new SAXReader();
  84             Document document = reader.read(aStream);
  85
  86             Element root = document.getRootElement();
  87
  88             for (Iterator i = root.elementIterator(ELEM_PROGRAM); i.hasNext();) {
  89                 Element program = (Element) i.next();
  90
  91                 Element categoryElem = program.element(ELEM_CATEGORY);
  92                 String category = "";
  93                 if (categoryElem != null) {
  94                     category = categoryElem.getText().trim();
  95                 }
  96
  97                 Element actionElem = program.element(ELEM_ACTION);
  98                 int priority = DEFAULT_PRIORITY;
  99                 String priorityString = program.elementTextTrim(ELEM_PRIORITY);
 100                 if ( priorityString != null ) {
 101                     priority = Integer.valueOf(priorityString);
 102                 }
 103                 ProgramAction action = new RecordProgramAction(priority);
 104                 if (actionElem != null) {
 105                     if (actionElem.getText().equals(ACTION_NOTIFY)) {
 106                         action = new InterestingProgramAction(category);
 107                     }
 108                 }
 109
 110                 List<Condition<Program>> regexConditions = new ArrayList<Condition<Program>>();
 111                 for (Iterator j = program.elementIterator(ELEM_PATTERN); j
 112                         .hasNext();) {
 113                     Element patternElem = (Element) j.next();
 114                     String fieldName = "name";
 115                     Attribute fieldAttribute = patternElem.attribute("field");
 116                     if (fieldAttribute != null) {
 117                         fieldName = fieldAttribute.getText();
 118                     }
 119                     String pattern = ".*(" + patternElem.getText() + ").*";
 120                     regexConditions.add(new PropertyRegexCondition<Program>(
 121                             fieldName, pattern, true));
 122                 }
 123                 Condition<Program> condition = new AndCondition<Program>(
 124                         regexConditions);
 125                 filters.add(new ProgramFilter(condition, action));
 126             }
 127             _filters = filters;
 128         } catch (DocumentException e) {
 129             throw new RuntimeException("Error parsing program configuraiton", e);
 130         }
 131     }
 132
 133     /**
 134      * Returns the list of program filters.
 135      *
 136      * @return Filter list.
 137      */
 138     public List<ProgramFilter> getFilters() {
 139         return _filters;
 140     }
 141 }