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
  34 /**
  35  * Parse the configuration of desired programs.
  36  */
  37 class ProgramConfigurationParser {
  38     private static final int DEFAULT_PRIORITY = 1;
  39
  40     // Configuration of interesting programs.
  41
  42     private static final String ELEM_PROGRAM = "program";
  43
  44     private static final String ELEM_PRIORITY = "priority";
  45
  46     private static final String ELEM_PATTERN = "match";
  47
  48     private static final String ELEM_ACTION = "action";
  49
  50     private static final String ELEM_CATEGORY = "category";
  51
  52     private static final String ACTION_NOTIFY = "notify";
  53
  54     private List<ProgramFilter> _filters;
  55
  56     ProgramConfigurationParser() {
  57         _filters = null;
  58     }
  59
  60     /**
  61      * Parses the condition used to match the desired programs.
  62      *
  63      * @param aStream
  64      *            Input stream to parse from.
  65      * @return Condition.
  66      */
  67     void parse(InputStream aStream) {
  68         List<ProgramFilter> filters = new ArrayList<ProgramFilter>();
  69         try {
  70             SAXReader reader = new SAXReader();
  71             Document document = reader.read(aStream);
  72
  73             Element root = document.getRootElement();
  74
  75             for (Iterator i = root.elementIterator(ELEM_PROGRAM); i.hasNext();) {
  76                 Element program = (Element) i.next();
  77
  78                 Element categoryElem = program.element(ELEM_CATEGORY);
  79                 String category = "";
  80                 if (categoryElem != null) {
  81                     category = categoryElem.getText().trim();
  82                 }
  83
  84                 Element actionElem = program.element(ELEM_ACTION);
  85                 int priority = DEFAULT_PRIORITY;
  86                 String priorityString = program.elementTextTrim(ELEM_PRIORITY);
  87                 if ( priorityString != null ) {
  88                     priority = Integer.valueOf(priorityString);
  89                 }
  90                 ProgramAction action = new RecordProgramAction(priority);
  91                 if (actionElem != null) {
  92                     if (actionElem.getText().equals(ACTION_NOTIFY)) {
  93                         action = new InterestingProgramAction(category);
  94                     }
  95                 }
  96
  97                 List<Condition<Program>> regexConditions = new ArrayList<Condition<Program>>();
  98                 for (Iterator j = program.elementIterator(ELEM_PATTERN); j
  99                         .hasNext();) {
 100                     Element patternElem = (Element) j.next();
 101                     String fieldName = "name";
 102                     Attribute fieldAttribute = patternElem.attribute("field");
 103                     if (fieldAttribute != null) {
 104                         fieldName = fieldAttribute.getText();
 105                     }
 106                     String pattern = ".*(" + patternElem.getText() + ").*";
 107                     regexConditions.add(new PropertyRegexCondition<Program>(
 108                             fieldName, pattern, true));
 109                 }
 110                 Condition<Program> condition = new AndCondition<Program>(
 111                         regexConditions);
 112                 filters.add(new ProgramFilter(condition, action));
 113             }
 114             _filters = filters;
 115         } catch (DocumentException e) {
 116             throw new RuntimeException("Error parsing program configuraiton", e);
 117         }
 118     }
 119
 120     /**
 121      * Returns the list of program filters.
 122      *
 123      * @return Filter list.
 124      */
 125     public List<ProgramFilter> getFilters() {
 126         return _filters;
 127     }
 128 }