(no commit message)

[utils] / crawler / kissweb / src / org / wamblee / crawler / kiss / scheduling / CrawlerSchedule.java

diff --git a/crawler/kissweb/src/org/wamblee/crawler/kiss/scheduling/CrawlerSchedule.java b/crawler/kissweb/src/org/wamblee/crawler/kiss/scheduling/CrawlerSchedule.java
index 09ada84da8d4f9bf0411ef1b6cb228a252ce4f27..5121f92058df76a437ca159e595c1b0e1f089866 100644 (file)
--- a/crawler/kissweb/src/org/wamblee/crawler/kiss/scheduling/CrawlerSchedule.java
+++ b/crawler/kissweb/src/org/wamblee/crawler/kiss/scheduling/CrawlerSchedule.java
@@ -20,6 +20,9 @@ import java.io.Serializable;
 import java.util.Calendar;
 import java.util.Date;
 
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+
 /**
  * This class encapsulates the logic for deciding whether to
  * run the crawler. This provides the mechanism to keep the 
@@ -29,22 +32,32 @@ import java.util.Date;
  */
 public class CrawlerSchedule implements Serializable {
     
+    private static final Log LOG = LogFactory.getLog(CrawlerSchedule.class);
+    
     private CrawlerExecutor _crawler; 
     private Date _lastExecuted;
-    private Exception _lastResult;
+    private boolean _lastResult;
+    private Exception _lastException;
     private int _hourMin; 
     private int _hourMax;
     
     /**
      * Constructs the scheduler.
-     * @param aCrawler The interface by which the crawler is executed.
+     * The crawler will run if it is triggered in the range between the minimum (included)
+     * and maximum (included) hour of the day if either
+     * <ul>
+     *   <li>it is triggered for the first time on the current day.</li>
+     *   <li>an earlier crawling attempt on the same day failed. </li>
+     * </ul>
+     * @param aCrawler The interface through which the crawler is executed.
      * @param aHourMin The crawler may only run if hour &gt;= <code>aHourMin</code>
      * @param aHourMax The crawler may only run if hour &lt;= <code>aHourMax</code>
      */
     public CrawlerSchedule(CrawlerExecutor aCrawler, int aHourMin, int aHourMax) { 
         _crawler = aCrawler; 
-        _lastExecuted = null; 
-        _lastResult = null; 
+        _lastExecuted = new Date(); 
+        _lastResult = true; // the crawler will automatically run the next day. 
+        _lastException = null; 
         _hourMin = aHourMin;
         _hourMax = aHourMax;
     }
@@ -55,16 +68,20 @@ public class CrawlerSchedule implements Serializable {
      * @param aDate Time at which we are executing now. 
      */
     public void execute(Date aDate) { 
+        
         if (mustExecute(aDate)) { 
-            try { 
-                _lastResult = null; 
+            LOG.info("Executing crawler at " + aDate);
+            try {  
                 _crawler.execute(aDate);
+                _lastResult = true; 
+                _lastException = null; 
             } catch (Exception e) {
-                _lastResult = e; 
+                _lastResult = false; 
+                _lastException = e; 
             } finally { 
                 _lastExecuted = aDate;
             }
-        }
+        } 
     }
     
     /**
@@ -75,13 +92,25 @@ public class CrawlerSchedule implements Serializable {
         return _lastExecuted; 
     }
     
+    public void setLastExecuted(Date aDate) { 
+        _lastExecuted = aDate;
+    }
+    
     /**
      * Gets the result of the last execution. 
+     * @return True iff last execution was a success. 
+     */
+    public boolean getLastResult() { 
+        return _lastResult; 
+    }
+    
+    /**
+     * Gets the exception thrown by the last execution. 
      * @return null if the last execution was successful or an exception 
      *   otherwise.
      */
-    public Exception getLastResult() { 
-        return _lastResult; 
+    public Exception getLastException() { 
+        return _lastException; 
     }
     
     /**
@@ -90,22 +119,51 @@ public class CrawlerSchedule implements Serializable {
      * @return True iff the crawler must be run. 
      */
     private boolean mustExecute(Date aDate) {
+        if ( _lastExecuted == null ) {
+            return false; // crawler must be started manually at least once after deployment.
+        }
         Calendar calendar = Calendar.getInstance();
         calendar.setTime(aDate);
         int hour = calendar.get(Calendar.HOUR_OF_DAY);
         if ( hour < _hourMin ) {
-            return false;
+            return false; 
         }
         if (hour > _hourMax ) { 
             return false; 
         }
-        if ( hour == _hourMin ) { 
+     
+        if ( !lastExecutionWasOnSameDay(aDate)) { 
             return true; // First execution of today. 
-        }
-        if ( _lastResult != null ) { 
+        } 
+        // last execution was on the same day.
+        if ( !_lastResult ) { 
             return true; // last execution of today was unsuccessful, retry. 
         }
-        
         return false; // already run successfully today.
     }
+    
+    /**
+     * Determines if the last execution was on the same day. 
+     * @param aDate Current time. 
+     * @return True iff last execution was on the same day. 
+     */
+    private boolean lastExecutionWasOnSameDay(Date aDate) {
+        if ( _lastExecuted == null ) { 
+            return false;
+        }
+        int curDay = getDayOfYear(aDate);
+        int lastDay = getDayOfYear(_lastExecuted);
+        return curDay == lastDay; // check can be invalid only if scheduling interval is one year,
+                                  // which is ridiculous.
+    }
+
+    /**
+     * Gets the day of the year
+     * @param aDate Date to compute day for. 
+     */
+    private int getDayOfYear(Date aDate) {
+        Calendar calendar = Calendar.getInstance();
+        calendar.setTime(aDate);
+        return calendar.get(Calendar.DAY_OF_YEAR);
+    }
 }