logs.jsp

这是个爬虫和lucece相结合最好了

<%@include file="/include/handler.jsp"%>

<%@page import="org.archive.crawler.datamodel.CrawlOrder" %>
<%@page import="org.archive.crawler.settings.SettingsHandler" %>
<%@page import="org.archive.crawler.settings.XMLSettingsHandler" %>
<%@page import="org.archive.crawler.admin.CrawlJob" %>
<%@page import="org.archive.crawler.util.LogReader" %>
<%@page import="java.io.File" %>
<%@page import="org.archive.util.TextUtils" %>

<%
    /* Various settings with default values (where applicable) */
    String mode = request.getParameter("mode");
    String[] log = null;
    String logText = "";
    String logInfo = "";
    int linesToShow = 50;
    int iTime = -1;
    int linenumber = 1;
    String timestamp = null;
    String regexpr = null;
    boolean ln = false;
    boolean grep = false;
    boolean indent = false;
    int linesToSkip = 1;
        
    /* Which log to display */
    String fileName = request.getParameter("log");
    if(fileName == null || fileName.length() <= 0) {
        fileName = "crawl.log";
    }
    
    if(request.getParameter("linesToShow") != null && request.getParameter("linesToShow").length()>0 ){
        try{
            linesToShow = Integer.parseInt(request.getParameter("linesToShow"));
        } catch(java.lang.NumberFormatException e){
            linesToShow = 50;
        }
    }

    /* Location of logs */
    SettingsHandler settingsHandler = null;
    CrawlJob theJob = null;
    if(request.getParameter("job") != null && request.getParameter("job").length() > 0){
        //Get logs for specific job. This assumes that the logs for each job are stored in a unique location.
        theJob = handler.getJob(request.getParameter("job"));
    }else{
        if(handler.getCurrentJob() != null){
            // If no specific job then assume current one
            theJob = handler.getCurrentJob();
        } else if(handler.getCompletedJobs().size() > 0){
            // If no current job, use the latest completed job.
            theJob = (CrawlJob)handler.getCompletedJobs().get(handler.getCompletedJobs().size()-1);
        }
    }
    
    if(theJob != null) {
        // Got a valid crawl order, find it's logs
        if(mode != null && mode.equalsIgnoreCase("number")) {
            /* Get log by line number */
            try {
                linenumber = Integer.
                    parseInt(request.getParameter("linenumber"));
            }
            catch(Exception e){/*Ignore*/}
            log = LogReader.getFromSeries(theJob.getLogPath(fileName),
                linenumber, linesToShow);
        } else if(mode != null && mode.equalsIgnoreCase("time")) {
            /* View by timestamp */
            timestamp = request.getParameter("timestamp");
        
            if(timestamp == null || timestamp.length() < 1)
            {
                // No data
                logText = "No timestamp!";
            }    
            else
            {
                int timestampLinenumber = LogReader.
                    findFirstLineBeginningFromSeries(theJob.getLogPath(fileName),
                        timestamp);
                log =  LogReader.getFromSeries(theJob.getLogPath(fileName),
                    timestampLinenumber, linesToShow);
            }
        }
        else if(mode != null && mode.equalsIgnoreCase("regexpr"))
        {
            /* View by regexpr */
            try
            {
                linesToSkip = Integer.parseInt(request.getParameter("linesToSkip"));
            }
            catch(Exception e){/*Ignore*/}
            
            regexpr = request.getParameter("regexpr");
            
            if(regexpr == null)
            {
                logText = "No regular expression";
            }
            else
            {
                ln = request.getParameter("ln")!=null&&request.getParameter("ln").equalsIgnoreCase("true");
                grep = request.getParameter("grep")!=null&&request.getParameter("grep").equalsIgnoreCase("true");
                indent = request.getParameter("indent")!=null&&request.getParameter("indent").equalsIgnoreCase("true");
                
                if(grep){
                    regexpr = ".*" + regexpr + ".*";
                }
                
                if(indent) {
                    log = LogReader.
                        getByRegExprFromSeries(theJob.getLogPath(fileName),
                            regexpr, " ", ln,linesToSkip-1, linesToShow);
                } else {
                    log = LogReader.
                        getByRegExprFromSeries(theJob.getLogPath(fileName),
                            regexpr, 0, ln,linesToSkip-1, linesToShow);
                }
            }
        } else {
            /* View by tail (default) */
            mode = "tail";
    
            try
            {
                iTime = Integer.parseInt(request.getParameter("time"));
            }
            catch(Exception e){/* Ignore - default value will do */}
            log = LogReader.tail(theJob.getLogPath(fileName), linesToShow);
        }
    } 
    else 
    {
        mode = "tail";
        log = new String[]{"Invalid or missing crawl order",""};
    }
    
    if(log != null && log.length>=2){
        logText = log[0];
        logInfo = log[1];
    } else {
        logText = "";
        logInfo = "";
    }
    
    
    String title = "View logs";
    int tab = 3;
    
%>

<%@include file="/include/head.jsp"%>
    <% if(iTime>0){ %>
        <meta http-equiv=Refresh content="<%=iTime%> URL=logs.jsp?time=<%=iTime%>&log=<%=fileName%>&linesToShow=<%=linesToShow%>">
    <% } %>
    
    <% 
        if(theJob == null){
            out.println("<br /><b>No job selected/available</b>");
            return;
        } 
    %>
    <script type="text/javascript">
        function viewLog(log)
        {
            document.frmLogs.log.value = log;
            document.frmLogs.submit();
        }
        
        function changeMode(mode)
        {
            document.frmLogs.mode.value = mode;
            document.frmLogs.submit();
        }
    </script>

    <form method="get" action="logs.jsp" name="frmLogs">
        <input type="hidden" name="job" value="<%=theJob.getUID()%>">
        <table border="0" cellspacing="0" cellpadding="0">
            <tr>
                <td height="3"></td>
            </tr>
            <tr>
                <td valign="top" width="210">
                    <table border="0" cellspacing="0" cellpadding="0">
                        <tr>
                            <td width="50" align="right" valign="top">
                                &nbsp;<b>View:</b>&nbsp;
                            </td>
                            <td align="left" valign="top" width="160">
                                <a href="javascript:viewLog('crawl.log')" <%=fileName.equalsIgnoreCase("crawl.log")?"style='text-decoration: none; color: #000000'":""%>>crawl.log</a><br>
                                <a href="javascript:viewLog('local-errors.log')" <%=fileName.equalsIgnoreCase("local-errors.log")?"style='text-decoration: none; color: #000000'":""%>>local-errors.log</a><br>
                                <a href="javascript:viewLog('progress-statistics.log')" <%=fileName.equalsIgnoreCase("progress-statistics.log")?"style='text-decoration: none; color: #000000'":""%>>progress-statistics.log</a><br>
                                <a href="javascript:viewLog('runtime-errors.log')" <%=fileName.equalsIgnoreCase("runtime-errors.log")?"style='text-decoration: none; color: #000000'":""%>>runtime-errors.log</a><br>
                                <a href="javascript:viewLog('uri-errors.log')" <%=fileName.equalsIgnoreCase("uri-errors.log")?"style='text-decoration: none; color: #000000'":""%>>uri-errors.log</a><br>