You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
247 lines
8.2 KiB
247 lines
8.2 KiB
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">
|
|
<html>
|
|
<head>
|
|
<title>YaCy '#[clientname]#': Index Import</title>
|
|
#%env/templates/metas.template%#
|
|
<meta http-equiv="REFRESH" content="30">
|
|
</head>
|
|
<body marginheight="0" marginwidth="0" leftmargin="0" topmargin="0">
|
|
#%env/templates/header.template%#
|
|
#%env/templates/submenuIndexControl.template%#
|
|
<br><br>
|
|
<h2>Index DB Import</h2>
|
|
<p>The local index currently consists of (at least) #[wcount]# reverse word indexes and #[ucount]# URL references.</p>
|
|
<hr>
|
|
#(error)#<!-- 0 -->
|
|
::<!-- 1 -->
|
|
<p><font color="red"><b>#[error_msg]#</b></font></p>
|
|
::<!-- 2 -->
|
|
<p><font color="red"><b>Import Job with the same path already started.</b></font></p>
|
|
::<!-- 3 -->
|
|
<p><font color="red"><b>#[error_msg]#</b></font></p>
|
|
<p><font color="red"><code>#[error_stackTrace]#</code></font></p>
|
|
#(/error)#
|
|
<form action="IndexImport_p.html" method="post" enctype="multipart/form-data">
|
|
<h3>Starting new Job</h3>
|
|
<table>
|
|
<tr><td>
|
|
<table border="0" cellpadding="2" cellspacing="1">
|
|
<tr class="TableCellLight">
|
|
<td>Import Type:</td>
|
|
<td title="the path to the database that should be imported"><select name="importType" size="1">
|
|
<option value="plasmaDB">PLASMA DB Import</option>
|
|
<option value="assortment">Assortment File Import</option>
|
|
<option value="NURL">Crawling Queue Import</option>
|
|
</select>
|
|
</td>
|
|
<td title="the cache size that should be used for the import db">Cache Size</td>
|
|
<td><select name="cacheSize" size="1">
|
|
<option value="2097152">2 MB</option>
|
|
<option value="4194304">4 MB</option>
|
|
<option value="8388608" selected="selected">8 MB</option>
|
|
<option value="16777216">16 MB</option>
|
|
<option value="33554432">32 MB</option>
|
|
<option value="67108864">64 MB</option>
|
|
<option value="134217728">128 MB</option>
|
|
</select>
|
|
</td>
|
|
<td><a href="#usage">Usage Examples</a></td>
|
|
</tr>
|
|
<tr class="TableCellLight">
|
|
<td title="Path to the PLASMADB directory of the foreign peer">Import Path:</td>
|
|
<td colspan="3"><input name="importPath" type="text" size="50" value=""></td>
|
|
<td><input type="submit" name="startIndexDbImport" value="Start Import"></td>
|
|
</tr>
|
|
</table>
|
|
</td>
|
|
<td><font color="red"><b>Attention:</b><br>Always do a backup of your source and destination database before starting to use this import function.</font>
|
|
</td>
|
|
</tr>
|
|
</table>
|
|
</form>
|
|
|
|
<hr>
|
|
<h3>Currently running jobs</h3>
|
|
<p>
|
|
<table border="0" cellpadding="2" cellspacing="1">
|
|
<tr class="TableHeader" valign="bottom">
|
|
<td class="small">Job Type</td>
|
|
<td class="small" width="150">Job Name</td>
|
|
<td class="small" >Status</td>
|
|
<td class="small" >%</td>
|
|
<td class="small" >Elapsed<br>Time</td>
|
|
<td class="small" >Time<br>Left</td>
|
|
<td class="small" >Import Status</td>
|
|
<td class="small" >Abort Import</td>
|
|
<td class="small" >Pause Import</td>
|
|
</tr>
|
|
#{running.jobs}#
|
|
<form action="IndexImport_p.html" method="post" enctype="multipart/form-data">
|
|
<input type="hidden" name="jobNr" value="#[job_nr]#">
|
|
<tr class="TableCellLight">
|
|
<td class="small">#[type]#</td>
|
|
<td class="small" title="#[fullName]#">#[shortName]#</td>
|
|
<td class="small"><font color="#(runningStatus)#red::green::blue#(/runningStatus)#">#(runningStatus)#Finished::Running::Paused#(/runningStatus)#</font></td>
|
|
<td class="small" align="right">#[percent]#</td>
|
|
<td class="small" align="right">#[elapsed]#</td>
|
|
<td class="small" align="right">#[estimated]#</td>
|
|
<td class="small" align="left"><tt>#[status]#</tt></td>
|
|
<td class="small">
|
|
#(stopped)#::
|
|
<input type="submit" name="stopIndexDbImport" value="Abort">
|
|
#(/stopped)#
|
|
</td>
|
|
<td class="small">
|
|
#(paused)#
|
|
<input type="submit" name="pauseIndexDbImport" value="Pause">
|
|
::
|
|
<input type="submit" name="continueIndexDbImport" value="Continue">
|
|
#(/paused)#
|
|
</td>
|
|
</tr>
|
|
</form>
|
|
#{/running.jobs}#
|
|
</table>
|
|
|
|
|
|
<hr>
|
|
<form action="IndexImport_p.html" method="post" enctype="multipart/form-data">
|
|
<h3>Finished jobs</h3>
|
|
<p>
|
|
<table border="0" cellpadding="2" cellspacing="1">
|
|
<tr class="TableHeader" valign="bottom">
|
|
<td class="small">Job Type</td>
|
|
<td class="small" width="150">Path</td>
|
|
<td class="small" >Status</td>
|
|
<td class="small" >%</td>
|
|
<td class="small" >Elapsed<br>Time</td>
|
|
<td class="small" >Import Status</td>
|
|
</tr>
|
|
#{finished.jobs}#
|
|
<tr class="TableCellLight">
|
|
<td class="small">#[type]#</td>
|
|
<td class="small" title="#[fullName]#">#[shortName]#</td>
|
|
<td class="small"><font color="#(runningStatus)#green::red::blue#(/runningStatus)#">#(runningStatus)#Finished::<b>Error:</b> #[errorMsg]#::Paused#(/runningStatus)#</font></td>
|
|
<td class="small" align="right">#[percent]#</td>
|
|
<td class="small" align="right">#[elapsed]#</td>
|
|
<td class="small" align="left"><tt>#[status]#</tt></td>
|
|
</tr>
|
|
#{/finished.jobs}#
|
|
</table>
|
|
<input type="submit" name="clearFinishedJobList" value="Clear List">
|
|
</form>
|
|
<p><font size="-3"><i>Last Refresh:</i> #[date]#</font></p>
|
|
<hr>
|
|
<p><div id="usage"><h2>Usage Examples:</h2></div></p>
|
|
|
|
<p><h3>Plasma DB Import:</h3></p>
|
|
<p>
|
|
<b>Example Path:</b> <tt>E:\PLASMADB\</tt><br>
|
|
</p>
|
|
<p>
|
|
<b>Requirements:</b><br>
|
|
You need to have at least the following directories and files in this path:
|
|
<table border="1" cellpadding="2" cellspacing="1">
|
|
<tr style="background-color: lightgrey">
|
|
<td>Name</td>
|
|
<td>Type</td>
|
|
<td>Writeable</td>
|
|
<td>Description</td>
|
|
</tr>
|
|
<tr>
|
|
<td><tt>urlHash.db</tt></td>
|
|
<td>File</td>
|
|
<td>No</td>
|
|
<td>The LoadedURL Database containing all loaded and indexed URLs</td>
|
|
</tr>
|
|
<tr>
|
|
<td><tt>ACLUSTER</tt></td>
|
|
<td>Directory</td>
|
|
<td>Yes</td>
|
|
<td>The assortment directory containing parts of the word index.</td>
|
|
</tr>
|
|
<tr>
|
|
<td><tt>WORDS</tt></td>
|
|
<td>Directory</td>
|
|
<td>Yes</td>
|
|
<td>The words directory containing parts of the word index.</td>
|
|
</tr>
|
|
</table>
|
|
</p>
|
|
|
|
|
|
<p><h3>Assortment Import:</h3></p>
|
|
<p>
|
|
<b>Example Path:</b> <tt>E:\PLASMADB\ACLUSTER\indexAssortment001.db</tt>
|
|
</p>
|
|
<p>
|
|
<b>Requirements:</b><br>
|
|
You need to have at least the following directories and files in this path:
|
|
<table border="1" cellpadding="2" cellspacing="1">
|
|
<tr style="background-color: lightgrey">
|
|
<td>Name</td>
|
|
<td>Type</td>
|
|
<td>Writeable</td>
|
|
<td>Description</td>
|
|
</tr>
|
|
<tr>
|
|
<td><tt>indexAssortment001.db</tt></td>
|
|
<td>File</td>
|
|
<td>No</td>
|
|
<td>The assortment file that should be imported.<br>
|
|
<b>Attention:</b> The assortment file must have the postfix "[0-9]{3}\.db".
|
|
If you would like to import an assortment file from the <tt>PLASMADB\ACLUSTER\ABKP</tt>
|
|
you have to rename it first.</td>
|
|
</tr>
|
|
</table>
|
|
</p>
|
|
<p>
|
|
<b>Notes:</b><br>
|
|
Please note that the imported words are useless if the destination peer doesn't know
|
|
the URLs the imported words belongs to.
|
|
</p>
|
|
|
|
<p><h3>Crawling Queue Import:</h3></p>
|
|
<p>
|
|
<b>Example Path:</b> <tt>E:\PLASMADB\</tt>
|
|
</p>
|
|
<b>Requirements:</b><br>
|
|
You need to have at least the following directories and files in this path:
|
|
<table border="1" cellpadding="2" cellspacing="1">
|
|
<tr style="background-color: lightgrey">
|
|
<td>Name</td>
|
|
<td>Type</td>
|
|
<td>Writeable</td>
|
|
<td>Description</td>
|
|
</tr>
|
|
<tr>
|
|
<td><tt>crawlProfiles0.db</tt></td>
|
|
<td>File</td>
|
|
<td>No</td>
|
|
<td>Contains data about the crawljob an URL belongs to</td>
|
|
</tr>
|
|
<tr>
|
|
<td><tt>urlNotice1.db</tt></td>
|
|
<td>File</td>
|
|
<td>Yes</td>
|
|
<td>The crawling queue</td>
|
|
</tr>
|
|
<tr>
|
|
<td><tt>urlNoticeImage0.stack</tt></td>
|
|
<td rowspan="8">File</td>
|
|
<td rowspan="8">Yes</td>
|
|
<td rowspan="8">Various stack files that belong to the crawling queue</td>
|
|
</tr>
|
|
<tr><td><tt>urlNoticeImage0.stack</tt></td></tr>
|
|
<tr><td><tt>urlNoticeLimit0.stack</tt></td></tr>
|
|
<tr><td><tt>urlNoticeLocal0.stack</tt></td></tr>
|
|
<tr><td><tt>urlNoticeMovie0.stack</tt></td></tr>
|
|
<tr><td><tt>urlNoticeMusic0.stack</tt></td></tr>
|
|
<tr><td><tt>urlNoticeOverhang0.stack</tt></td></tr>
|
|
<tr><td><tt>urlNoticeRemote0.stack</tt></td></tr>
|
|
</table>
|
|
</p>
|
|
|
|
#%env/templates/footer.template%#
|
|
</body>
|
|
</html> |