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.
167 lines
9.1 KiB
167 lines
9.1 KiB
<!DOCTYPE html>
|
|
<html lang="en">
|
|
<head>
|
|
<title>YaCy '#[clientname]#': Field Re-Indexing</title>
|
|
#%env/templates/metas.template%#
|
|
<script type="text/javascript" src="js/IndexReIndexMonitor.js"></script>
|
|
</head>
|
|
<body id="IndexReindexMonitor">
|
|
#%env/templates/header.template%#
|
|
#%env/templates/submenuIndexControl.template%#
|
|
<div class="container-fluid">
|
|
<div class="row">
|
|
<h2>Field Re-Indexing</h2>
|
|
<p>In case that an index schema of the embedded/local index has changed, all documents with missing field entries can be indexed again with a reindex job.</p>
|
|
<form action="IndexReIndexMonitor_p.html" method="post" enctype="multipart/form-data" accept-charset="UTF-8" class="col-md-10 col-lg-6">
|
|
<input type="hidden" name="transactionToken" value="#[transactionToken]#"/>
|
|
<fieldset>
|
|
<dl>
|
|
<dt>Documents in current queue</dt>
|
|
<dd>#[querysize]#</dd>
|
|
<dt>Documents processed</dt>
|
|
<dd>#[docsprocessed]#</dd>
|
|
<dt>current select query </dt>
|
|
<dd>#[currentselectquery]#</dd>
|
|
</dl>
|
|
#(reindexjobrunning)#::<input type="submit" value="refresh page" class="btn btn-success"/>#(/reindexjobrunning)#
|
|
#(reindexjobrunning)#
|
|
<input type="submit" name="reindexnow" value="start reindex job now" class="btn btn-primary"/>
|
|
::<input type="submit" name="stopreindex" value="stop reindexing" class="btn btn-danger"/>
|
|
#(/reindexjobrunning)#
|
|
<p class="info">#[infomessage]#</p>
|
|
</fieldset>
|
|
#(reindexjobrunning)#::
|
|
<fieldset><legend>Remaining field list</legend>
|
|
<p>reindex documents containing these fields: </p>
|
|
<table>
|
|
<tr><th>Field</th><th>count</th></tr>
|
|
#{fieldlist}#
|
|
<tr>
|
|
<td>#[fieldname]#</td> <td align="right">#[fieldscore]#</td>
|
|
</tr>
|
|
#{/fieldlist}#
|
|
</table>
|
|
</fieldset>
|
|
#(/reindexjobrunning)#
|
|
</form>
|
|
</div>
|
|
<div class="row">
|
|
<h2>Re-Crawl Index Documents</h2>
|
|
<p>Searches the local index and selects documents to add to the crawler (recrawl the document).
|
|
This runs transparent as background job. Documents are added to the crawler only if no other crawls are active
|
|
and are added in small chunks.</p>
|
|
<form action="IndexReIndexMonitor_p.html?setup=recrawljob" method="post" enctype="multipart/form-data" accept-charset="UTF-8" class="col-md-10 col-lg-6">
|
|
<input type="hidden" name="transactionToken" value="#[transactionToken]#" />
|
|
|
|
#(recrawljobrunning)#
|
|
<fieldset id="newJobFieldset">
|
|
#(error)#::
|
|
<div class="alert alert-danger" role="alert">Re-crawl works only with an embedded local Solr index!</div>
|
|
#(/error)#
|
|
<div class="form-group">
|
|
<label>Solr query <input type="text" name="recrawlquerytext" size="40" value="#[recrawlquerytext]#" /></label>
|
|
<input type="submit" name="simulateRecrawl" value="Simulate" class="btn btn-default" title="Check only how many documents would be selected for recrawl"/>
|
|
</div>
|
|
#(simulationResult)#
|
|
::
|
|
<div class="alert alert-success" role="alert">#[docCount]# #(showSelectLink)#document(s)::<a href="#[browseSelectedUrl]#" target="recrawlCandidates" title="Browse metadata of the #[rows]# first selected documents">document(s)</a>#(/showSelectLink)# selected for recrawl.</div>
|
|
::
|
|
<div class="alert alert-danger" role="alert">An error occurred when trying to run the selection query.</div>
|
|
::
|
|
<div class="alert alert-danger" role="alert">The Solr index is not connected. Please restart your peer.</div>
|
|
#(/simulationResult)#
|
|
<div class="form-group">
|
|
<label>Include failed URLs <input type="checkbox" name="includefailedurls" #(includefailedurls)#::checked="checked"#(/includefailedurls)# /></label>
|
|
</div>
|
|
<div class="form-group">
|
|
<label>Delete URLs <input type="checkbox" name="deleteOnRecrawl" #(deleteOnRecrawl)#::checked="checked"#(/deleteOnRecrawl)# /></label>
|
|
</div>
|
|
<input type="submit" name="recrawlDefaults" value="Set defaults" class="btn btn-default" title="Reset to default values"/>
|
|
<input type="submit" name="recrawlnow" value="start recrawl job now" class="btn btn-primary"/>
|
|
to re-crawl documents selected with the given query.
|
|
</fieldset>
|
|
::
|
|
<fieldset id="updateJobFieldset"><legend>Re-Crawl Query Details</legend>
|
|
<dl>
|
|
<dt>Documents to process</dt>
|
|
<dd id="urlsToRecrawlCount">#[docCount]#</dd>
|
|
<dt>Current Query</dt>
|
|
<dd>#[recrawlquerytext]#</dd>
|
|
</dl>
|
|
<div class="form-group">
|
|
<label>Edit Solr Query <input type="text" name="recrawlquerytext" size="40" value="#[recrawlquerytext]#" /></label><input type="submit" name="updquery" value="update" class="btn btn-sm btn-default"/>
|
|
</div>
|
|
<div class="form-group">
|
|
<label>Include failed urls <input type="checkbox" name="includefailedurls" onchange="this.form.submit()" #(includefailedurls)#::checked="checked"#(/includefailedurls)# /></label>
|
|
</div>
|
|
<div class="form-group">
|
|
<label>Delete urls <input type="checkbox" name="deleteOnRecrawl" onchange="this.form.submit()" #(deleteOnRecrawl)#::checked="checked"#(/deleteOnRecrawl)# /></label>
|
|
</div>
|
|
<input type="submit" name="stoprecrawl" value="stop recrawl job" class="btn btn-danger"/>
|
|
</fieldset>
|
|
#(/recrawljobrunning)#
|
|
</form>
|
|
|
|
#(recrawlReport)#::
|
|
<div class="col-md-10 col-lg-6">
|
|
<div class="panel panel-info">
|
|
<div class="panel-heading">
|
|
<h3 class="panel-title">#(jobStatus)#::::Last #(/jobStatus)#Re-Crawl job report
|
|
<span id="refreshingIcon" class="glyphicon glyphicon-transfer hidden" title="Automatically refreshing"></span>
|
|
<span id="refreshFailureIcon" class="glyphicon glyphicon-warning-sign warning hidden" title="An error occurred while trying to refresh automatically"></span>
|
|
</h3>
|
|
</div>
|
|
<div class="panel-body">
|
|
<div id="earlyRecrawlTermination" class="alert alert-danger #(error)#hidden::#(/error)#" role="alert">The job terminated early due to an error when requesting the Solr index.</div>
|
|
<table class="table">
|
|
<tbody>
|
|
<tr>
|
|
<th scope="row">Status</th>
|
|
<td id="recrawlStatus"
|
|
arial-live="polite"
|
|
data-status="#(jobStatus)#0::1::2#(/jobStatus)#"
|
|
data-status0="Running" data-status1="Shutdown in progress"
|
|
data-status2="Terminated">#(jobStatus)#Running::Shutdown in progress::Terminated#(/jobStatus)#</td>
|
|
</tr>
|
|
<tr>
|
|
<th scope="row">Query</th>
|
|
<td id="recrawlQueryText">#[recrawlquerytext]#</td>
|
|
</tr>
|
|
<tr>
|
|
<th scope="row">Start time</th>
|
|
<td id="recrawlStartTime">#[startTime]#</td>
|
|
</tr>
|
|
<tr>
|
|
<th scope="row">End time</th>
|
|
<td id="recrawlEndTime">#[endTime]#</td>
|
|
</tr>
|
|
<tr>
|
|
<th scope="row" title="URLs added to the crawler queue for recrawl">Recrawled URLs</th>
|
|
<td id="recrawledUrlsCount" aria-live="polite">#[recrawledUrlsCount]#</td>
|
|
</tr>
|
|
<tr>
|
|
<th scope="row" title="URLs rejected for some reason by the crawl stacker or the crawler queue. Please check the logs for more details.">Rejected URLs</th>
|
|
<td id="rejectedUrlsCount">#[rejectedUrlsCount]#</td>
|
|
</tr>
|
|
<tr>
|
|
<th scope="row">Malformed URLs</th>
|
|
<td id="malformedUrlsCount" title="#[malformedUrlsDeletedCount]# deleted from the index">#[malformedUrlsCount]#</td>
|
|
</tr>
|
|
</tbody>
|
|
</table>
|
|
#(jobStatus)#
|
|
<a id="recrawlRefreshBtn" class="btn btn-default" href="IndexReIndexMonitor_p.html" role="button"><span class="glyphicon glyphicon-refresh"></span> Refresh</a>
|
|
::
|
|
<a id="recrawlRefreshBtn" class="btn btn-default" href="IndexReIndexMonitor_p.html" role="button"><span class="glyphicon glyphicon-refresh"></span> Refresh</a>
|
|
::
|
|
#(/jobStatus)#
|
|
</div>
|
|
</div>
|
|
</div>
|
|
#(/recrawlReport)#
|
|
</div>
|
|
</div>
|
|
#%env/templates/footer.template%#
|
|
</body>
|
|
</html>
|