René Kriegler
René Kriegler
With the split of the Apache Lucene and Apache Solr projects, we can no longer rely on a Solr version becoming available soon after a Lucene release. Currently, this is...
Querqy 5 changed how rewriters can be configured by moving to an HTTP-based API. In order to facilitate the transition we allowed a temporary solution in solrconfig.xml via searcherListeners that...
It can happen that rewriters create the same query clause more than once. For example, a WordBreakCompoundRewriter might break up the query `sofabed` by adding a synonym `sofa bed` and,...
Some Querqy applications use thousands of rewrite rules. Rewrite chains become increasingly complex. As this makes understanding the rewriting process hard, we should add a number of features to inspect...
Remove from pom.xml - we don't use it
When we turn a Querqy query into a Lucene query, we are ignoring the positionLength attribute from Lucene's analysis chain. So far, we've hardly noticed this. A first case, where...
We currently have `AbstractLoggingRewriter extends AbstractNodeVisitor` On the other hand, we never assume that the visit() methods return a Node object. To provide greater flexibility to subclasses of the AbstractLoggingRewriter...
We get a number of deprecation warnings in our tests. Let's review them and avoid them where possible.
We see a number of stacktraces when running our tests. This is okay - some of our test cases test the error case and the stacktraces are produced by Solr's...
Currently Querqy produces only term-centric queries, for example: ```Common rules rewriter => SYNONYM: iphone x ``` produces: ```(field1:iphone | field2:iphone) AND (field1:x | field2:x)``` In some cases we want the...