Hello, I've been using auto-phrasing. I believe it was my company's query to LucidWorks that got that initial implementation created. In working with it I found a few issues and forked the repo and simplified some code (where I didn't need features) and expanded the testing quite a bit. I've got my version deployed successfully in our Solr implementation. One key difference is that I only squash the phrases by removing whitespace. So, "seat cushions" becomes "seatcushions".
My fork is at https://github.com/jstrassburg/auto-phrase-tokenfilter Also, it does support configuring different downstream parsers. If you want edismax simply add defType dismax to your query parser config like this: <queryParser name="autophrasingParser" class=" com.lucidworks.analysis.AutoPhrasingQParserPlugin"> <str name="phrases">autophrases.txt</str> <str name="defType">dismax</str> </queryParser> Ted, I'd appreciate if you took a look at my fork and provide any feedback. JiM On Fri, Dec 12, 2014 at 2:31 PM, Ted Sullivan <ted.sulli...@lucidworks.com> wrote: > > Hi Shamik: > > Can you send me a JSON output using debugQuery=true so I can help > troubleshoot this? > > As to the question about edismax features - yes I *think* so :) but it > would > be great if you could give me some specific examples of queries as I am > currently writing the test cases for this. General disclaimer - I haven't > got this into shape for submission to Solr/Lucene yet so this feedback > would > be of great help to me. Thanks. > > Ted > > > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/Have-anyone-used-Automatic-Phrase-Tokenization-AutoPhrasingTokenFilterFactory-tp4173808p4174087.html > Sent from the Solr - User mailing list archive at Nabble.com. >
