> We construct our lucene query after certain analysis(ex :
> words
> segmentation, category identification).
By analysis, I referring charfilter(s)+tokenizer+tokenfilter(s) combination.
> Do you mean we
> plugin those analysis
> logic and query construction part onto solr, and solr takes
> the very
> beginning input.
>
I didn't understand what "very beginning input" is.
Lets say you have pure java program that takes a String as an input, and
returns org.apache.lucene.search.Query as output. You can embed this into solr.
e.g. Query constructMyMagicQuery(String)
public QParser createParser(String qstr, SolrParams localParams, SolrParams
params, SolrQueryRequest req) {
return new QParser(qstr, localParams, params, req) {
public Query parse() throws ParseException {
String query = params.get(CommomParams.Q);
return constructMyMagicQuery(query);
}
}
}
Your custom program can read/use any key-value pairs from the search URL, if
required.
