Apache Solr: Indexing, Queries and Safer Operation
An introduction to a free search platform and the boundary between a search box and its server.
עברית
A search platform for your own data
Apache Solr is free, open-source search software built on Apache Lucene. It indexes documents and answers queries over that index. Developers can use it for a website, a catalogue or an internal document collection. It is a server component; the public search page is an application built around it.
From documents to results
A project needs to decide which fields represent a title, body, date or category. Indexing and query analysis must suit the content and its languages. Relevance is something to evaluate with representative queries and expected results, rather than judging a system by one impressive demonstration.
Read the syntax carefully
With the standard parser, double quotes express a phrase and uppercase AND, OR and NOT express Boolean combinations. A field prefix limits a term to that field. Field names depend on the application schema; the following examples assume a field named title.
"free software"
linux AND server
title:linuxQueries, filters and interface design
In the server API, q carries the query while fq adds filter queries. rows controls the requested result count and sort selects ordering. An application should expose only the options its users need, validate input and limit expensive requests. These API parameters are not a promise that every public search box accepts them.
Operate it behind a controlled boundary
Keep administrative endpoints away from the public internet. Use network restrictions, authentication, authorization and TLS as appropriate. Monitor the service, maintain updates and test recovery. A search interface should mediate public requests rather than handing visitors unrestricted access to the Solr administration API.