net.fortytwo.stream.caching

Class QueryIndex<T,C>

java.lang.Object

net.fortytwo.stream.caching.QueryIndex<T,C>

public class QueryIndex<T,C>
extends java.lang.Object

A recursive data structure which indexes tuple-based queries and matches incoming tuples against stored queries in an incremental fashion.

Author:

Joshua Shinavier (http://fortytwo.net)

Constructor Summary

Constructors

Constructor and Description
QueryIndex(int tupleSize) Constructs a new query index for tuples of a given size

Method Summary

Modifier and Type	Method and Description
void	add(Query<T,C> query) Indexes a query for matching of future tuples
boolean	add(T[] tuple, java.util.function.BiConsumer<C,Bindings<T>> handler, int ttl, long now) Consumes a tuple and matches it against any and all applicable queries, possibly producing one or more solutions.
void	clear() Returns this index to its original, empty state, removing all queries and all solutions
void	remove(Query<T,C> query) Removes a previously added query.
void	removeExpired(long now) Enforce query and tuple time-to-live by removing expired graph patterns and solutions.
void	renew(Query<T,C> query, int ttl, long now) Renews a previously submitted query, applying a new time-to-live with respect to the current moment.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

QueryIndex

public QueryIndex(int tupleSize)

Constructs a new query index for tuples of a given size

Parameters:

tupleSize - the size of tuples to index and match. For RDF, an appropriate tuple size is 3, or 4 for RDF with Named Graphs

Method Detail

clear

public void clear()

Returns this index to its original, empty state, removing all queries and all solutions

add

public void add(Query<T,C> query)

Indexes a query for matching of future tuples

Parameters:

query - the query to index

remove

public void remove(Query<T,C> query)

Removes a previously added query. The given query will no longer match incoming tuples, and the memory it previously occupied will be freed.

Parameters:

query - the query to remove

renew

public void renew(Query<T,C> query,
                  int ttl,
                  long now)

Renews a previously submitted query, applying a new time-to-live with respect to the current moment. The query will expire only after its new expiration time has been reached.

Parameters:

query - the query to refresh

ttl - a time-to-live, in seconds. If the special value 0 is supplied, the query will never expire.

now - the current time, in milliseconds since the Unix epoch

add

public boolean add(T[] tuple,
                   java.util.function.BiConsumer<C,Bindings<T>> handler,
                   int ttl,
                   long now)

Consumes a tuple and matches it against any and all applicable queries, possibly producing one or more solutions. Note: this method is thread-safe; any number of threads may add tuples concurrently.

Parameters:

tuple - the input tuple

handler - a handler for any solutions produced

ttl - the time-to-live of the tuple and any derived solutions, in seconds. Use 0 for infinite time-to-live.

now - the current Unix time, in milliseconds (not seconds)

Returns:

whether at least one tuple pattern is applied to the given tuple, changing the state of the query index. If no tuple pattern is applied, the tuple simply passes through.

removeExpired

public void removeExpired(long now)

Enforce query and tuple time-to-live by removing expired graph patterns and solutions. The latter is a space-saving operation which has no effect on query evaluation; it forces all expired solutions to be removed at once, rather than merely becoming irrelevant and being gradually removed upon discovery.

Parameters:

now - the current time, in milliseconds since the Unix epoch