So I'm starting to get my head around instance_/cast_ ... Basically,
sometimes Java sees an object as a less-specific type than it actually is,
and being a static language, the attr/method names available are those of the
less-specific type.
For example:
In [3]: q=QueryParser('pants', StandardAnalyzer()).parse('shirts AND -shoes')
In [4]: q
Out[4]: <Query: +pants:shirts -pants:shoes>
In [5]: BooleanQuery.instance_(q)
Out[5]: True
In [6]: q.getClauses()
<type 'exceptions.AttributeError'>: 'Query' object has no
attribute 'getClauses'
In [7]: bq=BooleanQuery.cast_(q)
In [8]: bq.getClauses()
Out[8]: [<BooleanClause: +pants:shirts>, <BooleanClause: -pants:shoes>]
While this is probably perfectly sensible to a Java/C++ programmer, as a
Python coder this behavior is somewhat foreign. Hopefully the above example
will help make it clearer - the one in the README wasn't quite clear to me.
My question is: when should we use isinstance(obj, Klass) vs.
Klass.instance_(obj)? From below, it looks like python's isinstance can look
up the Java class hierarchy but not down. As a general rule, should we
always be using Klass.instance_ for such tests? Will this work correctly
with pure-python types / Python subclasses of Java classes?
In [12]: isinstance(q, BooleanQuery)
Out[12]: False
In [23]: BooleanQuery.instance_(q)
Out[23]: True
In [13]: isinstance(bq, BooleanQuery)
Out[13]: True
In [14]: isinstance(bq, Query)
Out[14]: True
--
Peter Fein || 773-575-0694 || [EMAIL PROTECTED]
http://www.pobox.com/~pfein/ || PGP: 0xCCF6AE6B
irc: [EMAIL PROTECTED] || jabber: [EMAIL PROTECTED]
_______________________________________________
pylucene-dev mailing list
[email protected]
http://lists.osafoundation.org/mailman/listinfo/pylucene-dev
