Mailing-List: contact user-help@uima.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@uima.apache.org
Received-SPF: pass (nike.apache.org: domain of burnlewis@gmail.com designates
 74.125.82.51 as permitted sender)
MIME-Version: 1.0
In-Reply-To: <7DEAB9C7-A98A-46DA-A03E-9D48E71C7589@googlemail.com>
References: <7DEAB9C7-A98A-46DA-A03E-9D48E71C7589@googlemail.com>
Date: Thu, 13 Dec 2012 10:21:44 -0500
Message-ID: 
 <CA+U9pLt9yyUFU5c6XCTiQXc7aFTNjz5Oaqod86evVpKv8ZSCDg@mail.gmail.com>
Subject: Re: How to clone a FeatureStructure?
From: Burn Lewis <burnlewis@gmail.com>
To: user@uima.apache.org
Content-Type: multipart/alternative; boundary=089e0122f17003eefe04d0bd7937

--089e0122f17003eefe04d0bd7937
Content-Type: text/plain; charset=ISO-8859-1

I don't quite understand your question.  What do you mean by cloning?  Are
you creating another array that references the same FSs or are you also
replicating each FS in the array?
What appears to be wrong?  Are you expecting a different parse tree when
the rules are different but instead getting identical trees?  If so, why do
you think that's a UIMA problem.

It may not be relevant, but my view is that CASes should be treated as
places to store public information that is to be shared amongst
annotators.  Any intermediate results computed by an annotator should not
be stored in the CAS ... basically because there is no delete method.
FeatureStructures can be removed from indexes but that does not delete
them.  The only form of garbage collection occurs when a CAS is serialized
to a file or to a remote service, as then only indexed FeatureStructures
are sent.

~Burn


On Wed, Dec 12, 2012 at 5:33 PM, Administrator <
roman.alexandrovich.mueller@gmail.com> wrote:

> Dear uima user,
>
> here is what i am trying. One analysis engine receives an array of feature
> structures. It clones all feature structures of the array and produces a
> parse tree. The cloning and parsing happens several times but with
> different rules. However the result is always a parse forrest in witch all
> parse trees are the same (the last one). Creating new FeatureStructures
> with the new keyword has the same result. What can i do?
>
> Thanks in advance
> Roman

--089e0122f17003eefe04d0bd7937--