mged-mage-java — Development discussion of the MAGE API for Java

[Mged-mage-java] Mage STK & Maven

[Trond N. <xo...@gm...>]

Hi,

Is anyone using Mage-STK in a project managed with Maven? More specifically,
does anyone know of a public repository using it?

I'm converting a pre-existing project to use Maven that relies on Mage-STK,
and am trying to work out which dependencies I'll need manage myself.

Thanks!

-- 
Thanks!
Trond.

Cell: (206) 913-3331
Blogs:
http://www.meme-hazard.org
http://aigantighe.livejournal.com

Re: [Mged-mage-java] support for tab2mage doc builder

[Kjell P. <Kje...@bc...>]

As you're saying, what is needed is a mapping from the general MAGE 
structure to the
more restricted MAGE-TAB structure.

As the MAGE-TAB software from EBI can convert a MAGE-TAB doc to MAGE-ML, 
and
other people have been playing around with the more general mapping of 
going the other way,
some thoughts and work has been spent on the subject.

But as far as I know there is currently no initiative to get some code 
performing this mapping either
way into the current MAGEstk. Since the stk follows the standard 
specification, I don't think such tools
will be dealt with before MAGE v2 (including MAGE-TAB as a platform 
sepcific implementation) will
be more mature and ready for release.

Best,
Kjell

juan elvira wrote:
> I was thinking in the following scenario:
> My microarray data management application uses MAGE-stk to build a 
> MAGE-ML doc and now it should create a MAGE-TAB doc. The MAGE-stk is 
> based on MAGE-OM and MAGE-TAB is based on MAGE-OM too thus it's straight 
> to think in extending MAGE-stk to build the MAGE-TAB docs.
>
>
> Arto Viitanen wrote:
>   
>> juan elvira wrote:
>>     
>>> Hi all,
>>> I wonder if you have planned to extend the java MAGE-stk to enable 
>>> building a MAGE-TAB document from a MAGEJava object? Something like 
>>> MAGEJava.writeMAGETAB. It would be very welcome. Although I 
>>> understand it's not a straightforward task.
>>>
>>>       
>> Just a hint: if you need to handle Excel files, I suggest you check
>> Jakarta POI (http://jakarta.apache.org/poi/). I used it in generating
>> Excel files from queries across Arraysxpress database, but it should
>> work fine in  reading the Excel files also.
>>
>>
>>     
>
>
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> Mged-mage-java mailing list
> Mge...@li...
> https://lists.sourceforge.net/lists/listinfo/mged-mage-java
>   

Re: [Mged-mage-java] support for tab2mage doc builder

[juan e. <jua...@in...>]

I was thinking in the following scenario:
My microarray data management application uses MAGE-stk to build a 
MAGE-ML doc and now it should create a MAGE-TAB doc. The MAGE-stk is 
based on MAGE-OM and MAGE-TAB is based on MAGE-OM too thus it's straight 
to think in extending MAGE-stk to build the MAGE-TAB docs.


Arto Viitanen wrote:
> juan elvira wrote:
>> Hi all,
>> I wonder if you have planned to extend the java MAGE-stk to enable 
>> building a MAGE-TAB document from a MAGEJava object? Something like 
>> MAGEJava.writeMAGETAB. It would be very welcome. Although I 
>> understand it's not a straightforward task.
>>
> Just a hint: if you need to handle Excel files, I suggest you check
> Jakarta POI (http://jakarta.apache.org/poi/). I used it in generating
> Excel files from queries across Arraysxpress database, but it should
> work fine in  reading the Excel files also.
>
>

Re: [Mged-mage-java] support for tab2mage doc builder

[Arto V. <art...@cs...>]

juan elvira wrote:
> Hi all,
> I wonder if you have planned to extend the java MAGE-stk to enable 
> building a MAGE-TAB document from a MAGEJava object? Something like 
> MAGEJava.writeMAGETAB. It would be very welcome. Although I understand 
> it's not a straightforward task.
>
Just a hint: if you need to handle Excel files, I suggest you check
Jakarta POI (http://jakarta.apache.org/poi/). I used it in generating
Excel files from queries across Arraysxpress database, but it should
work fine in  reading the Excel files also.


-- 
Arto Viitanen, CSC Ltd.
Espoo, Finland

[Mged-mage-java] support for tab2mage doc builder

[juan e. <jua...@in...>]

Hi all,
I wonder if you have planned to extend the java MAGE-stk to enable 
building a MAGE-TAB document from a MAGEJava object? Something like 
MAGEJava.writeMAGETAB. It would be very welcome. Although I understand 
it's not a straightforward task.

regards
Juan

[Mged-mage-java] Fwd: MAGE JAVA

[Jason S. <jas...@gm...>]

Dear Bhuwan,

It is better to write directly to the list: mge...@li...

not to me or the -owner account.

Can someon help with the question?

Cheers, jas.

---------- Forwarded message ----------
From: Tiwari, Bhuwan <bhu...@im...>
Date: Aug 31, 2006 4:43 PM
Subject: MAGE JAVA
To: mge...@li...





Hi Jason/ Admin,



I am first time user of MAGE-Java.

I am interested in reading ArrayExpress MAGE-ML file and using sample
code for this, modified code goes like as follows:



I am passing MAGE-ML file which I downloaded from ArrayExpress as
parameter. For Audit Security package it returns me right count for
CONTACT which was 7. But it doesn't load actual contact objects. So
when I read individual and tried to print there identifier or Name it
returns null. I used debugger it shows null as well.



Could you please guide me through this?



/*

 * example.java

 *

 * Created on March 24, 2005, 12:10 AM

 */



//package org.biomage.examples;



import org.biomage.Common.MAGEReader;

import org.biomage.Common.MAGEWriter;

import org.biomage.Common.MAGEJava;

import java.io.FileWriter;





import org.biomage.Experiment.Experiment_package;

import org.biomage.Experiment.Experiment;



import org.biomage.AuditAndSecurity.AuditAndSecurity_package;

import org.biomage.AuditAndSecurity.Contact;

import org.biomage.AuditAndSecurity.AuditAndSecurity_package.Contact_list;

import org.biomage.AuditAndSecurity.Person;

import org.biomage.AuditAndSecurity.Organization;







/**

 *

 * @author  bccs

 */

public class MyTestExample {



    /** Creates a new instance of example */

    public MyTestExample() {

    }



    /**

     * @param args the command line arguments

     */

    public static void main(String[] args) throws java.lang.Exception {



        // Read in a MAGE-ML file

        System.out.println("In Simple MAGE Example");

        MAGEReader mr = new MAGEReader(args[0]);



        System.out.println("Mage reader created");



        // Here's the Object model in memory for you :)

        MAGEJava mageom = mr.getMAGEobj();



        System.out.println("Mage object created");



        // Progress report to screen

        if (mageom != null) {

            System.out.println("MAGE-ML file "+args[0]+" read into
object model in memory.");

        } else {

            System.out.println("An error occured when trying to read
in "+args[0]+

                               ". Please check that the path/filename
"+args[0]+" is correct.");

        }







             AuditAndSecurity_package auditPackage =
mageom.getAuditAndSecurity_package();

             Contact_list contactList = auditPackage.getContact_list();

             System.out.println("contact List size is " + contactList.size());

             ///

             //System.out.println("person1 Last Name" + person1.getName());



             for(int i=0;i<contactList.size();i++)

             {

                        if (contactList.get(i) instanceof Organization )

                        {

                                    System.out.println("I am
Organisation " + i);

                        }

                        else if (contactList.get(i) instanceof Person)

                        {

                                    Person person1 = (Person)contactList.get(i);

                                    System.out.println("I am person "
+ i + " My first name is " + person1.getFirstName());

                                    System.out.println("I am person "
+ i + " My identifier is " + person1.getIdentifier());

                        }



             }





    }



}



Kind Regards



Bhuwan Tiwari

Software Developer

Imperial College,  London

[Mged-mage-java] RE: [Fuge-devel] STK and database schema for FuGE and extensions

[Miller, M. D (Rosetta) <Mic...@Ro...>]

Hi Andy,

Sounds good, figured it was good to raise this use case sooner than
later.

> The code at present can also work without a database back end=20
> at all (I
> think - although I haven't tested it).=20

that's the main requirement, if there is some extra unused code or
packages, that wouldn't be a big deal.

> ..., because Hibernate can be configured to
> work with any schema ...

as an aside, only...

I'm not sure that the effort to do use Hibernate for some databases that
are radically different from the FuGE/MAGE structure would be worth it.
But I am not an expert in Hibernate, though it looks like I'll probably
need to learn soon enough.

cheers,
Michael

> -----Original Message-----
> From: fug...@li...=20
> [mailto:fug...@li...] On Behalf Of=20
> Andy Jones
> Sent: Monday, March 20, 2006 9:01 AM
> To: fug...@li...
> Cc: mge...@li...
> Subject: RE: [Fuge-devel] STK and database schema for FuGE=20
> and extensions
>=20
>=20
> Hi Michael,
> Yes this occurred to me that most users will want their own=20
> database back
> end. I think this will be okay, because Hibernate can be configured to
> work with any schema - it's just a matter of changing the Hibernate
> mapping file. However, this does place a restriction that=20
> Hibernate only
> works with MySQL, Oracle and Postgres.
>=20
> The code at present can also work without a database back end=20
> at all (I
> think - although I haven't tested it). I'm fairly sure that=20
> the transform
> operation to convert from objects to XML, has no reliance on=20
> the database.
> So far, I've only tested converting directly from objects=20
> retrieved from
> the database to XML.
>=20
> I've invested a few weeks work in this, but much of this is
> understanding what Hibernate/AndroMDA can do. The time coding spent on
> actual coding has been fairly small so I don't have that much invested
> that we can't re-code if need be, to keep Hibernate and the XML in-out
> in completely separate modules.
>=20
> cheers,
> Andy
>=20
>=20
> On Mon, 20 Mar 2006, Miller, Michael D (Rosetta) wrote:
>=20
> > Hi Andy,
> >
> > Great stuff here--the only concern I have is that the hibernate and
> > database persistence layer is not linked at all to the code=20
> that goes
> > between the FuGE-ML representation and the in-memory=20
> FuGE-java (or what
> > ever languages we support).
> >
> > Many users of the toolkit will have their own database schemas that
> > aren't necessarily modeled on FuGE/MAGE (us, for instance!)=20
> and so would
> > not need/want this functionality.
> >
> > cheers,
> > Michael
> >
> > 	-----Original Message-----
> > 	From: fug...@li...
> > [mailto:fug...@li...] On Behalf=20
> Of Andy Jones
> > 	Sent: Monday, March 20, 2006 7:39 AM
> > 	To: fug...@li...;
> > mge...@li...
> > 	Subject: [Fuge-devel] STK and database schema for FuGE and
> > extensions
> >
> >
> >
> > 	Hi all,
> >
> > 	I've now completed a first pass at a software toolkit for FuGE
> > milestone 2. The code performs database persistence and=20
> writing out of
> > valid FuGE XML (see status below). Support for reading FuGE XML is
> > pretty limited at this stage but could be developed into a=20
> full parser
> > fairly quickly I think.
> >
> >
> >
> > 	I'm not going to put any more time into it now until we are
> > close to the UML for FuGE version 1. The code is now posted=20
> on the FuGE
> > site for anyone to test out as a complete bundle
> > (http://fuge.sourceforge.net/dev/index.php). I've also=20
> pulled out the
> > database schema for MySQL and the Hibernate mapping file=20
> which are also
> > posted as separate files. This means that theoretically=20
> someone could
> > use the code without using AndroMDA (I think...).
> >
> >
> >
> > 	I've written some fairly detailed notes describing what I've
> > done and how it can be used, which is also posted on the website.
> >
> >
> >
> > 	The status of the code is as follows:
> >
> > 	-        Supports creating Java objects for every entity in FuGE
> > milestone 2 (allows associations, attributes etc. to be set).
> >
> > 	-        These objects can be stored, updated and deleted from a
> > database.
> >
> > 	-        Methods have been implemented that allow all
> > Describable objects to be retrieved from the database by=20
> their unique
> > database ID (different from the FuGE identifier) and then=20
> cast to the
> > correct type.
> >
> > 	-        An entity can be converted to a DOM element, and then
> > written to a file as XML. For instance if a FuGEroot element is
> > retrieved, this will process all associations to create a=20
> complete FuGE
> > file, so long as all associations have been created correctly.
> >
> > 	-        The native XML produced is close but not an exact match
> > of the FuGE M2 XML Schema. Therefore, there are helper methods that
> > perform this conversion to output valid FuGE M2. If we make=20
> any major
> > changes to the XML Schema, these helper methods may need to=20
> be updated.
> >
> > 	-        Reading XML is only supported so far in that an XML
> > file can be parsed, and the methods supplied can be used to create
> > objects as required. However, there is no code for=20
> processing objects
> > and their associations in the correct order etc. It would=20
> probably take
> > a week or two of coding to create a full FuGE parser. I'm=20
> not sure if
> > the native XML persistence offered by Hibernate can be made=20
> to work, I
> > think it is really intended for simpler schemas than FuGE.
> >
> >
> >
> > 	It's only been tested with MySQL but the AndroMDA properties can
> > be adapted to work with Oracle or Postgres. The Oracle schema didn't
> > work straight out of the box, but I can probably generate one fairly
> > easily if anyone wants to try this out instead.
> >
> >
> >
> > 	There are obviously still lots of unresolved issues and the code
> > has not been tested rigorously at all, so if anyone wants=20
> to contribute
> > some time, please get in touch.
> >
> > 	Best wishes,
> >
> > 	Andy
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > ________________________________
> >
> >
> > 	From: mge...@li...
> > [mailto:mge...@li...] On Behalf=20
> Of Andy Jones
> > 	Sent: 10 March 2006 15:52
> > 	To: fug...@li...; 'mged-mage2';
> > mge...@li...
> > 	Subject: [Mged-mage2] RE: [Fuge-devel] STK and database schema
> > for FuGE and extensions
> >
> >
> >
> > 	Hi all,
> >
> > 	I thought I'd give a quick update on this issue. I've spent a
> > few days looking at Hibernate and AndroMDA and I think this=20
> is going to
> > be a really good way to go. I've managed to generate a=20
> framework from a
> > cut down FuGE (just FuGE Common). The framework creates=20
> Java objects for
> > each entity and a database schema. It provides methods for=20
> creating all
> > Java objects, associations and attributes, then writing or=20
> retrieving
> > them from the database. I don't think there will be any major issues
> > scaling up to complete FuGE (and then extensions).
> >
> >
> >
> > 	I've also started experimenting with the XML persistence in
> > Hibernate 3. This seems to work pretty well for the small tests I've
> > performed so far, although there will be a few issues to iron out.
> > Basically, I can create objects in the database, get them=20
> back as Java
> > objects, which can then be converted to a Dom4J Entity=20
> (in-memory XML
> > nodes). The native mapping does not quite match our XML=20
> Schema, but this
> > can be customized pretty easily through the AndroMDA cartridge that
> > generates the Hibernate mapping file. I should be able to get this
> > sorted in a day or two.
> >
> >
> >
> > 	If we customize the EntityToHibernate mapping files (rather than
> > writing bespoke code) we should be able to use the=20
> Hibernate framework
> > also for reading XML directly into Java objects, in other=20
> words I think
> > we are quite close to having a basic API for FuGE.
> >
> >
> >
> > 	A customized mapping could also be developed for a different
> > database schema relatively easily by manually altering one=20
> XML file, so
> > the same API will work with in-house databases rather than the
> > auto-generated database schema.
> >
> >
> >
> > 	The only thing that will be a bit of an issue is unique
> > identifiers. Hibernate seems to want to have integer IDs on every
> > database table/entity, and in the XML for referencing other=20
> items in the
> > document. In FuGE we have our own identification system. However, I
> > think we can get around this by having a server that=20
> stores/converts all
> > database integer IDs and FuGE IDs (although at the moment I=20
> have no idea
> > how to make it work if FuGE documents are split across files...).
> >
> >
> >
> > 	I'm going to spend a few more days on this then I'll post it all
> > on the FuGE website, including instructions for how to=20
> setup and install
> > AndroMDA and Hibernate (which is the hardest part of=20
> this!). If anyone
> > wants to contribute some coding time, let me know,
> >
> >
> >
> > 	Cheers
> >
> > 	Andy
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > ________________________________
> >
> >
> > 	From: fug...@li...
> > [mailto:fug...@li...] On Behalf=20
> Of Andy Jones
> > 	Sent: 02 March 2006 11:32
> > 	To: fug...@li...; 'mged-mage2';
> > mge...@li...
> > 	Subject: [Fuge-devel] STK and database schema for FuGE and
> > extensions
> >
> >
> >
> > 	Hi all,
> >
> > 	I've been taking a look at what else AndroMDA can do for us from
> > the UML. Using the Hibernate cartridges AndroMDA generates a fairly
> > sensible looking relational schema (for MySQL, Oracle or=20
> PostGres). So
> > far I've only tested with a simpler test model but I didn't=20
> encounter
> > any problems that should prevent this working with FuGE. I don't see
> > that a relational schema will be an essential part of FuGE=20
> but I think
> > getting a relational schema for free for FuGE and=20
> extensions is a bonus.
> >
> >
> >
> >
> > 	More interestingly, Hibernate generates Java code for performing
> > object-relational persistence. Basically, for every entity=20
> in the model
> > it provides code for reading and writing objects to the=20
> database (and
> > methods can be added in UML for querying the database very easily).
> >
> >
> >
> > 	I found an article saying that Hibernate 3 also supports
> > object-XML persistence (http://www.devx.com/Java/Article/27896 and
> > http://www.hibernate.org/hib_docs/v3/reference/en/html/xml.html).
> > However, this feature seems to be in development and I'm not sure if
> > it's advanced enough for our purposes. So alternatively, I think the
> > Hibernate cartridges could be adapted fairly easily to=20
> generate code for
> > reading and writing XML as well.
> >
> >
> >
> > 	Ideally, the code should be modular so that we can read and
> > write XML to and from Java objects, and use this in=20
> conjunction with the
> > auto-generated Hibernate code, such that if developers=20
> wished they could
> > create their own in-house applications based on this=20
> framework. What do
> > people think?
> >
> >
> >
> > 	I've looked at the existing Java MAGEstk and I'm starting to
> > understand the basics. If anyone has an advice on how we=20
> might re-use
> > code from MAGEstk, in conjunction with what can be auto-generated by
> > AndroMDA, it would be really appreciated. I think we should=20
> be able to
> > get a basic API together fairly quickly, given that=20
> AndroMDA does a lot
> > for us automatically. I would like to encode as much of the logic as
> > possible in the UML and in the cartridges i.e. as little=20
> bespoke coding
> > as possible, such that the same process can be performed=20
> for FuGE, MAGE
> > 2, PSI models and so on...
> >
> >
> >
> > 	Cheers
> >
> > 	Andy
> >
> >
>=20
>=20
>=20
> -------------------------------------------------------
> This SF.Net email is sponsored by xPML, a groundbreaking=20
> scripting language
> that extends applications into web and mobile media. Attend=20
> the live webcast
> and join the prime developer group breaking into this new=20
> coding territory!
> http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D110944&bid=3D241720&
dat=3D121642
_______________________________________________
Fuge-devel mailing list
Fug...@li...
https://lists.sourceforge.net/lists/listinfo/fuge-devel

[Mged-mage-java] RE: [Fuge-devel] STK and database schema for FuGE and extensions

[Andy J. <aj...@cs...>]

Hi Michael,
Yes this occurred to me that most users will want their own database back
end. I think this will be okay, because Hibernate can be configured to
work with any schema - it's just a matter of changing the Hibernate
mapping file. However, this does place a restriction that Hibernate only
works with MySQL, Oracle and Postgres.

The code at present can also work without a database back end at all (I
think - although I haven't tested it). I'm fairly sure that the transform
operation to convert from objects to XML, has no reliance on the database.
So far, I've only tested converting directly from objects retrieved from
the database to XML.

I've invested a few weeks work in this, but much of this is
understanding what Hibernate/AndroMDA can do. The time coding spent on
actual coding has been fairly small so I don't have that much invested
that we can't re-code if need be, to keep Hibernate and the XML in-out
in completely separate modules.

cheers,
Andy


On Mon, 20 Mar 2006, Miller, Michael D (Rosetta) wrote:

> Hi Andy,
>
> Great stuff here--the only concern I have is that the hibernate and
> database persistence layer is not linked at all to the code that goes
> between the FuGE-ML representation and the in-memory FuGE-java (or what
> ever languages we support).
>
> Many users of the toolkit will have their own database schemas that
> aren't necessarily modeled on FuGE/MAGE (us, for instance!) and so would
> not need/want this functionality.
>
> cheers,
> Michael
>
> 	-----Original Message-----
> 	From: fug...@li...
> [mailto:fug...@li...] On Behalf Of Andy Jones
> 	Sent: Monday, March 20, 2006 7:39 AM
> 	To: fug...@li...;
> mge...@li...
> 	Subject: [Fuge-devel] STK and database schema for FuGE and
> extensions
>
>
>
> 	Hi all,
>
> 	I've now completed a first pass at a software toolkit for FuGE
> milestone 2. The code performs database persistence and writing out of
> valid FuGE XML (see status below). Support for reading FuGE XML is
> pretty limited at this stage but could be developed into a full parser
> fairly quickly I think.
>
>
>
> 	I'm not going to put any more time into it now until we are
> close to the UML for FuGE version 1. The code is now posted on the FuGE
> site for anyone to test out as a complete bundle
> (http://fuge.sourceforge.net/dev/index.php). I've also pulled out the
> database schema for MySQL and the Hibernate mapping file which are also
> posted as separate files. This means that theoretically someone could
> use the code without using AndroMDA (I think...).
>
>
>
> 	I've written some fairly detailed notes describing what I've
> done and how it can be used, which is also posted on the website.
>
>
>
> 	The status of the code is as follows:
>
> 	-        Supports creating Java objects for every entity in FuGE
> milestone 2 (allows associations, attributes etc. to be set).
>
> 	-        These objects can be stored, updated and deleted from a
> database.
>
> 	-        Methods have been implemented that allow all
> Describable objects to be retrieved from the database by their unique
> database ID (different from the FuGE identifier) and then cast to the
> correct type.
>
> 	-        An entity can be converted to a DOM element, and then
> written to a file as XML. For instance if a FuGEroot element is
> retrieved, this will process all associations to create a complete FuGE
> file, so long as all associations have been created correctly.
>
> 	-        The native XML produced is close but not an exact match
> of the FuGE M2 XML Schema. Therefore, there are helper methods that
> perform this conversion to output valid FuGE M2. If we make any major
> changes to the XML Schema, these helper methods may need to be updated.
>
> 	-        Reading XML is only supported so far in that an XML
> file can be parsed, and the methods supplied can be used to create
> objects as required. However, there is no code for processing objects
> and their associations in the correct order etc. It would probably take
> a week or two of coding to create a full FuGE parser. I'm not sure if
> the native XML persistence offered by Hibernate can be made to work, I
> think it is really intended for simpler schemas than FuGE.
>
>
>
> 	It's only been tested with MySQL but the AndroMDA properties can
> be adapted to work with Oracle or Postgres. The Oracle schema didn't
> work straight out of the box, but I can probably generate one fairly
> easily if anyone wants to try this out instead.
>
>
>
> 	There are obviously still lots of unresolved issues and the code
> has not been tested rigorously at all, so if anyone wants to contribute
> some time, please get in touch.
>
> 	Best wishes,
>
> 	Andy
>
>
>
>
>
>
>
>
>
>
>
>
> ________________________________
>
>
> 	From: mge...@li...
> [mailto:mge...@li...] On Behalf Of Andy Jones
> 	Sent: 10 March 2006 15:52
> 	To: fug...@li...; 'mged-mage2';
> mge...@li...
> 	Subject: [Mged-mage2] RE: [Fuge-devel] STK and database schema
> for FuGE and extensions
>
>
>
> 	Hi all,
>
> 	I thought I'd give a quick update on this issue. I've spent a
> few days looking at Hibernate and AndroMDA and I think this is going to
> be a really good way to go. I've managed to generate a framework from a
> cut down FuGE (just FuGE Common). The framework creates Java objects for
> each entity and a database schema. It provides methods for creating all
> Java objects, associations and attributes, then writing or retrieving
> them from the database. I don't think there will be any major issues
> scaling up to complete FuGE (and then extensions).
>
>
>
> 	I've also started experimenting with the XML persistence in
> Hibernate 3. This seems to work pretty well for the small tests I've
> performed so far, although there will be a few issues to iron out.
> Basically, I can create objects in the database, get them back as Java
> objects, which can then be converted to a Dom4J Entity (in-memory XML
> nodes). The native mapping does not quite match our XML Schema, but this
> can be customized pretty easily through the AndroMDA cartridge that
> generates the Hibernate mapping file. I should be able to get this
> sorted in a day or two.
>
>
>
> 	If we customize the EntityToHibernate mapping files (rather than
> writing bespoke code) we should be able to use the Hibernate framework
> also for reading XML directly into Java objects, in other words I think
> we are quite close to having a basic API for FuGE.
>
>
>
> 	A customized mapping could also be developed for a different
> database schema relatively easily by manually altering one XML file, so
> the same API will work with in-house databases rather than the
> auto-generated database schema.
>
>
>
> 	The only thing that will be a bit of an issue is unique
> identifiers. Hibernate seems to want to have integer IDs on every
> database table/entity, and in the XML for referencing other items in the
> document. In FuGE we have our own identification system. However, I
> think we can get around this by having a server that stores/converts all
> database integer IDs and FuGE IDs (although at the moment I have no idea
> how to make it work if FuGE documents are split across files...).
>
>
>
> 	I'm going to spend a few more days on this then I'll post it all
> on the FuGE website, including instructions for how to setup and install
> AndroMDA and Hibernate (which is the hardest part of this!). If anyone
> wants to contribute some coding time, let me know,
>
>
>
> 	Cheers
>
> 	Andy
>
>
>
>
>
>
>
>
>
>
>
>
> ________________________________
>
>
> 	From: fug...@li...
> [mailto:fug...@li...] On Behalf Of Andy Jones
> 	Sent: 02 March 2006 11:32
> 	To: fug...@li...; 'mged-mage2';
> mge...@li...
> 	Subject: [Fuge-devel] STK and database schema for FuGE and
> extensions
>
>
>
> 	Hi all,
>
> 	I've been taking a look at what else AndroMDA can do for us from
> the UML. Using the Hibernate cartridges AndroMDA generates a fairly
> sensible looking relational schema (for MySQL, Oracle or PostGres). So
> far I've only tested with a simpler test model but I didn't encounter
> any problems that should prevent this working with FuGE. I don't see
> that a relational schema will be an essential part of FuGE but I think
> getting a relational schema for free for FuGE and extensions is a bonus.
>
>
>
>
> 	More interestingly, Hibernate generates Java code for performing
> object-relational persistence. Basically, for every entity in the model
> it provides code for reading and writing objects to the database (and
> methods can be added in UML for querying the database very easily).
>
>
>
> 	I found an article saying that Hibernate 3 also supports
> object-XML persistence (http://www.devx.com/Java/Article/27896 and
> http://www.hibernate.org/hib_docs/v3/reference/en/html/xml.html).
> However, this feature seems to be in development and I'm not sure if
> it's advanced enough for our purposes. So alternatively, I think the
> Hibernate cartridges could be adapted fairly easily to generate code for
> reading and writing XML as well.
>
>
>
> 	Ideally, the code should be modular so that we can read and
> write XML to and from Java objects, and use this in conjunction with the
> auto-generated Hibernate code, such that if developers wished they could
> create their own in-house applications based on this framework. What do
> people think?
>
>
>
> 	I've looked at the existing Java MAGEstk and I'm starting to
> understand the basics. If anyone has an advice on how we might re-use
> code from MAGEstk, in conjunction with what can be auto-generated by
> AndroMDA, it would be really appreciated. I think we should be able to
> get a basic API together fairly quickly, given that AndroMDA does a lot
> for us automatically. I would like to encode as much of the logic as
> possible in the UML and in the cartridges i.e. as little bespoke coding
> as possible, such that the same process can be performed for FuGE, MAGE
> 2, PSI models and so on...
>
>
>
> 	Cheers
>
> 	Andy
>
>

[Mged-mage-java] RE: [Fuge-devel] STK and database schema for FuGE and extensions

[Miller, M. D (Rosetta) <Mic...@Ro...>]

Hi Andy,
=20
Great stuff here--the only concern I have is that the hibernate and
database persistence layer is not linked at all to the code that goes
between the FuGE-ML representation and the in-memory FuGE-java (or what
ever languages we support).
=20
Many users of the toolkit will have their own database schemas that
aren't necessarily modeled on FuGE/MAGE (us, for instance!) and so would
not need/want this functionality.
=20
cheers,
Michael

	-----Original Message-----
	From: fug...@li...
[mailto:fug...@li...] On Behalf Of Andy Jones
	Sent: Monday, March 20, 2006 7:39 AM
	To: fug...@li...;
mge...@li...
	Subject: [Fuge-devel] STK and database schema for FuGE and
extensions
=09
=09

	Hi all,

	I've now completed a first pass at a software toolkit for FuGE
milestone 2. The code performs database persistence and writing out of
valid FuGE XML (see status below). Support for reading FuGE XML is
pretty limited at this stage but could be developed into a full parser
fairly quickly I think.=20

	=20

	I'm not going to put any more time into it now until we are
close to the UML for FuGE version 1. The code is now posted on the FuGE
site for anyone to test out as a complete bundle
(http://fuge.sourceforge.net/dev/index.php). I've also pulled out the
database schema for MySQL and the Hibernate mapping file which are also
posted as separate files. This means that theoretically someone could
use the code without using AndroMDA (I think...).

	=20

	I've written some fairly detailed notes describing what I've
done and how it can be used, which is also posted on the website.

	=20

	The status of the code is as follows:

	-        Supports creating Java objects for every entity in FuGE
milestone 2 (allows associations, attributes etc. to be set).

	-        These objects can be stored, updated and deleted from a
database.

	-        Methods have been implemented that allow all
Describable objects to be retrieved from the database by their unique
database ID (different from the FuGE identifier) and then cast to the
correct type.

	-        An entity can be converted to a DOM element, and then
written to a file as XML. For instance if a FuGEroot element is
retrieved, this will process all associations to create a complete FuGE
file, so long as all associations have been created correctly.

	-        The native XML produced is close but not an exact match
of the FuGE M2 XML Schema. Therefore, there are helper methods that
perform this conversion to output valid FuGE M2. If we make any major
changes to the XML Schema, these helper methods may need to be updated.

	-        Reading XML is only supported so far in that an XML
file can be parsed, and the methods supplied can be used to create
objects as required. However, there is no code for processing objects
and their associations in the correct order etc. It would probably take
a week or two of coding to create a full FuGE parser. I'm not sure if
the native XML persistence offered by Hibernate can be made to work, I
think it is really intended for simpler schemas than FuGE.

	=20

	It's only been tested with MySQL but the AndroMDA properties can
be adapted to work with Oracle or Postgres. The Oracle schema didn't
work straight out of the box, but I can probably generate one fairly
easily if anyone wants to try this out instead.

	=20

	There are obviously still lots of unresolved issues and the code
has not been tested rigorously at all, so if anyone wants to contribute
some time, please get in touch.

	Best wishes,

	Andy

	=20

	=20

	=20

	=20

	=20

=09
________________________________


	From: mge...@li...
[mailto:mge...@li...] On Behalf Of Andy Jones
	Sent: 10 March 2006 15:52
	To: fug...@li...; 'mged-mage2';
mge...@li...
	Subject: [Mged-mage2] RE: [Fuge-devel] STK and database schema
for FuGE and extensions

	=20

	Hi all,

	I thought I'd give a quick update on this issue. I've spent a
few days looking at Hibernate and AndroMDA and I think this is going to
be a really good way to go. I've managed to generate a framework from a
cut down FuGE (just FuGE Common). The framework creates Java objects for
each entity and a database schema. It provides methods for creating all
Java objects, associations and attributes, then writing or retrieving
them from the database. I don't think there will be any major issues
scaling up to complete FuGE (and then extensions).=20

	=20

	I've also started experimenting with the XML persistence in
Hibernate 3. This seems to work pretty well for the small tests I've
performed so far, although there will be a few issues to iron out.
Basically, I can create objects in the database, get them back as Java
objects, which can then be converted to a Dom4J Entity (in-memory XML
nodes). The native mapping does not quite match our XML Schema, but this
can be customized pretty easily through the AndroMDA cartridge that
generates the Hibernate mapping file. I should be able to get this
sorted in a day or two.

	=20

	If we customize the EntityToHibernate mapping files (rather than
writing bespoke code) we should be able to use the Hibernate framework
also for reading XML directly into Java objects, in other words I think
we are quite close to having a basic API for FuGE.=20

	=20

	A customized mapping could also be developed for a different
database schema relatively easily by manually altering one XML file, so
the same API will work with in-house databases rather than the
auto-generated database schema.

	=20

	The only thing that will be a bit of an issue is unique
identifiers. Hibernate seems to want to have integer IDs on every
database table/entity, and in the XML for referencing other items in the
document. In FuGE we have our own identification system. However, I
think we can get around this by having a server that stores/converts all
database integer IDs and FuGE IDs (although at the moment I have no idea
how to make it work if FuGE documents are split across files...).

	=20

	I'm going to spend a few more days on this then I'll post it all
on the FuGE website, including instructions for how to setup and install
AndroMDA and Hibernate (which is the hardest part of this!). If anyone
wants to contribute some coding time, let me know,=20

	=20

	Cheers

	Andy

	=20

	=20

	=20

	=20

	=20

=09
________________________________


	From: fug...@li...
[mailto:fug...@li...] On Behalf Of Andy Jones
	Sent: 02 March 2006 11:32
	To: fug...@li...; 'mged-mage2';
mge...@li...
	Subject: [Fuge-devel] STK and database schema for FuGE and
extensions

	=20

	Hi all,

	I've been taking a look at what else AndroMDA can do for us from
the UML. Using the Hibernate cartridges AndroMDA generates a fairly
sensible looking relational schema (for MySQL, Oracle or PostGres). So
far I've only tested with a simpler test model but I didn't encounter
any problems that should prevent this working with FuGE. I don't see
that a relational schema will be an essential part of FuGE but I think
getting a relational schema for free for FuGE and extensions is a bonus.


	=20

	More interestingly, Hibernate generates Java code for performing
object-relational persistence. Basically, for every entity in the model
it provides code for reading and writing objects to the database (and
methods can be added in UML for querying the database very easily).=20

	=20

	I found an article saying that Hibernate 3 also supports
object-XML persistence (http://www.devx.com/Java/Article/27896 and
http://www.hibernate.org/hib_docs/v3/reference/en/html/xml.html).
However, this feature seems to be in development and I'm not sure if
it's advanced enough for our purposes. So alternatively, I think the
Hibernate cartridges could be adapted fairly easily to generate code for
reading and writing XML as well.=20

	=20

	Ideally, the code should be modular so that we can read and
write XML to and from Java objects, and use this in conjunction with the
auto-generated Hibernate code, such that if developers wished they could
create their own in-house applications based on this framework. What do
people think?

	=20

	I've looked at the existing Java MAGEstk and I'm starting to
understand the basics. If anyone has an advice on how we might re-use
code from MAGEstk, in conjunction with what can be auto-generated by
AndroMDA, it would be really appreciated. I think we should be able to
get a basic API together fairly quickly, given that AndroMDA does a lot
for us automatically. I would like to encode as much of the logic as
possible in the UML and in the cartridges i.e. as little bespoke coding
as possible, such that the same process can be performed for FuGE, MAGE
2, PSI models and so on...=20

	=20

	Cheers

	Andy

[Mged-mage-java] STK and database schema for FuGE and extensions

[Andy J. <aj...@cs...>]

Hi all,

I've now completed a first pass at a software toolkit for FuGE milestone 2.
The code performs database persistence and writing out of valid FuGE XML
(see status below). Support for reading FuGE XML is pretty limited at this
stage but could be developed into a full parser fairly quickly I think. 

 

I'm not going to put any more time into it now until we are close to the UML
for FuGE version 1. The code is now posted on the FuGE site for anyone to
test out as a complete bundle (http://fuge.sourceforge.net/dev/index.php).
I've also pulled out the database schema for MySQL and the Hibernate mapping
file which are also posted as separate files. This means that theoretically
someone could use the code without using AndroMDA (I think.).

 

I've written some fairly detailed notes describing what I've done and how it
can be used, which is also posted on the website.

 

The status of the code is as follows:

-        Supports creating Java objects for every entity in FuGE milestone 2
(allows associations, attributes etc. to be set).

-        These objects can be stored, updated and deleted from a database.

-        Methods have been implemented that allow all Describable objects to
be retrieved from the database by their unique database ID (different from
the FuGE identifier) and then cast to the correct type.

-        An entity can be converted to a DOM element, and then written to a
file as XML. For instance if a FuGEroot element is retrieved, this will
process all associations to create a complete FuGE file, so long as all
associations have been created correctly.

-        The native XML produced is close but not an exact match of the FuGE
M2 XML Schema. Therefore, there are helper methods that perform this
conversion to output valid FuGE M2. If we make any major changes to the XML
Schema, these helper methods may need to be updated.

-        Reading XML is only supported so far in that an XML file can be
parsed, and the methods supplied can be used to create objects as required.
However, there is no code for processing objects and their associations in
the correct order etc. It would probably take a week or two of coding to
create a full FuGE parser. I'm not sure if the native XML persistence
offered by Hibernate can be made to work, I think it is really intended for
simpler schemas than FuGE.

 

It's only been tested with MySQL but the AndroMDA properties can be adapted
to work with Oracle or Postgres. The Oracle schema didn't work straight out
of the box, but I can probably generate one fairly easily if anyone wants to
try this out instead.

 

There are obviously still lots of unresolved issues and the code has not
been tested rigorously at all, so if anyone wants to contribute some time,
please get in touch.

Best wishes,

Andy

 

 

 

 

 

  _____  

From: mge...@li...
[mailto:mge...@li...] On Behalf Of Andy Jones
Sent: 10 March 2006 15:52
To: fug...@li...; 'mged-mage2';
mge...@li...
Subject: [Mged-mage2] RE: [Fuge-devel] STK and database schema for FuGE and
extensions

 

Hi all,

I thought I'd give a quick update on this issue. I've spent a few days
looking at Hibernate and AndroMDA and I think this is going to be a really
good way to go. I've managed to generate a framework from a cut down FuGE
(just FuGE Common). The framework creates Java objects for each entity and a
database schema. It provides methods for creating all Java objects,
associations and attributes, then writing or retrieving them from the
database. I don't think there will be any major issues scaling up to
complete FuGE (and then extensions). 

 

I've also started experimenting with the XML persistence in Hibernate 3.
This seems to work pretty well for the small tests I've performed so far,
although there will be a few issues to iron out. Basically, I can create
objects in the database, get them back as Java objects, which can then be
converted to a Dom4J Entity (in-memory XML nodes). The native mapping does
not quite match our XML Schema, but this can be customized pretty easily
through the AndroMDA cartridge that generates the Hibernate mapping file. I
should be able to get this sorted in a day or two.

 

If we customize the EntityToHibernate mapping files (rather than writing
bespoke code) we should be able to use the Hibernate framework also for
reading XML directly into Java objects, in other words I think we are quite
close to having a basic API for FuGE. 

 

A customized mapping could also be developed for a different database schema
relatively easily by manually altering one XML file, so the same API will
work with in-house databases rather than the auto-generated database schema.

 

The only thing that will be a bit of an issue is unique identifiers.
Hibernate seems to want to have integer IDs on every database table/entity,
and in the XML for referencing other items in the document. In FuGE we have
our own identification system. However, I think we can get around this by
having a server that stores/converts all database integer IDs and FuGE IDs
(although at the moment I have no idea how to make it work if FuGE documents
are split across files.).

 

I'm going to spend a few more days on this then I'll post it all on the FuGE
website, including instructions for how to setup and install AndroMDA and
Hibernate (which is the hardest part of this!). If anyone wants to
contribute some coding time, let me know, 

 

Cheers

Andy

 

 

 

 

 

  _____  

From: fug...@li...
[mailto:fug...@li...] On Behalf Of Andy Jones
Sent: 02 March 2006 11:32
To: fug...@li...; 'mged-mage2';
mge...@li...
Subject: [Fuge-devel] STK and database schema for FuGE and extensions

 

Hi all,

I've been taking a look at what else AndroMDA can do for us from the UML.
Using the Hibernate cartridges AndroMDA generates a fairly sensible looking
relational schema (for MySQL, Oracle or PostGres). So far I've only tested
with a simpler test model but I didn't encounter any problems that should
prevent this working with FuGE. I don't see that a relational schema will be
an essential part of FuGE but I think getting a relational schema for free
for FuGE and extensions is a bonus. 

 

More interestingly, Hibernate generates Java code for performing
object-relational persistence. Basically, for every entity in the model it
provides code for reading and writing objects to the database (and methods
can be added in UML for querying the database very easily). 

 

I found an article saying that Hibernate 3 also supports object-XML
persistence (http://www.devx.com/Java/Article/27896 and
http://www.hibernate.org/hib_docs/v3/reference/en/html/xml.html). However,
this feature seems to be in development and I'm not sure if it's advanced
enough for our purposes. So alternatively, I think the Hibernate cartridges
could be adapted fairly easily to generate code for reading and writing XML
as well. 

 

Ideally, the code should be modular so that we can read and write XML to and
from Java objects, and use this in conjunction with the auto-generated
Hibernate code, such that if developers wished they could create their own
in-house applications based on this framework. What do people think?

 

I've looked at the existing Java MAGEstk and I'm starting to understand the
basics. If anyone has an advice on how we might re-use code from MAGEstk, in
conjunction with what can be auto-generated by AndroMDA, it would be really
appreciated. I think we should be able to get a basic API together fairly
quickly, given that AndroMDA does a lot for us automatically. I would like
to encode as much of the logic as possible in the UML and in the cartridges
i.e. as little bespoke coding as possible, such that the same process can be
performed for FuGE, MAGE 2, PSI models and so on. 

 

Cheers

Andy

[Mged-mage-java] RE: [Fuge-devel] STK and database schema for FuGE and extensions

[Andy J. <aj...@cs...>]

Hi all,

I thought I'd give a quick update on this issue. I've spent a few days
looking at Hibernate and AndroMDA and I think this is going to be a really
good way to go. I've managed to generate a framework from a cut down FuGE
(just FuGE Common). The framework creates Java objects for each entity and a
database schema. It provides methods for creating all Java objects,
associations and attributes, then writing or retrieving them from the
database. I don't think there will be any major issues scaling up to
complete FuGE (and then extensions). 

 

I've also started experimenting with the XML persistence in Hibernate 3.
This seems to work pretty well for the small tests I've performed so far,
although there will be a few issues to iron out. Basically, I can create
objects in the database, get them back as Java objects, which can then be
converted to a Dom4J Entity (in-memory XML nodes). The native mapping does
not quite match our XML Schema, but this can be customized pretty easily
through the AndroMDA cartridge that generates the Hibernate mapping file. I
should be able to get this sorted in a day or two.

 

If we customize the EntityToHibernate mapping files (rather than writing
bespoke code) we should be able to use the Hibernate framework also for
reading XML directly into Java objects, in other words I think we are quite
close to having a basic API for FuGE. 

 

A customized mapping could also be developed for a different database schema
relatively easily by manually altering one XML file, so the same API will
work with in-house databases rather than the auto-generated database schema.

 

The only thing that will be a bit of an issue is unique identifiers.
Hibernate seems to want to have integer IDs on every database table/entity,
and in the XML for referencing other items in the document. In FuGE we have
our own identification system. However, I think we can get around this by
having a server that stores/converts all database integer IDs and FuGE IDs
(although at the moment I have no idea how to make it work if FuGE documents
are split across files.).

 

I'm going to spend a few more days on this then I'll post it all on the FuGE
website, including instructions for how to setup and install AndroMDA and
Hibernate (which is the hardest part of this!). If anyone wants to
contribute some coding time, let me know, 

 

Cheers

Andy

 

 

 

 

 

  _____  

From: fug...@li...
[mailto:fug...@li...] On Behalf Of Andy Jones
Sent: 02 March 2006 11:32
To: fug...@li...; 'mged-mage2';
mge...@li...
Subject: [Fuge-devel] STK and database schema for FuGE and extensions

 

Hi all,

I've been taking a look at what else AndroMDA can do for us from the UML.
Using the Hibernate cartridges AndroMDA generates a fairly sensible looking
relational schema (for MySQL, Oracle or PostGres). So far I've only tested
with a simpler test model but I didn't encounter any problems that should
prevent this working with FuGE. I don't see that a relational schema will be
an essential part of FuGE but I think getting a relational schema for free
for FuGE and extensions is a bonus. 

 

More interestingly, Hibernate generates Java code for performing
object-relational persistence. Basically, for every entity in the model it
provides code for reading and writing objects to the database (and methods
can be added in UML for querying the database very easily). 

 

I found an article saying that Hibernate 3 also supports object-XML
persistence (http://www.devx.com/Java/Article/27896 and
http://www.hibernate.org/hib_docs/v3/reference/en/html/xml.html). However,
this feature seems to be in development and I'm not sure if it's advanced
enough for our purposes. So alternatively, I think the Hibernate cartridges
could be adapted fairly easily to generate code for reading and writing XML
as well. 

 

Ideally, the code should be modular so that we can read and write XML to and
from Java objects, and use this in conjunction with the auto-generated
Hibernate code, such that if developers wished they could create their own
in-house applications based on this framework. What do people think?

 

I've looked at the existing Java MAGEstk and I'm starting to understand the
basics. If anyone has an advice on how we might re-use code from MAGEstk, in
conjunction with what can be auto-generated by AndroMDA, it would be really
appreciated. I think we should be able to get a basic API together fairly
quickly, given that AndroMDA does a lot for us automatically. I would like
to encode as much of the logic as possible in the UML and in the cartridges
i.e. as little bespoke coding as possible, such that the same process can be
performed for FuGE, MAGE 2, PSI models and so on. 

 

Cheers

Andy

[Mged-mage-java] Re: [Fuge-devel] STK and database schema for FuGE and extensions

[Junmin L. <ju...@pc...>]

Hi, Andy,
Those features in the MDA are such appealing. But I wonder in practice how 
they work for us.

> I found an article saying that Hibernate 3 also supports object-XML
> persistence (http://www.devx.com/Java/Article/27896 and
> http://www.hibernate.org/hib_docs/v3/reference/en/html/xml.html). However,
> this feature seems to be in development and I'm not sure if it's advanced
> enough for our purposes. So alternatively, I think the Hibernate cartridges
> could be adapted fairly easily to generate code for reading and writing XML
> as well.

There is also a hyperJAXB tool coming from JAXB side. I was thinking about 
using it to load some of the proteomics data into database a while ago.

---junmin

[Mged-mage-java] Re: [Mged-mage2] STK and database schema for FuGE and extensions

[Jason S. <jas...@gm...>]

Hi Andy,

I haven't butted in much lately - so let me introduce myself. I help
build MAGEstk we back when we were getting the first MAGE object model
put together. I've been asked by MGED to help continue MAGEstk work to
keep up with the new developments with MAGE and FuGE.

 I think you beautiful summarized my goal with MAGEstk:

> I would like
> to encode as much of the logic as possible in the UML and in the cartridg=
es
> i.e. as little bespoke coding as possible, such that the same process can=
 be
> performed for FuGE, MAGE 2, PSI models and so on=85

When we started MAGEstk there where no Open Source UML compilers that
could produce all we needed - which then was minimally a Perl and Java
and XML/DTD language bindings. Since then things have changed quite a
bit and there are a lot more choices for tools and technologies.

It is possible that we will completely throw away all the old MAGEstk
codebase and rebuild it using better tools - I know the Perl code
sucks (I wrote it), and the java code can also be better written
nowadays. Perhaps as you suggest AndroMDA can do much of the work for
us - Perl and Python support would be nice to have, tho.

I've used Hibernate3 on some java projects over the past year, and I
think it's a great technology.  I'm still learning how best to use it
in production systems, it's a big system - so I am no expert.

It will take me a bit to get up to speed again with all the changes,
so please forgive me if I ask many silly questions.

Cheers,
jas.

[Mged-mage-java] STK and database schema for FuGE and extensions

[Andy J. <aj...@cs...>]

Hi all,

I've been taking a look at what else AndroMDA can do for us from the UML.
Using the Hibernate cartridges AndroMDA generates a fairly sensible looking
relational schema (for MySQL, Oracle or PostGres). So far I've only tested
with a simpler test model but I didn't encounter any problems that should
prevent this working with FuGE. I don't see that a relational schema will be
an essential part of FuGE but I think getting a relational schema for free
for FuGE and extensions is a bonus. 

 

More interestingly, Hibernate generates Java code for performing
object-relational persistence. Basically, for every entity in the model it
provides code for reading and writing objects to the database (and methods
can be added in UML for querying the database very easily). 

 

I found an article saying that Hibernate 3 also supports object-XML
persistence (http://www.devx.com/Java/Article/27896 and
http://www.hibernate.org/hib_docs/v3/reference/en/html/xml.html). However,
this feature seems to be in development and I'm not sure if it's advanced
enough for our purposes. So alternatively, I think the Hibernate cartridges
could be adapted fairly easily to generate code for reading and writing XML
as well. 

 

Ideally, the code should be modular so that we can read and write XML to and
from Java objects, and use this in conjunction with the auto-generated
Hibernate code, such that if developers wished they could create their own
in-house applications based on this framework. What do people think?

 

I've looked at the existing Java MAGEstk and I'm starting to understand the
basics. If anyone has an advice on how we might re-use code from MAGEstk, in
conjunction with what can be auto-generated by AndroMDA, it would be really
appreciated. I think we should be able to get a basic API together fairly
quickly, given that AndroMDA does a lot for us automatically. I would like
to encode as much of the logic as possible in the UML and in the cartridges
i.e. as little bespoke coding as possible, such that the same process can be
performed for FuGE, MAGE 2, PSI models and so on. 

 

Cheers

Andy

[Mged-mage-java] Measurement inner class enumerations

[Joe W. <jw...@ji...>]

Hi Folks,

Can anyone send me a snippet of code showing how to set the KindCV and 
Type attributes of a Measurement object?  This is the most 
counterintuitive part of MAGE-Java--and something I'ld like to see 
chaanged to an enumeration in Measurement (rather than an inner class). 

For the time being, I just need to find out how to use the damn thing.  
Can anyone help? 

Thanks, and please parden my French.

Joe White

[Mged-mage-java] Does MAGEReader support multiple xml source files?

[Juan E. <jua...@gm...>]

Hi everyone!
I would like to know if the current JAVA MAGE-stk 2004-11-20 v1.1 is
able to handle the scenario where different xml documents contains the
MAGE-ML experiment information and if the MAGEReader is able to
resolve references from one document to the other. For instance, one
document contains experiment_package, bioassay_package, etc, and the
other one contains the protocol_package.

regards
--Juan Elvira

Re: [Mged-mage-java] MAGE relations

[<ko...@vb...>]

Something like this ....

Person p = new Person();
Experiment e = new Experiment();
e.addToProviders(p);

Hope this helps ....

>
>
> I'm starting using the MAGEstk package and although and already able to
> build a few examples some doubts still remain.
>
>
>
> My actual problem is how i can represent packages relations in the
> MAGE-ML.
> For example, using the MAGEstk, how can i relate one experiment with one
> person?
>
>
>
>
>
>
>
> Person p = new Person();
>
> (add information about the person)
>
>
>
> Experiment e = new Experiment();
>
> (add information about the Experiment)
>
>
>
> (Relate the experiment with the person)
>
>
>
>
>
>
>
> Thanks in advance,
>
>
>
> Joel Arrais
>
>

[Mged-mage-java] MAGE relations

[Joel A. <jp...@ie...>]

 

I'm starting using the MAGEstk package and although and already able to
build a few examples some doubts still remain.

 

My actual problem is how i can represent packages relations in the MAGE-ML.
For example, using the MAGEstk, how can i relate one experiment with one
person?

 

 

 

Person p = new Person();

(add information about the person)

 

Experiment e = new Experiment();

(add information about the Experiment)

 

(Relate the experiment with the person)

 

 

 

Thanks in advance,

 

Joel Arrais

[Mged-mage-java] MAGEstk installation problems

[joelarrais <joe...@gm...>]

I'm trying to install and use the java version of the MAGEstk but I'm facing
some problems. Does anyone have a manual that I can follow to get the
MAGEstk work.

 

Thanks in advance.

 

 

[Mged-mage-java] MAGEstk installation

[joelarrais <joe...@gm...>]

I'm trying to install and use the java version of the MAGEstk but I'm facing
some problems. Does anyone have a manual that I can follow to get the
MAGEstk work.

 

Thanks in advance.

 

Re: [Mged-mage-java] xslt for MAGE

[<ia...@co...>]

Juan,

Attached is a simple file that extracts experiment details from a MAGE-ML file (if they are present) to a text file. But as Angel suggests the possibilities of what you can do are enormous and really require an understanding of MAGE-ML structure.

XMLSpy has a graphical tool for creating XSL transformations. The cost may be justified by the time it saves you. The XSL that it produces might not be what one might choose if one did it manually. It works from an XML Schema the last time I looked at it - but it would help you create that from the MAGE-ML DTD. There are other such tools that can be found by searching the web.

Ian


> Keep dreaming Juan ;)
> Seriously though, if your issue is learning XSLT & XPATH, then here is a 
> quick template you can follow to learn, but is not related to MAGE-ML. 
> For that, you must learn the MAGE model as well as XSLT and XPATH.
> -angel
> 
> 
> 
> Juan Elvira wrote:
> 
> > Hi everyone!
> > I need to transform a MAGE xml document into a HTML.
> > I would appreciante if  anybody could provide me with a basic XSLT 
> > file for MAGE  to do the transformation or maybe  recommend some 
> > application to edit o generate the XSLT files, in an easy way.
> >
> >
> > best regards
> > //Juan
> >
> >
> > -------------------------------------------------------
> > This SF.Net email is sponsored by Oracle Space Sweepstakes
> > Want to be the first software developer in space?
> > Enter now for the Oracle Space Sweepstakes!
> > http://ads.osdn.com/?ad_id=7393&alloc_id=16281&op=click
> > _______________________________________________
> > Mged-mage-java mailing list
> > Mge...@li...
> > https://lists.sourceforge.net/lists/listinfo/mged-mage-java
> >
> 
> -- 
> 
> Angel Pizarro
> Director, Bioinformatics Facility
> Institute for Translational Medicine and Therapeutics
> University of Pennsylvania
> 806 BRB II/III
> 421 Curie Blvd.
> Philadelphia, PA 19104-6160
> 
> P: 215-573-3736
> F: 215-573-9004
> 
> <?xml version="1.0" encoding="utf-8"?>
> <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"  version="1.0">
> <xsl:output method="xml" indent="yes" />
> <xsl:template match="/">
> 
>   <xsl:for-each select="GBSet/GBSeq">
>    <xsl:variable name="id"  select="GBSeq_locus" /> 
>    <xsl:for-each select="GBSeq_feature-table/GBFeature[GBFeature_key = 'CDS']" >
>      <xsl:for-each  select="GBFeature_quals/GBQualifier[GBQualifier_name = 
> 'note' ]" >
>        <xsl:variable name="notes" select="GBQualifier_value" />"<xsl:value-of 
> select="$id"/>"  "<xsl:value-of select="$notes"/>"
>      </xsl:for-each>
>    </xsl:for-each>
>   </xsl:for-each>
> </xsl:template>
> </xsl:stylesheet>

Re: [Mged-mage-java] xslt for MAGE

[Angel P. <an...@ma...>]

Keep dreaming Juan ;)
Seriously though, if your issue is learning XSLT & XPATH, then here is a 
quick template you can follow to learn, but is not related to MAGE-ML. 
For that, you must learn the MAGE model as well as XSLT and XPATH.
-angel



Juan Elvira wrote:

> Hi everyone!
> I need to transform a MAGE xml document into a HTML.
> I would appreciante if  anybody could provide me with a basic XSLT 
> file for MAGE  to do the transformation or maybe  recommend some 
> application to edit o generate the XSLT files, in an easy way.
>
>
> best regards
> //Juan
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by Oracle Space Sweepstakes
> Want to be the first software developer in space?
> Enter now for the Oracle Space Sweepstakes!
> http://ads.osdn.com/?ad_id=7393&alloc_id=16281&op=click
> _______________________________________________
> Mged-mage-java mailing list
> Mge...@li...
> https://lists.sourceforge.net/lists/listinfo/mged-mage-java
>

-- 

Angel Pizarro
Director, Bioinformatics Facility
Institute for Translational Medicine and Therapeutics
University of Pennsylvania
806 BRB II/III
421 Curie Blvd.
Philadelphia, PA 19104-6160

P: 215-573-3736
F: 215-573-9004

[Mged-mage-java] xslt for MAGE

[Juan E. <jua...@in...>]

Hi everyone!
I need to transform a MAGE xml document into a HTML.
I would appreciante if  anybody could provide me with a basic XSLT file 
for MAGE  to do the transformation or maybe  recommend some application 
to edit o generate the XSLT files, in an easy way.


best regards
//Juan

[Mged-mage-java] (no subject)

[Juan E. <jua...@in...>]

jua...@gm...

[Mged-mage-java] "argument type mismatch" in MAGEContentHandler.makeAssociation

[Paul P. <pav...@db...>]

I'm using the MAGEContentHandler with code pretty much copied from the MAGEReader and am 
having problems. I'm using the 1.1 version.

I've been getting the above error with MAGE-ML files that seem valid and are of recent 
vintage, when they contain an ArrayDesign package. My debugger tells me that the state of 
the program includes the following setup for the call at line 606 of the MAGEContentHandler:

(which reads: set_method.invoke( parent, new Object[] { child } );)

parent is a CompositeSequence
child is a BioSequence
method_name is "addToReporterCompositeMaps"

In other words, a call is being made like 'addToReporterCompositeMaps(BioSequence)', which 
is not valid, the argument ('child') should of course be a ReporterCompositeMap.

I'm sure there is a simple explanation for what I am doing wrong.

A related question is: where can I find some substantial (e.g., non-trivial) _valid_ 
example documents to use as tests. The documents on the MAGE site seem to be old and 
invalid (how confusing!). The EBI 'best practice' files are too small to be really useful 
for unit tests etc., and 'real' files from ArrayExpress are too big for that purpose. I'd 
appreciate any pointers.

Thanks,
Paul