Content MS - Architecture/Implementation

Question

Hi
Can someone point me to a document which explanin the underlying architecture
of the CMS. Whethr it uses Ejbs/classes/is the Data cached for the Content Operation
(add/delete). I want to write my own interface for BEA Content Management (do
not want to use BulkLoader OR myContent/Portal App tool). I am planning to develop
my own server side classes using the Content API/classes but before that wanted
to undestand the underlying arch of the CMS.
Any help would be highly appreciated
Pete

Answer 1

pete, below is an example of how to create a content node, with a binary property.
it also includes creating the content type (objectclass), which i don't believe
you will need to do. hopefully this will help.
repositoryMgr.connect();
NodeOps nops = repositoryMgr.getNodeOps();
ObjectClassOps ocops = repositoryMgr.getObjectClassOps();
ID repId = new ID();
repId.setRepositoryName(repositoryName);
PropertyDefinition[] pds = new PropertyDefinition[1];
pds[0] = new PropertyDefinition(repId, "binaryType", null, Property.BINARY,
false, false, false, false, false, null);
ObjectClass oc = new ObjectClass(repId, "schema", null, pds);
oc = ocops.createObjectClass(oc);
Node newNode = nops.createHierarchyNode(repId, "A");
Property[] props = new Property[1];
BinaryValue bv = new BinaryValue(mimeType, fileName, size, inputStream);
props[0] = new Property("binaryType", new Value(bv));
Node updatedNode = nops.addNodeContent(newNode.getId(), oc.getId(),
props);
"pete" <[email protected]> wrote:
>
James
Thanx for your reply.
I have been trying to take a look at the myContent portlet code but to
be very
honest with you it is too cluttered for me and I could not follow the
jpf code
however hard I try.
It would be extremely helpful to me if you can provide some sample code
which
just connects to a repository and then does some Add/Update function
on the CMs
tables.
Thanks, Pete
"James Owen" <[email protected]> wrote:
Pete, ok, so you want to configure your own repository (using the Portal
CM Repository),
then from a jsp, call your own classes that call the cm api, which will
store/retrieve
the content in the Portal CM Repository.
In that case, you don't need to do anything with the SPI, you'll just
be using
the CM API.
Just to clarify, when you "create" your own repository (using the Portal
CM Repository),
it doesn't really create a Repository, it just configures it, more like
a jdbc
connection. So if you delete the BEA Repository and then create your
own, using
the same Repository class (shift plus / will fill it in for you) your
newly configured
repository will point to the same tables that the BEA Repository did,
so it is
more like a rename.
For api examples, I would suggest you take a look at the sample cm portlet
code
- not for best practices ;) - but it does give good examples of howto
use the
api to create content, search, etc. if it is doesn't address your specific
questions
then let me know.
James
"pete" <[email protected]> wrote:
James, Thanks for the answers, Your Question
sorry, i'm not following what you would like to do. if you are developing
your
own ui, then all you'll just need to use the API, if you would liketo
expose
your own repository to the Virtual Repository then you will need toimplement
the SPI. what exactly is your goal?My goal is to have a UI(JSP) which in turn will call some functionsin
a Server
Side class for CRUD operation on a new Repository(with my <Application
name> Repository
instead of BEA Repository) in BEA PORTAL DB. Currently I do not have
my Virtual
repository, I just will use the BEA's virual repository and either
rename
it to
<app>Repository OR will create a new repository in the BEA Portal DB.
Do you think
If I just write some functions in a server side class for Content Oprations
using
the API that should suffice OR I will have to implement SPI
Also it would be extremely helpful if you could provide me with some
sample code
for connecting to a repository in BEA PORTAL DB and some function like
adding/updating
content in the that repository. You have been very helpful so far,thanks
a bunch
"James Owen" <[email protected]> wrote:
"pete" <[email protected]> wrote:
James
Thanks for the diagram.
1. From this it becomes clear that the Content SPI is a combinationof
Entity
and Session beans and the property set is a Facade pattern, am I
correct
in my
understanding?
this is true of the BEA Content Repository, which implements the Content
SPI.
2. If the Content Database it implemented as a Entity Bean (probably
CMP) then
is it safe to assume that it is cached?Yes, it is CMP and yes, the beans are cached with the default setting
(1000).
You can up the settings by cracking open the descriptor in content_repo.jar.
3. what is the diff between BEA content SPI and Content SPI I am
planning
to develop
the server side classes for Adding/Updating/Deleting Content in BEACMS
(developing
my own methods using content API), do you think it is safe to do
it
this
way.sorry, i'm not following what you would like to do. if you are developing
your
own ui, then all you'll just need to use the API, if you would liketo
expose
your own repository to the Virtual Repository then you will need toimplement
the SPI. what exactly is your goal?
Pete
"James Owen" <[email protected]> wrote:
Pete,
here is a diagram showing the architecture. are you planning on
implementing
the
SPI? if so, i submitted a doc on dev2dev that describes the SPI.it
hasn't
been
posted yet, so i can send it to you separately.
James
"pete" <[email protected]> wrote:
Hi
Can someone point me to a document which explanin the underlying
architecture
of the CMS. Whethr it uses Ejbs/classes/is the Data cached for
the
Content
Operation
(add/delete). I want to write my own interface for BEA Content
Management
(do
not want to use BulkLoader OR myContent/Portal App tool). I amplanning
to develop
my own server side classes using the Content API/classes but beforethat
wanted
to undestand the underlying arch of the CMS.
Any help would be highly appreciated
Pete