Read Me

The MCAS Project ( Metrics Correlation and Analysis service ) provides integral solution for system operators or VO users to uniformly access, transform and correlate disjoint data sources or metrics. As part of the goal to correlate and summarize information, the scope of the project also includes tools and services for efficient refactoring of existing informational portals.

Env variables required for building entire project:
JBOSS_HOME , example  /home/abaranov/jboss-portal-2.7.0.GA
BERKELEY_DB , example  /home/abaranov/dbxml-2.5.13/install

(unfortunately, build process requires exact, dbxml-2.5.13 version of BerkleyDB )

targets:
unpack,deploy,portlets,admin,warehouse,services.mule

Note that admin,warehouse,services.mule targets implicitly depend on
unpack.Thus in order to build admin,warehouse,services.mule targets
incrementally, unpack target must be invoked by hand once after svn co. 

Mule transformers are built under services.mule target.
build scripts will create build directory where all war, sar and jar files
will be placed at the directory root level
ant deploy target will unpack mule into build/rpm directory, compile and
deploy all work flow and warehouse related codes such that build/rpm can be packaged into
functioning rpm target - workflow engine, warehouse and data source api frontends

--------------
** Installation of the Mule based data WareHouse 

- deploy the berkeley db on a system directory 
download the source code, e.g. for v2.5.13, from
http://www.oracle.com/technology/software/products/berkeley-db/xml/index.html

Follow the build instruction. In short, from the untarred directory, do
> ./buildall.sh --enable-debug --prefix=/scratch/dbxml-2.5.13/ --enable-java 
This will install the code in /scratch/dbxml-2.5.13/

ant deploy

cd build/rpm/

. ./setup.sh

configure.sh

data-source-admin --add=./bin/testInput.ds 

deployConfig.sh

startMcas.sh &

Write a table with
> curl --data-binary @./tbl.xml http://localhost:9090/test/Input/Sink
Where tbl.xml is a file like
<table>
 <row>
   <cell>andrew</cell>
   <cell>bar</cell>
   <cell>naper</cell>
   <cell>rhodes</cell>
   <cell>630</cell>
   <cell>840</cell>
 </row>
  <row>
   <cell>jandrew</cell>
   <cell>12bar</cell>
   <cell>45naper</cell>
   <cell>66rhodes</cell>
   <cell>630</cell>
   <cell>840</cell>
 </row>
</table>

Get the latest document
> curl http://localhost:9090/test/Input/Get

Query the whole table with
> curl --data-binary '_doc_//table' 'http://localhost:9091/xmldb/query/Input'

Query all 
> curl --data-binary '_doc_/p/table/row/cell[ . = "andrew" and position() = 1 ]/..' 'http://localhost:9091/xmldb/query/Input'

--------------
** Building and testing of the Data Source Admin tool
- Modify the setup script to point to your installation of the products.
See documentation in the script itself for details of what to change

- Source setup.sh

- From the root of the MCAS distribution:
> ant admin
The mcas-data-source-admin.jar is stored at modules/admin/lib/
Should you want to change the amount of debugging output information, edit 
modules/admin/lib/log4j.properties

- Run the test of the tool. 
-- Edit and customize the configuration file at $MCAS_DS_ADMIN_CONFIG
-- Create the directory for the local file system storage (to store 
transformation and validation files)
-- From the root of the MCAS distribution:
> cd modules/admin/test/
> ./data-source-test.sh -clean # NOTE: clean up ALL files in the database