Nathan, I've answered a little bit, but I need to forward this to ethan to
answer most of these questions, I hope he's watching his email to answer soon.
Nathan Potter wrote:
John,
So I made this catalog.xml:
<?xml version="1.0" encoding="UTF-8"?>
<catalog name="TestData_NC" version="1.0.1"
xmlns="http://www.unidata.ucar.edu/namespaces/thredds/
InvCatalog/v1.0">
<service name="TestData" serviceType="OpenDAP" base="/thredds/
contextPath/"/>
<datasetScan location="/nc" path="nc" name="Test netCDF Data"
serviceName="OpenDAP">
<crawlableDatasetImpl className="opendap.olfs.S4CrawlableDataset">
</datasetScan>
</catalog>
And I have questions:
- In the <service> element what is the attribute "base" set to, the
location of the servlet:
$CATALINA_HOME/webapps/myservlet
Or is it set to:
$CATALINA_HOME/content/thredds
The service base is the (first part of the) external URL of the dataset, that is the URL
that your clients will see. Since this determines which servlet gets called when it comes
into Tomcat, you need to make sure it matches. I prefer using
"/contextName/servletMapping/", which then makes it reletive to the catalog
URL. The contextName matches your war file, and the servletMapping matches what you put
into the web.xml. So if your war file is opendap.war, and you have
<servlet-mapping>
<servlet-name>DefaultHandlerServlet</servlet-name>
<url-pattern>/*</url-pattern>
</servlet-mapping>
in your web.xml,
then use /opendap/ as the service base, and the DefaultHandlerServlet will bet called when any URLS of the form
http://servername:8080/opendap/
get sent.
- Given that I identify my implementation of CrawlableDataset to the
THREDDS code as above, how do I get one built? The constructor requires
both a path and a config object, and that config object is something I
have and THREDDS doesn't. Seems like I either need to pass an instance
of my CrawlableDataset, or my config object, into CatalogRootHandler.
I think the code will instantiate an object for you.
ethan?
- Logging: I read the basic docs on the Log4j and the Simple Logging
Facade. I like this. I don't understand how to use it. I am concerned
that my hacks to get THREDDS into my servlet don't address a number of
startup issues surrounding logging that THREDDS may require. You're
ThreddsDefault.makeDebugActions() is pretty involved, and I don't
really get it.
There's a question of how much of the thredds framework you want to use. If you
want to use all of it, thats great, we can leverage each others work. The
DebugActions allow remote testing and info. Very useful. I can give you
permssion to look at how the TDS on motherlode works if you want. Anyway, you
need to decide whether to use all or a part of it.
The SLF/Log4j is worth useing in any case.
- Caching: It looks like caching is an integral part of your code base.
Is it isolated in the servlet related code, or are there
implementations of caching related activities in the catalog stuff that
I need to be aware of?
Im using caching deep in my data server, but not in the servlet handling part.
So you still have a choice.
I am NOT talking to a filesystem. I can't give THREDDS a file root, or
a file type/format. I can give provide a CrawlableDataset instance that
can traverse the data store. I can hand you all of the traditional
OPeNDAP/DODS responses, although from what I got from Ethan's
description once I get the catalog working I would have my servlet to
respond to a DODS request for, say a DDS, by calling
CatalogRootHandler.findRequestedDataset(...) for the data set, then if
a CRawlableDataset comes back (indicating that this data set exists and
is not filtered) I can use the CrawlableDataset to do what I need to
get the stuff to send to the client. Is that true?
We undersstand that, thats what CrawlableDataset does for you.
Ethan can you explain his next steps?
Lets just incrementally get this working one step at a time. First thing is to get catalogs generated using CrawlableDataset.
Dont worry, we will get all this working ASAP.
