Hi Faranak, I'm a bit confused by your email because it sounds as though this question is more for the constraints of the ruby middleware you are interacting with. There's no reason the crawler couldn't send the metadata to the ruby middleware as a post ingestion success action or that the ruby middleware couldn't read the metadata from the filemgr. Writing the metadata file to disk seems like an unnecessary step but I could be wrong if those are system requirements. That said the ruby middleware can query the filemgr for the product metadata via xmlrpc or via REST (if you install the curation webapp which comes with a set of REST services). The answer really depends on your project system requirements.
I believe there is documentation to the REST services I'm talking about but I forget the link so hopefully someone will chime in. Although from what I can gather from your statement it sounds like you may want to try writing a CrawlerAction. In which case you also may want to write a failure action too as I'm assuming that the ruby middleware should know when stuff has gone awry. I've done a very similar thing on another project and simply used the Apache HTTP client API within a crawler action to interact with a REST service. Thanks, Paul Ramirez On Jan 10, 2012, at 2:46 PM, Davoodi, Faranak (388J) wrote: > Hi All, > > I have a system that consists of a crawler and a filemgr that reads and > extracts few metadata from the files submitted to the staging area, then > submits the extracted metadata keys and their values over to a ruby > middleware. The system is already kind of slow. I am wondering if sending the > metadata keys and values via a set of restful commands are faster or leaving > the metadata file somewhere for the middleware to grab. Let me know what you > think. > > Thanks, > Faranak
