Esri/geoportal-server-harvester

Automatically harvest upon detected changes or new XML

Closed this issue · 3 comments

Automating the process of harvesting would be great. Would it be possible to connect harvester to a script that detects and changes in the folder?

For example if a new XML file is added to a harvested folder it would automatically detect it and begin harvesting.

Yes (but it's not there yet). The basic idea is that scheduling is also a broader concept based on triggers paradigm. One type of the trigger is such an object which starts harvesting according to the elapsed time (as available now). Another is a listener to the changes of the observed folder; once the content of the folder has changed it starts harvesting. Java has already everything to implement such an observer so implementing trigger wouldn't be that hard.

That's great to hear! Any hints on when an folder observer trigger would be implemented?

If it is quite a bit down the line would you be able to point me to which files would need to be changed with Java so that I could try this in our environment? Thanks.

Latest release of the Harvester has a "Sink" input broker. This input broker allows to declare a designated folder to be a drop folder where harvester will observer and listen to any new items in it. When combined into a task it allows to publish any xml metadata into the output broker. To get it working, a task must be started (this task is a perpetual task; it stays active until an operator cancels it). A metadata could be dragged and dropped into that folder (or copied); after publishing into the destination, metadata will be deleted from the sink.