I am wondering if anyone has used the Elastic Search opensource product with a Laserfiche repository before? If so, could you give some details of how you got it to work?
Has anyone used Elastic Search with a Laserfiche repository?
Could you provide more context behind this inquiry? What would you be aiming to accomplish by integrating Elastic Search?
The concept is much like Federated Search, but using Elastic Search instead. It wouldn't require the end-users to have a Laserfiche license.
I see. So if Elasticsearch somehow indexes a Laserfiche repository, and there is some UI hooked up to it, what is the search returning that gets you out of end users needing a Laserfiche license (in one form or another)?
If it returns web client URLs, users still need a license to actually access the docs.
If it somehow returns the actual repository files, you're violating the Laserfiche license agreement by circumventing licensing users unless you have an integration license for the use case (which is just a different way of licensing the users).
If it returns WebLink URLs, that's fair game, but again, just a different way of licensing the users (via the Public Portal license pool).
I believe the client's goal is to be able to return the WebLink URLs. They are already using Elastic Search for their website and it is tied into other resources they own, so they are just trying to see if they could also integrate it with Laserfiche in some way. The licensing questions are something we have brought to their attention, as well as security concerns.
I should also note that they own the SDK and are interested in ways to use it to accomplish what they need. They are wary of just crawling the WebLink site because of issues with it possibly taking up licensing and causing slowness.
In that case, I'd enable the Performance Counters for public portal license consumption in LFS and run a test crawl after-hours. That's by far the most straightforward thing to try and it's easy to see what kind of impact is has on portal license consumption and LFS and WebLink performance.
If they know exactly what content they want to surface through the search, it's not difficult to write a Workflow/script/etc. to construct WebLink URLs and shovel them plus any relevant metadata into an Elasticsearch index.
We will pass that information to the client. Thank you for the feedback!