>> Just wanted to point out that you should not have the start up overhead if >> you wrap FOP in a servlet
Yes that's exactly what I was recommending (and did). Using a servlet or some equivalent technology that is running in a 'always on' JVM or at least one that hangs around for a while ( in case your load is occasional spikes you can use a simple start-on-demand-but-wait-a-while server) ----------------------------------------------------------------------------- David Lee Lead Engineer MarkLogic Corporation [email protected] Phone: +1 812-482-5224 Cell: +1 812-630-7622 www.marklogic.com<http://www.marklogic.com/> From: [email protected] [mailto:[email protected]] On Behalf Of Florent Georges Sent: Wednesday, September 02, 2015 8:52 AM To: MarkLogic Developer Discussion <[email protected]> Subject: Re: [MarkLogic Dev General] Generate PDF from Marklogic Hi David, Just wanted to point out that you should not have the start up overhead if you wrap FOP in a servlet, as the JVM is not loading the classes over and over again, it's done only once. From experience, the critical part is to connect the various components properly (that is, the servlet HTTP layer and FOP). I've seen a lot of examples on the web serializing XML as a String to parse it again, or saving FOP result in a byte array to pass it as a result of the servlet, instead of streaming properly between HTTP and FOP. Regards, -- Florent Georges http://fgeorges.org/ http://h2oconsulting.be/ On 2 September 2015 at 14:20, David Lee wrote: I've done the same thing for PDFS. For a previous $mployeer I used a locally running embedded tomcat server with a simple servlet that ran a simple xmlsh script (could be any language) that ran the FOP processing and then returned the PDF as a binary. If you need high throughput, its critical to not have to start heavyweight processes for each invocation. Many FOP implementations are Java based which have a high startup time (200ms+). That's fine for 10 jobs not 10,000 ... hence using a service that keeps the FOP processor 'loaded'. For high performance, or maximizing licensing value you may want to move the FOP processing off-server -- Modern alternatives may appropriate, such as a pool of Docker and/or web services shared by all the ML servers to offload the FOP processing and scale it independent of the ML processing needs. ----------------------------------------------------------------------------- David Lee Lead Engineer MarkLogic Corporation [email protected]<mailto:[email protected]> Phone: +1 812-482-5224<tel:%2B1%20812-482-5224> Cell: +1 812-630-7622<tel:%2B1%20812-630-7622> www.marklogic.com<http://www.marklogic.com/> From: [email protected]<mailto:[email protected]> [mailto:[email protected]<mailto:[email protected]>] On Behalf Of vimal c Sent: Wednesday, September 02, 2015 2:09 AM To: MarkLogic Developer Discussion <[email protected]<mailto:[email protected]>> Subject: Re: [MarkLogic Dev General] Generate PDF from Marklogic Thanks Florent and David for your inputs. Regards, Vimal C On Tue, Sep 1, 2015 at 8:20 PM, David Ennis <[email protected]<mailto:[email protected]>> wrote: I've done something similar to what Florent suggests in the past with the Java flying saucer Library (XML or XHTML + CSS2 = PDF). I think no matter how you wrap it up and what the details are, its the same pattern used over and over with MarkLogic for external functionality - create an HTTP service in the language of your choice to do the thing you want done and post to that service. Kind Regards, David Ennis On 1 September 2015 at 16:43, Florent Georges <[email protected]<mailto:[email protected]>> wrote: Hi, As far as I know, there is no built-in way. I think the usual way is to use XSLT on MarkLogic to generate XSL-FO, and use xdmp:http-post() to send it to an endpoint an XSL-FO processor listens to (typically Apache FOP, wrapped in Cocoon or in an in-house Java HTTP endpoint). This setup is reasonably easy. I guess searching for "fop marklogic" should give a few links about that technique. Regards, -- Florent Georges http://fgeorges.org/ http://h2oconsulting.be/ On 1 September 2015 at 16:00, vimal c wrote: Hi All, Is it possible to generate .pdf file from marklogic? I see that we can convert a pdf document to xhtml files and parts by using xdmp:pdf-convert. But Do we have any way in marklogic to create a .pdf file out of the xml content saved in marklogic? Any idea would be really helpful. Thanks in advance. Regards, Vimal C _______________________________________________ General mailing list [email protected]<mailto:[email protected]> Manage your subscription at: http://developer.marklogic.com/mailman/listinfo/general -- <http://developer.marklogic.com/mailman/listinfo/general> Florent Georges <http://developer.marklogic.com/mailman/listinfo/general>http://fgeorges.org/ http://h2oconsulting.be/ _______________________________________________ General mailing list [email protected]<mailto:[email protected]> Manage your subscription at: http://developer.marklogic.com/mailman/listinfo/general _______________________________________________ General mailing list [email protected]<mailto:[email protected]> Manage your subscription at: http://developer.marklogic.com/mailman/listinfo/general _______________________________________________ General mailing list [email protected]<mailto:[email protected]> Manage your subscription at: http://developer.marklogic.com/mailman/listinfo/general
_______________________________________________ General mailing list [email protected] Manage your subscription at: http://developer.marklogic.com/mailman/listinfo/general
