hi processes gitter, does the group have ideas on the concept of job runs, can a job be configured to run at intervals, and can the results of individual job runs be listed and retrieved?
or is a job actually intended as a 'run' of a process? I'd see a job more as a specific configuration of a process, e.g process is calculate-height, job would be calculate height of Mont Blanc, runs would be height in 1950, 1980, 2000 and 2020...
Francis Charette Migneault
Hi. The current definition of Job in OGC API Processes is more to represent an execution run of a process. The results of that job can be retrieved with this method: https://docs.ogc.org/DRAFTS/18-062.html#sc_retrieve_job_results Processes are typically intended for single run per call, but you could have a process implementation that has for execution steps to start further processes with different values, or simply call the desired process multiple times with each of the desired values.
Hi, thinking out loud about a new API we want to implement - it'd have 2 functions, to validate borehole data files (.txt) conform to standard schema & libraries and convert between file formats txt <> xlsx . Is this the sort of thing which could come under OGCAPI-Processes?
Looking to use OGCAPI where ever possible
@KoalaGeo we’ve done exactly this (in WPS) in the context of quality assessment (file checking) and format translation, so I would say this is in scope.
paul van genuchten
thanks @fmigneault. Our use case is a metadata harvester which runs at intervals, grabbing metadata from a remote csw, oai-pmh or opensearch. I hoped to use the concept of process for a harvester, have a certain csw target as parameter to define the job and then schedule that job to run at intervals. But your approach could also work; set up a new process dynamically containing the parameters (harvest target, run schedule) and then use the jobs and job results for retrieving individual runs. eg /api/processes/worldbank-csw-harvest/jobs/2021-05-12 18:00/results