I want to write an application that solely relies on API keys in order to automate some tasks using ArcGIS pro.
I am using this library to execute some functions. I am not using notebooks to run it or run it within arcgis pro and I am developing on windows using conda for the environment management.
My problem is that I am finding that some of the functions rely on the local install of arcgis pro
- enriching a buffer study area,
- converting from shapely to arcgis polygons.
Since we have a paid license to use it, and we have supplied a api key to the GIS object and then supplied the GIS to the functions in question, why do I still need the arcgis application installed locally and logged in?
This makes deployment much harder than it needs to be for us as we are running the application as a service on a VM and the API key owner has to login and change the password every time the password needs to be changed due to company policy. It would be nice to eventually package the whole application as a single executable.
Is it possible that functions that demand the Arcgis Pro software be labeled? It is very frustrating using a function and realizing it is not just a wrapper for the api calls. Is using docker enough or will it demand that Arcgis Pro be installed and logged in too?