Is a good idea for RBC to know if the server has support for armadillo/mlpack?

[pearu]

Also, do you think it is a good idea for RBC to know if the server has support for armadillo/mlpack?

Originally posted by @guilhermeleobas in https://github.com/xnd-project/rbc/issues/317#issuecomment-814626142

Yes, in general, the omnscidb clients like RBC should be provided a method for querying various parameters of the server. Currently, omniscidb thrift interface provides the following relevant end-points that could be used for the task:

TServerStatus get_server_status(1: TSessionId session) throws (1: TOmniSciException e)
list<TServerStatus> get_status(1: TSessionId session) throws (1: TOmniSciException e)
TClusterHardwareInfo get_hardware_info(1: TSessionId session) throws (1: TOmniSciException e)
string get_version() throws (1: TOmniSciException e)
TRowDescriptor sql_validate(1: TSessionId session, 2: string query) throws (1: TOmniSciException e)
map<string, string> get_device_parameters(1: TSessionId session) throws (1: TOmniSciException e)

Notice:

struct TServerStatus {
  1: bool read_only;
  2: string version;
  3: bool rendering_enabled;
  4: i64 start_time;
  5: string edition;
  6: string host_name;
  7: bool poly_rendering_enabled;
  8: TRole role;
}

that is, get_status might be a viable option if TServerStatus is extended with a member containing information about cmake flags, for instance.