robinl / costmodelr Goto Github PK

Rather than the pattern like this

add_oneoff_costs <- function(cost_model, oneoff_cost_assumptions) {

  # If this is the first time we've called add_oneoff_costs, then register a new module, otherwise append new data
  if (!("oneoff_cost" %in% names(cost_model$registered_modules))) {
    cost_model$registered_modules$oneoff_cost <- list()
    cost_model$registered_modules$oneoff_cost$oneoff_cost_raw <- oneoff_cost_assumptions

    # Note we don't process at this point, we just say how to proess.
    # Appending is therefore just a case of adding more assumption rows.  The only difficulty is the id column.
    cost_model$registered_modules$oneoff_cost$process_module <- process_oneoff_costs
  } else {
    cost_model$registered_modules$oneoff_cost$oneoff_cost_raw %<>%
      dplyr::bind_rows(oneoff_cost_assumptions)
  }

  cost_model
}

instead it appends this oneoff cost to list of registered models, which could include multiple oneof costs. This means we can easily add multiple oneoff costs, but is also a more flexible pattern. So for instance, adding user variable costs multiple times, you might need different num users each time. By treating each add operation independently, it is more flexible

When the user specifies day rates in the rate card, these get multiplied by 7 rather than 5.

Somewhere - possible in this or another repository - we want a template project, with template assumptions files read to be edited

This will mean the user can later create charts with labeled axis that show how the quantity consumed or quantity in use varies through time

run_cost_model should erase previous results rather than causing duplication

When trying to install the costsmodelr package i am encountering an error "ERROR: dependency 'hierarchicalbubble' is not available for package 'costmodelr'". I can't seem to find any documentation that shows this package even exists or previously existed.

The things you should be able to control are:

(1) When date does this cost start?
(2) What is the initial quantity at the time the costs start?
(3) How does the quantity increase through time, and how is this related to the number of users?
(4) Is there a fixed quantity increase associated with a new user?

At the moment there's no way of adding costs which are fixed, but scale with the number of users - such as if each user needs 4gb of EBS pre-allocated