In this guide we'll discuss some best practices for how to structure your project. These recommendations align well with [Elixir conventions](https://hexdocs.pm/elixir/1.16.2/naming-conventions.html#casing) around file and module naming. These conventions allow for a logical coupling of module and file names, and help keep your project organized and easy to navigate.
Generally speaking, your Ash application lives in the standard place within your elixir application, i.e `lib/my_app`. Within that folder, you create one folder for each context that you have. Each context has an `Ash.Domain` module within it, and the resources that live within that context. All resource interaction ultimately goes through a domain module.
Alongside the domain module, you have your resources, as well as any other files used in the context. If a resource has any additional files that are used to implement it, they should be placed in a folder with the same name as the resource, in subfolders grouping the files by type. Feel free to choose another logical grouping, but we've found by-type to be effective.
The purpose of Ash is to be both the model of and the interface to your domain logic (A.K.A business logic). Applying this generally looks like building as much of your domain logic "behind" your resources. This does not mean, however, that everything has to go _inside of_ your resources. For example, if you have a `Purchase` resource, and you want to be able to display a list of purchases that were taxable, and also calculate the percentage of the purchase that was taxable. You might have an action called `:taxable` and a calculation called `:percentage_tax`.
In practice, you may not need the `taxable` action, i.e perhaps you simply want a "taxable" checkbox on a list view in your application, in which case you may use the primary read, or some other read like `:transaction_report`. You would then, on the consumer, provide the filter for `taxable == true`, and load the `:percentage_tax` calculation.
## Example 2: Using external data in create actions
Lets say you want the user to fill in a github issue id, and you will fetch information from that github issue to use as part of creating a "ticket" in your system.. You might be tempted to do something like this in a LiveView:
```elixir
def handle_event("link_ticket", %{"issue_id" => issue_id}, socket) do