AF
Ash Framework
The Elixir backend framework for unparalleled productivity. Declarative tools that let you stop wasting time. Use with Phoenix LiveView or build APIs in minutes for your front-end of choice.
JoinAF
Ash Framework
The Elixir backend framework for unparalleled productivity. Declarative tools that let you stop wasting time. Use with Phoenix LiveView or build APIs in minutes for your front-end of choice.
JoinHow to write a policy to ensure the user belongs to the tenant through a membership relation?
I have read many examples on multitenancy implementation with Ash, but most of these examples assume that the user belongs to one tenant.
In my case, a user can belong to one or more tenants and can switch from one to another using a select list.
We therefore have three resources, Organization, User, and Membership, to make the connection. Our application mainly uses a JSON API where we pass an authentication token in each request, as well as a custom X-Org-Id header with the current tenant ID the user selected....
Anonymous Function Queries (Preparations)
I can't seem to get the syntax right, what am I doing wrong?
```
preparations do
prepare fn query, context ->...
Loading data from an external API for a calculation
Hello! I have a resource that I would like to enhance in certain reads with a calculation that needs to be based off data from an external API.
I'm mindful of not wanting to make the API request from within a transaction if possible.
I thought of adding a preparation to the read action that will fetch the API data and put it into the context, which the calculation can use.
...
Flattening Polymorphic Relationships in Ash GraphQL
Hey folks 👋
I’m working with polymorphic relationships in Ash and followed the guide here: https://hexdocs.pm/ash/polymorphic-relationships.html. I tried exposing it through the GraphQL API, and I’m a bit unsure if there’s a way to shape the schema the way I want.
Right now, when I query
bankAccounts, I get something like this:...Ash Auth For Inertia Setup & Ash Auth vs Ash Phoenix Auth
Hi everyone! So I have been playing with the
mix phx.gen.auth in my Inertia setup and was wondering:
a) Am I losing anything if I don't use Ash Auth yet if I already has my auth configured from the phx.gen.auth since I can just pass the right scope for the actors when using Ash? I am migrating my models to Ash resources now but I need to understand how the ash auth works in tandem with the ash resources as a whole.
b) For the auth vs phoenix auth, I assume phoenix auth is just the UI/liveview + routes scope on top of the ash auth? I can still build my own routes with Inertia rendering and use the auth module to trigger the auth internally right?...Solution:
you can just use your generated auth, Ash doesn't really care to much about what you pass in as an actor, your policies just need to know how to work with it. I've used ash with our own custom auth before ash_authentication was a thing, worked without a hitch.
ash_authentication_live_session dups
After implementing
Ash.Scope, it occurred to me that I'm sort of duplicating assignments with current_user as it's also set via ash_authentication_live_session. Is this generally just how folks do things for convenient access in the view or should I be deduplicating somehow? I feel like I shouldn't care but it bothers me on some level lol. Thanks!Cross schema many_to_many
Hi, I have two resources let's say A and B, A is using :context multitenancy, B is in the public schema.
I want to have a many_to_many relationship between A and B, using a AB resource that also is using :context multitenancy.
The issue is that when I try to load the B records directly from A (MyApp.get_a(a_id, tenant: tenant, load: [:bs]), The query generated tries to join with an AB table from the public schema, which does not exist as it is a multitenant resource. Did I miss something or is it a bug ? Loading through the join resource does work though (MyApp.get_a(a_id, tenant: tenant, load: [abs: [:b]])...
TypedStruct with array and default
I have a typed_struct with an
{:array, OtherStruct} field. Passing data to new() works as expected but after adding default: [] to the field, the field only returns []. I'm not sure if this is a bug or intended.
Structs:
```...
Ash.Notifier
Is there a way to have a parametrized notifier, just like with validators/preparation/changes
AshJsonApi: Can a domain support routes for different endpoints/routers?
I'm looking into exposing some API routes only for my internal network, which won't be publicly accessible. Is it possible to configure routes to be exposed for one or more endpoints/routers? i.e. Some routes are public only, some private only, others are both.
Usage rules task not found
Hi! Did something break in
usage_rules? I'm getting this all of a sudden: ** (Mix) The task "usage_rules.sync" could not be found
It's in a clean new project. After running ash.setup, the next thing I always do before starting to work on the project is generating the AGENTS.md file. Always worked fine until now...Having a small policy that doesn't seem to ever be called
I have a policy where I need an actor to be present for three different actions, but the policies don't seem to be called.
The policies:
```elixir...
Solution:
yeah, in this context, the policies don't guarantee an actor. you could wrap this in before_action and than you'd know that if you reach the hook that there is an actor beccause of the policy, or you could just handle the actor being nil
case Map.get(context.actor || %{}, :group_id) do which i would probably doAsh AI not working
Failed to JSON decode error response. ERROR: %Jason.DecodeError{position: 1005, token: nil, data: "{\n \"error\": {\n \"code\": 400,\n \"message\": \"Invalid JSON payload received. Unknown name \\\"additionalProperties\\\" at 'tools[0].function_declarations[0].parameters
Failed to JSON decode error response. ERROR: %Jason.DecodeError{position: 1005, token: nil, data: "{\n \"error\": {\n \"code\": 400,\n \"message\": \"Invalid JSON payload received. Unknown name \\\"additionalProperties\\\" at 'tools[0].function_declarations[0].parameters
Missing __spark_metadata__ field in Ash.Notifier.PubSub.Publication struct
After upgrading to Spark 2.3.4 and Elixir 1.18.4, I’m seeing the following warning:
warning: Entity without spark_metadata field is deprecated. Entity Ash.Notifier.PubSub.Publication does not define a
__spark_metadata__ field. This field is required to access source annotations. Add __spark_metadata__: nil to the defstruct for Ash.Notifier.PubSub.Publication.
(elixir 1.18.4) lib/enum.ex:987: Enum."-each/2-lists^foreach/1-0-"/2
(spark 2.3.4) lib/spark/dsl/extension.ex:2135: Spark.Dsl.Extension.after_verify/1...no function clause matching in Pgvector.new/1
This is what I'm doing:
```elixir
case Req.post(url, json: body) do
{:ok, %{status: 200, body: %{"embeddings" => embeddings}}} ->...
Solution:
i think pg vector package is the issue
Ash AI writeable error:
```
vectorize do
attributes description: :description_vector
embedding_model CoffeeViewer.EmbeddingModels.Gemini...
Solution:
Well, you made it
writable? falsespark warning on compilation
I get bombarded by these warnings:
``
warning: Entity without __spark_metadata__ field is deprecated. Entity Reactor.Dsl.Middleware does not define a spark_metadata field. This field is required to acces
s source annotations. Add spark_metadata: nil` to the defstruct for Reactor.Dsl.Middleware.
(elixir 1.18.4) lib/enum.ex:987: Enum."-each/2-lists^foreach/1-0-"/2...usage rules req me to change igniter `only`
In a default app
igniter is not in prod but when i try to add usage rules via the expert LS it says that there's a version conflict.Solution:
But I'm telling you to add that to your usage rules dep lol
Composite primary keys vs through
It looks like
destination/source_attribute_on_join_resource both expect an atom. Is it possible to reference a composite key on a join table? I am trying to set up a many_to_many relationship.Solution:
But you can use a has_many with a filter
Can I instruct Ash to force update a CiString value regardless of comparison result?
I'm using
Ash.CiString on a resource field (backed by postgres citext) for all the uniqueness constraint validation goodness, but recently discovered that Ash is skipping updates to the field value when only the case has changed.
I know, I know: why am I using a CiString if I don't want case-insensitive comparison? Because I don't want to clobber other records with the same case-insensitive name. But I do want to give the user control to update the casing of the value on an individual record. I have a NameCollissionPrevention validation that ignores the current record in question when deciding if the record update is valid, so I don't think this is programmer error, it feels like Ash is ignoring the field in the update b/c it compares as equal.
Maybe CiString isn't the right tool for the job in this scenario. I'm curious if Ash provides a way to say "you can update this value to anything (and any case) you want". ...Solution:
I ended up modifying it a bit b/c the
:name arg isn't required:
```
change(fn changeset, _ctx ->
case changeset.arguments[:name] do...