fix: use apply/3 to avoid undefined-module warnings with Conditional ML module

Microwaveprop.Propagation.Model only exists in lib_ml/ which is excluded
from prod elixirc_paths. Using apply/3 instead of direct @ml_module
calls avoids compile-time warnings when the module is absent. Added
credo:disable-for-next-line inline annotations since Credo flags
apply/3 with known arity but the runtime dispatch is intentional.

Also adds --warnings-as-errors to Dockerfile RUN mix compile so the
build fails on any project-level warning.

Dockerfile

@@ -47,7 +47,7 @@ COPY priv/agent-skills priv/agent-skills
 # /docs/api artifacts are read at compile time by ApiDocsController
 # and ApiDocsLive (baked into module attributes).
 COPY docs/api docs/api
-RUN mix compile
+RUN mix compile --warnings-as-errors
 
 # copy assets and build (JS/CSS changes don't require recompilation)
 COPY assets assets

lib/microwaveprop/propagation.ex

@@ -26,9 +26,11 @@ defmodule Microwaveprop.Propagation do
   @spec load_ml_model() :: :ok
   def load_ml_model do
     if Code.ensure_loaded?(@ml_module) do
-      case @ml_module.load() do
+      # credo:disable-for-next-line Credo.Check.Refactor.Apply
+      case apply(@ml_module, :load, []) do
         {:ok, params} ->
-          predict_fn = @ml_module.compile_predict()
+          # credo:disable-for-next-line Credo.Check.Refactor.Apply
+          predict_fn = apply(@ml_module, :compile_predict, [])
           :persistent_term.put(@ml_key, {predict_fn, params})
           Logger.info("PropagationML: model loaded and compiled")
           :ok