feat: re-work distro-codegen

each *.py file in the various templates now has to use `Provider`s rather than the stringified provider_types in the DistributionTemplate. Adjust that, regenerate all templates, docs, etc.

Signed-off-by: Charlie Doern <cdoern@redhat.com>

llama_stack/templates/nvidia/build.yaml

@@ -3,27 +3,39 @@ distribution_spec:
   description: Use NVIDIA NIM for running LLM inference, evaluation and safety
   providers:
     inference:
-    - remote::nvidia
+    - provider_id: nvidia
+      provider_type: remote::nvidia
     vector_io:
-    - inline::faiss
+    - provider_id: faiss
+      provider_type: inline::faiss
     safety:
-    - remote::nvidia
+    - provider_id: nvidia
+      provider_type: remote::nvidia
     agents:
-    - inline::meta-reference
+    - provider_id: meta-reference
+      provider_type: inline::meta-reference
     telemetry:
-    - inline::meta-reference
+    - provider_id: meta-reference
+      provider_type: inline::meta-reference
     eval:
-    - remote::nvidia
+    - provider_id: nvidia
+      provider_type: remote::nvidia
     post_training:
-    - remote::nvidia
+    - provider_id: nvidia
+      provider_type: remote::nvidia
     datasetio:
-    - inline::localfs
-    - remote::nvidia
+    - provider_id: localfs
+      provider_type: inline::localfs
+    - provider_id: nvidia
+      provider_type: remote::nvidia
     scoring:
-    - inline::basic
+    - provider_id: basic
+      provider_type: inline::basic
     tool_runtime:
-    - inline::rag-runtime
+    - provider_id: rag-runtime
+      provider_type: inline::rag-runtime
 image_type: conda
+image_name: nvidia
 additional_pip_packages:
 - aiosqlite
 - sqlalchemy[asyncio]