Update the "InterleavedTextMedia" type (#635)

## What does this PR do?

This is a long-pending change and particularly important to get done
now.

Specifically:
- we cannot "localize" (aka download) any URLs from media attachments
anywhere near our modeling code. it must be done within llama-stack.
- `PIL.Image` is infesting all our APIs via `ImageMedia ->
InterleavedTextMedia` and that cannot be right at all. Anything in the
API surface must be "naturally serializable". We need a standard `{
type: "image", image_url: "<...>" }` which is more extensible
- `UserMessage`, `SystemMessage`, etc. are moved completely to
llama-stack from the llama-models repository.

See https://github.com/meta-llama/llama-models/pull/244 for the
corresponding PR in llama-models.

## Test Plan

```bash
cd llama_stack/providers/tests

pytest -s -v -k "fireworks or ollama or together" inference/test_vision_inference.py
pytest -s -v -k "(fireworks or ollama or together) and llama_3b" inference/test_text_inference.py
pytest -s -v -k chroma memory/test_memory.py \
  --env EMBEDDING_DIMENSION=384 --env CHROMA_DB_PATH=/tmp/foobar

pytest -s -v -k fireworks agents/test_agents.py  \
   --safety-shield=meta-llama/Llama-Guard-3-8B \
   --inference-model=meta-llama/Llama-3.1-8B-Instruct
```

Updated the client sdk (see PR ...), installed the SDK in the same
environment and then ran the SDK tests:

```bash
cd tests/client-sdk
LLAMA_STACK_CONFIG=together pytest -s -v agents/test_agents.py
LLAMA_STACK_CONFIG=ollama pytest -s -v memory/test_memory.py

# this one needed a bit of hacking in the run.yaml to ensure I could register the vision model correctly
INFERENCE_MODEL=llama3.2-vision:latest LLAMA_STACK_CONFIG=ollama pytest -s -v inference/test_inference.py
```

llama_stack/providers/tests/memory/test_memory.py

@@ -46,13 +46,13 @@ def sample_documents():
 
 
 async def register_memory_bank(
-    banks_impl: MemoryBanks, inference_model: str
+    banks_impl: MemoryBanks, embedding_model: str
 ) -> MemoryBank:
     bank_id = f"test_bank_{uuid.uuid4().hex}"
     return await banks_impl.register_memory_bank(
         memory_bank_id=bank_id,
         params=VectorMemoryBankParams(
-            embedding_model=inference_model,
+            embedding_model=embedding_model,
             chunk_size_in_tokens=512,
             overlap_size_in_tokens=64,
         ),
@@ -61,11 +61,11 @@ async def register_memory_bank(
 
 class TestMemory:
     @pytest.mark.asyncio
-    async def test_banks_list(self, memory_stack, inference_model):
+    async def test_banks_list(self, memory_stack, embedding_model):
         _, banks_impl = memory_stack
 
         # Register a test bank
-        registered_bank = await register_memory_bank(banks_impl, inference_model)
+        registered_bank = await register_memory_bank(banks_impl, embedding_model)
 
         try:
             # Verify our bank shows up in list
@@ -86,7 +86,7 @@ class TestMemory:
         )
 
     @pytest.mark.asyncio
-    async def test_banks_register(self, memory_stack, inference_model):
+    async def test_banks_register(self, memory_stack, embedding_model):
         _, banks_impl = memory_stack
 
         bank_id = f"test_bank_{uuid.uuid4().hex}"
@@ -96,7 +96,7 @@ class TestMemory:
             await banks_impl.register_memory_bank(
                 memory_bank_id=bank_id,
                 params=VectorMemoryBankParams(
-                    embedding_model=inference_model,
+                    embedding_model=embedding_model,
                     chunk_size_in_tokens=512,
                     overlap_size_in_tokens=64,
                 ),
@@ -111,7 +111,7 @@ class TestMemory:
             await banks_impl.register_memory_bank(
                 memory_bank_id=bank_id,
                 params=VectorMemoryBankParams(
-                    embedding_model=inference_model,
+                    embedding_model=embedding_model,
                     chunk_size_in_tokens=512,
                     overlap_size_in_tokens=64,
                 ),
@@ -129,14 +129,14 @@ class TestMemory:
 
     @pytest.mark.asyncio
     async def test_query_documents(
-        self, memory_stack, inference_model, sample_documents
+        self, memory_stack, embedding_model, sample_documents
     ):
         memory_impl, banks_impl = memory_stack
 
         with pytest.raises(ValueError):
             await memory_impl.insert_documents("test_bank", sample_documents)
 
-        registered_bank = await register_memory_bank(banks_impl, inference_model)
+        registered_bank = await register_memory_bank(banks_impl, embedding_model)
         await memory_impl.insert_documents(
             registered_bank.memory_bank_id, sample_documents
         )