make model manager v2 ready for PR review

- Replace legacy model manager service with the v2 manager.

- Update invocations to use new load interface.

- Fixed many but not all type checking errors in the invocations. Most
  were unrelated to model manager

- Updated routes. All the new routes live under the route tag
  `model_manager_v2`. To avoid confusion with the old routes,
  they have the URL prefix `/api/v2/models`. The old routes
  have been de-registered.

- Added a pytest for the loader.

- Updated documentation in contributing/MODEL_MANAGER.md

tests/aa_nodes/test_graph_execution_state.py

@@ -62,9 +62,7 @@ def mock_services() -> InvocationServices:
         invocation_cache=MemoryInvocationCache(max_cache_size=0),
         logger=logging,  # type: ignore
         model_manager=None,  # type: ignore
-        model_records=None,  # type: ignore
         download_queue=None,  # type: ignore
-        model_install=None,  # type: ignore
         names=None,  # type: ignore
         performance_statistics=InvocationStatsService(),
         processor=DefaultInvocationProcessor(),

tests/aa_nodes/test_invoker.py

@@ -65,9 +65,7 @@ def mock_services() -> InvocationServices:
         invocation_cache=MemoryInvocationCache(max_cache_size=0),
         logger=logging,  # type: ignore
         model_manager=None,  # type: ignore
-        model_records=None,  # type: ignore
         download_queue=None,  # type: ignore
-        model_install=None,  # type: ignore
         names=None,  # type: ignore
         performance_statistics=InvocationStatsService(),
         processor=DefaultInvocationProcessor(),

tests/backend/model_manager_2/model_loading/test_model_load.py

@@ -0,0 +1,22 @@
+"""
+Test model loading
+"""
+
+from pathlib import Path
+
+from invokeai.app.services.model_install import ModelInstallServiceBase
+from invokeai.backend.embeddings.textual_inversion import TextualInversionModelRaw
+from invokeai.backend.model_manager.load import AnyModelLoader
+from tests.backend.model_manager_2.model_manager_2_fixtures import *  # noqa F403
+
+
+def test_loading(mm2_installer: ModelInstallServiceBase, mm2_loader: AnyModelLoader, embedding_file: Path):
+    store = mm2_installer.record_store
+    matches = store.search_by_attr(model_name="test_embedding")
+    assert len(matches) == 0
+    key = mm2_installer.register_path(embedding_file)
+    loaded_model = mm2_loader.load_model(store.get_model(key))
+    assert loaded_model is not None
+    assert loaded_model.config.key == key
+    with loaded_model as model:
+        assert isinstance(model, TextualInversionModelRaw)

tests/backend/model_manager_2/model_manager_2_fixtures.py

@@ -20,6 +20,7 @@ from invokeai.backend.model_manager.config import (
     ModelFormat,
     ModelType,
 )
+from invokeai.backend.model_manager.load import AnyModelLoader, ModelCache, ModelConvertCache
 from invokeai.backend.model_manager.metadata import ModelMetadataStore
 from invokeai.backend.util.logging import InvokeAILogger
 from tests.backend.model_manager_2.model_metadata.metadata_examples import (
@@ -89,6 +90,16 @@ def mm2_app_config(mm2_root_dir: Path) -> InvokeAIAppConfig:
     return app_config
 
 
+@pytest.fixture
+def mm2_loader(mm2_app_config: InvokeAIAppConfig, mm2_record_store: ModelRecordServiceSQL) -> AnyModelLoader:
+    logger = InvokeAILogger.get_logger(config=mm2_app_config)
+    ram_cache = ModelCache(
+        logger=logger, max_cache_size=mm2_app_config.ram_cache_size, max_vram_cache_size=mm2_app_config.vram_cache_size
+    )
+    convert_cache = ModelConvertCache(mm2_app_config.models_convert_cache_path)
+    return AnyModelLoader(app_config=mm2_app_config, logger=logger, ram_cache=ram_cache, convert_cache=convert_cache)
+
+
 @pytest.fixture
 def mm2_record_store(mm2_app_config: InvokeAIAppConfig) -> ModelRecordServiceSQL:
     logger = InvokeAILogger.get_logger(config=mm2_app_config)