mirror of
https://github.com/invoke-ai/InvokeAI.git
synced 2026-02-11 05:55:23 -05:00
c238a7f18b
Upgrade pydantic and fastapi to latest.
- pydantic~=2.4.2
- fastapi~=103.2
- fastapi-events~=0.9.1
**Big Changes**
There are a number of logic changes needed to support pydantic v2. Most changes are very simple, like using the new methods to serialized and deserialize models, but there are a few more complex changes.
**Invocations**
The biggest change relates to invocation creation, instantiation and validation.
Because pydantic v2 moves all validation logic into the rust pydantic-core, we may no longer directly stick our fingers into the validation pie.
Previously, we (ab)used models and fields to allow invocation fields to be optional at instantiation, but required when `invoke()` is called. We directly manipulated the fields and invocation models when calling `invoke()`.
With pydantic v2, this is much more involved. Changes to the python wrapper do not propagate down to the rust validation logic - you have to rebuild the model. This causes problem with concurrent access to the invocation classes and is not a free operation.
This logic has been totally refactored and we do not need to change the model any more. The details are in `baseinvocation.py`, in the `InputField` function and `BaseInvocation.invoke_internal()` method.
In the end, this implementation is cleaner.
**Invocation Fields**
In pydantic v2, you can no longer directly add or remove fields from a model.
Previously, we did this to add the `type` field to invocations.
**Invocation Decorators**
With pydantic v2, we instead use the imperative `create_model()` API to create a new model with the additional field. This is done in `baseinvocation.py` in the `invocation()` wrapper.
A similar technique is used for `invocation_output()`.
**Minor Changes**
There are a number of minor changes around the pydantic v2 models API.
**Protected `model_` Namespace**
All models' pydantic-provided methods and attributes are prefixed with `model_` and this is considered a protected namespace. This causes some conflict, because "model" means something to us, and we have a ton of pydantic models with attributes starting with "model_".
Forunately, there are no direct conflicts. However, in any pydantic model where we define an attribute or method that starts with "model_", we must tell set the protected namespaces to an empty tuple.
```py
class IPAdapterModelField(BaseModel):
model_name: str = Field(description="Name of the IP-Adapter model")
base_model: BaseModelType = Field(description="Base model")
model_config = ConfigDict(protected_namespaces=())
```
**Model Serialization**
Pydantic models no longer have `Model.dict()` or `Model.json()`.
Instead, we use `Model.model_dump()` or `Model.model_dump_json()`.
**Model Deserialization**
Pydantic models no longer have `Model.parse_obj()` or `Model.parse_raw()`, and there are no `parse_raw_as()` or `parse_obj_as()` functions.
Instead, you need to create a `TypeAdapter` object to parse python objects or JSON into a model.
```py
adapter_graph = TypeAdapter(Graph)
deserialized_graph_from_json = adapter_graph.validate_json(graph_json)
deserialized_graph_from_dict = adapter_graph.validate_python(graph_dict)
```
**Field Customisation**
Pydantic `Field`s no longer accept arbitrary args.
Now, you must put all additional arbitrary args in a `json_schema_extra` arg on the field.
**Schema Customisation**
FastAPI and pydantic schema generation now follows the OpenAPI version 3.1 spec.
This necessitates two changes:
- Our schema customization logic has been revised
- Schema parsing to build node templates has been revised
The specific aren't important, but this does present additional surface area for bugs.
**Performance Improvements**
Pydantic v2 is a full rewrite with a rust backend. This offers a substantial performance improvement (pydantic claims 5x to 50x depending on the task). We'll notice this the most during serialization and deserialization of sessions/graphs, which happens very very often - a couple times per node.
I haven't done any benchmarks, but anecdotally, graph execution is much faster. Also, very larges graphs - like with massive iterators - are much, much faster.
168 lines
5.8 KiB
Python
168 lines
5.8 KiB
Python
import inspect
|
|
from enum import Enum
|
|
from typing import Literal, get_origin
|
|
|
|
from pydantic import BaseModel, ConfigDict, create_model
|
|
|
|
from .base import ( # noqa: F401
|
|
BaseModelType,
|
|
DuplicateModelException,
|
|
InvalidModelException,
|
|
ModelBase,
|
|
ModelConfigBase,
|
|
ModelError,
|
|
ModelNotFoundException,
|
|
ModelType,
|
|
ModelVariantType,
|
|
SchedulerPredictionType,
|
|
SilenceWarnings,
|
|
SubModelType,
|
|
)
|
|
from .clip_vision import CLIPVisionModel
|
|
from .controlnet import ControlNetModel # TODO:
|
|
from .ip_adapter import IPAdapterModel
|
|
from .lora import LoRAModel
|
|
from .sdxl import StableDiffusionXLModel
|
|
from .stable_diffusion import StableDiffusion1Model, StableDiffusion2Model
|
|
from .stable_diffusion_onnx import ONNXStableDiffusion1Model, ONNXStableDiffusion2Model
|
|
from .t2i_adapter import T2IAdapterModel
|
|
from .textual_inversion import TextualInversionModel
|
|
from .vae import VaeModel
|
|
|
|
MODEL_CLASSES = {
|
|
BaseModelType.StableDiffusion1: {
|
|
ModelType.ONNX: ONNXStableDiffusion1Model,
|
|
ModelType.Main: StableDiffusion1Model,
|
|
ModelType.Vae: VaeModel,
|
|
ModelType.Lora: LoRAModel,
|
|
ModelType.ControlNet: ControlNetModel,
|
|
ModelType.TextualInversion: TextualInversionModel,
|
|
ModelType.IPAdapter: IPAdapterModel,
|
|
ModelType.CLIPVision: CLIPVisionModel,
|
|
ModelType.T2IAdapter: T2IAdapterModel,
|
|
},
|
|
BaseModelType.StableDiffusion2: {
|
|
ModelType.ONNX: ONNXStableDiffusion2Model,
|
|
ModelType.Main: StableDiffusion2Model,
|
|
ModelType.Vae: VaeModel,
|
|
ModelType.Lora: LoRAModel,
|
|
ModelType.ControlNet: ControlNetModel,
|
|
ModelType.TextualInversion: TextualInversionModel,
|
|
ModelType.IPAdapter: IPAdapterModel,
|
|
ModelType.CLIPVision: CLIPVisionModel,
|
|
ModelType.T2IAdapter: T2IAdapterModel,
|
|
},
|
|
BaseModelType.StableDiffusionXL: {
|
|
ModelType.Main: StableDiffusionXLModel,
|
|
ModelType.Vae: VaeModel,
|
|
# will not work until support written
|
|
ModelType.Lora: LoRAModel,
|
|
ModelType.ControlNet: ControlNetModel,
|
|
ModelType.TextualInversion: TextualInversionModel,
|
|
ModelType.ONNX: ONNXStableDiffusion2Model,
|
|
ModelType.IPAdapter: IPAdapterModel,
|
|
ModelType.CLIPVision: CLIPVisionModel,
|
|
ModelType.T2IAdapter: T2IAdapterModel,
|
|
},
|
|
BaseModelType.StableDiffusionXLRefiner: {
|
|
ModelType.Main: StableDiffusionXLModel,
|
|
ModelType.Vae: VaeModel,
|
|
# will not work until support written
|
|
ModelType.Lora: LoRAModel,
|
|
ModelType.ControlNet: ControlNetModel,
|
|
ModelType.TextualInversion: TextualInversionModel,
|
|
ModelType.ONNX: ONNXStableDiffusion2Model,
|
|
ModelType.IPAdapter: IPAdapterModel,
|
|
ModelType.CLIPVision: CLIPVisionModel,
|
|
ModelType.T2IAdapter: T2IAdapterModel,
|
|
},
|
|
BaseModelType.Any: {
|
|
ModelType.CLIPVision: CLIPVisionModel,
|
|
# The following model types are not expected to be used with BaseModelType.Any.
|
|
ModelType.ONNX: ONNXStableDiffusion2Model,
|
|
ModelType.Main: StableDiffusion2Model,
|
|
ModelType.Vae: VaeModel,
|
|
ModelType.Lora: LoRAModel,
|
|
ModelType.ControlNet: ControlNetModel,
|
|
ModelType.TextualInversion: TextualInversionModel,
|
|
ModelType.IPAdapter: IPAdapterModel,
|
|
ModelType.T2IAdapter: T2IAdapterModel,
|
|
},
|
|
# BaseModelType.Kandinsky2_1: {
|
|
# ModelType.Main: Kandinsky2_1Model,
|
|
# ModelType.MoVQ: MoVQModel,
|
|
# ModelType.Lora: LoRAModel,
|
|
# ModelType.ControlNet: ControlNetModel,
|
|
# ModelType.TextualInversion: TextualInversionModel,
|
|
# },
|
|
}
|
|
|
|
MODEL_CONFIGS = list()
|
|
OPENAPI_MODEL_CONFIGS = list()
|
|
|
|
|
|
class OpenAPIModelInfoBase(BaseModel):
|
|
model_name: str
|
|
base_model: BaseModelType
|
|
model_type: ModelType
|
|
|
|
model_config = ConfigDict(protected_namespaces=())
|
|
|
|
|
|
for base_model, models in MODEL_CLASSES.items():
|
|
for model_type, model_class in models.items():
|
|
model_configs = set(model_class._get_configs().values())
|
|
model_configs.discard(None)
|
|
MODEL_CONFIGS.extend(model_configs)
|
|
|
|
# LS: sort to get the checkpoint configs first, which makes
|
|
# for a better template in the Swagger docs
|
|
for cfg in sorted(model_configs, key=lambda x: str(x)):
|
|
model_name, cfg_name = cfg.__qualname__.split(".")[-2:]
|
|
openapi_cfg_name = model_name + cfg_name
|
|
if openapi_cfg_name in vars():
|
|
continue
|
|
|
|
api_wrapper = create_model(
|
|
openapi_cfg_name,
|
|
__base__=(cfg, OpenAPIModelInfoBase),
|
|
model_type=(Literal[model_type], model_type), # type: ignore
|
|
)
|
|
vars()[openapi_cfg_name] = api_wrapper
|
|
OPENAPI_MODEL_CONFIGS.append(api_wrapper)
|
|
|
|
|
|
def get_model_config_enums():
|
|
enums = list()
|
|
|
|
for model_config in MODEL_CONFIGS:
|
|
if hasattr(inspect, "get_annotations"):
|
|
fields = inspect.get_annotations(model_config)
|
|
else:
|
|
fields = model_config.__annotations__
|
|
try:
|
|
field = fields["model_format"]
|
|
except Exception:
|
|
raise Exception("format field not found")
|
|
|
|
# model_format: None
|
|
# model_format: SomeModelFormat
|
|
# model_format: Literal[SomeModelFormat.Diffusers]
|
|
# model_format: Literal[SomeModelFormat.Diffusers, SomeModelFormat.Checkpoint]
|
|
|
|
if isinstance(field, type) and issubclass(field, str) and issubclass(field, Enum):
|
|
enums.append(field)
|
|
|
|
elif get_origin(field) is Literal and all(
|
|
isinstance(arg, str) and isinstance(arg, Enum) for arg in field.__args__
|
|
):
|
|
enums.append(type(field.__args__[0]))
|
|
|
|
elif field is None:
|
|
pass
|
|
|
|
else:
|
|
raise Exception(f"Unsupported format definition in {model_configs.__qualname__}")
|
|
|
|
return enums
|