Merge pull request #85 from LlmKira/dev

✨ refactor(generate_image): update field imports and type aliases

README.md

@@ -11,32 +11,12 @@ The goal of this repository is to use Pydantic to build legitimate requests to a
 
 > Python >= 3.9 is required.
 
-### Roadmap 🚧
+### 📰 News
 
-- [x] tool.random_prompt
-- [x] tool.paint_mask
-- [x] tool.image_metadata
-- [x] tokenizer
-- [x] /ai/generate-image
-- [x] /user/subscription
-- [x] /user/login
-- [x] /user/information
-- [x] /ai/upscale
-- [x] /ai/generate-image/suggest-tags
-- [x] /ai/generate-voice
-- [x] /ai/generate-stream
-- [x] /ai/generate
-- [x] /ai/augment-image
-- [ ] /ai/annotate-image
-- [ ] /ai/classify
-- [ ] /ai/generate-prompt
+- [Image Generation Model Release — NovelAI Anime Diffusion V4 Curated Preview (EN)](https://blog.novelai.net/release-novelai-anime-diffusion-v4-curated-preview-en-ca4b0b11e671)
+- [Tutorial: Creating Consistent Characters with NovelAI Diffusion Anime [Female]](https://blog.novelai.net/tutorial-en-creating-consistent-characters-with-novelai-diffusion-anime-female-538b4b678a4e)
 
-> GenerateImageInfer.calculate_cost is correct in most cases, but please request account information to get accurate
-> consumption information.
-
-> This repo is maintained by me personally now. If you have any questions, please feel free to open an issue.
-
-### Usage 🖥️
+### 📦 Usage
 
 ```shell
 pip -U install novelai-python
@@ -101,7 +81,7 @@ loop.run_until_complete(main())
 
 ```
 
-#### LLM
+#### 📦 LLM
 
 ```python
 import asyncio
@@ -152,7 +132,7 @@ loop = asyncio.get_event_loop()
 loop.run_until_complete(chat("Hello"))
 ```
 
-#### Random Prompt
+#### 📦 Random Prompt
 
 ```python
 from novelai_python.tool.random_prompt import RandomPromptGenerator
@@ -161,14 +141,14 @@ prompt = RandomPromptGenerator(nsfw_enabled=False).random_prompt()
 print(prompt)
 ```
 
-#### Run A Server
+#### 📦 Run A Server
 
 ```shell
 pip install novelai_python
 python3 -m novelai_python.server -h '127.0.0.1' -p 7888
 ```
 
-#### Tokenizer
+#### 📦 Tokenizer
 
 ```python
 from novelai_python._enum import get_tokenizer_model, TextLLMModel
@@ -183,15 +163,40 @@ print(tokenizer_package.decode(tokenizer_package.encode(t_text)))
 
 ```
 
-## About Nsfw 🚫
+### 🔨 Roadmap
+
+- [x] tool.random_prompt
+- [x] tool.paint_mask
+- [x] tool.image_metadata
+- [x] tokenizer
+- [x] /ai/generate-image
+- [x] /user/subscription
+- [x] /user/login
+- [x] /user/information
+- [x] /ai/upscale
+- [x] /ai/generate-image/suggest-tags
+- [x] /ai/generate-voice
+- [x] /ai/generate-stream
+- [x] /ai/generate
+- [x] /ai/augment-image
+- [ ] /ai/annotate-image
+- [ ] /ai/classify
+- [ ] /ai/generate-prompt
+
+> GenerateImageInfer.calculate_cost is correct in most cases, but please request account information to get accurate
+> consumption information.
+
+> This repo is maintained by me personally now. If you have any questions, please feel free to open an issue.
+
+## 🚫 About Nsfw
 
 You might need some solutions for identifying NSFW content and adding a mosaic to prevent operational mishaps.
 
 https://dghs-imgutils.deepghs.org/main/api_doc/detect/nudenet.html
 
 https://dghs-imgutils.deepghs.org/main/api_doc/operate/censor.html
 
-## Acknowledgements 🙏
+## 🙏 Acknowledgements
 
 [BackEnd](https://api.novelai.net/docs)
 

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "novelai-python"
-version = "0.6.0"
+version = "0.6.1"
 description = "NovelAI Python Binding With Pydantic"
 authors = [
     { name = "sudoskys", email = "[email protected]" },

src/novelai_python/sdk/ai/generate_image/__init__.py

@@ -7,7 +7,6 @@
 import math
 import random
 from copy import deepcopy
-from dataclasses import field
 from io import BytesIO
 from typing import Optional, Union, Tuple, List
 from urllib.parse import urlparse
@@ -28,7 +27,7 @@
 from novelai_python.sdk.ai._enum import Model, Sampler, NoiseSchedule, ControlNetModel, Action, UCPreset, \
     INPAINTING_MODEL_LIST, get_default_uc_preset, \
     ModelTypeAlias, UCPresetTypeAlias, get_default_noise_schedule, get_supported_noise_schedule, \
-    get_model_group, ModelGroups, get_supported_params, get_modifiers
+    get_model_group, ModelGroups, get_supported_params, get_modifiers, ImageBytesTypeAlias
 from .schema import Character, V4Prompt, V4NegativePrompt
 from ...schema import ApiBaseModel
 from ...._exceptions import APIError, AuthError, ConcurrentGenerationError, SessionHttpError
@@ -79,21 +78,21 @@ class Params(BaseModel):
     """Noise Schedule"""
     legacy_v3_extend: Optional[bool] = False
     """Legacy V3 Extend"""
-    mask: Union[bytes, str] = None
+    mask: ImageBytesTypeAlias = None
     """Mask for Inpainting"""
     seed: int = Field(
         default_factory=lambda: random.randint(0, 4294967295 - 7),
         gt=0,
         le=4294967295 - 7,
     )
     """Seed"""
-    image: Union[bytes, str] = None
+    image: ImageBytesTypeAlias = None
     """Image for img2img"""
     negative_prompt: Optional[str] = ''
     """Negative Prompt"""
-    reference_image_multiple: Optional[List[Union[str, bytes]]] = field(default_factory=list)
+    reference_image_multiple: Optional[List[Union[str, bytes]]] = Field(default_factory=list)
     """Reference Image For Vibe Mode"""
-    reference_information_extracted_multiple: Optional[List[float]] = field(default_factory=list)
+    reference_information_extracted_multiple: Optional[List[float]] = Field(default_factory=list)
     """Reference Information Extracted For Vibe Mode"""
     reference_strength_multiple: Optional[List[float]] = Field(
         default=list,
@@ -113,7 +112,7 @@ class Params(BaseModel):
     """
     use_coords: bool = Field(True, description="Use Coordinates")
     """Use Coordinates"""
-    characterPrompts: List[Character] = field(default_factory=list)
+    characterPrompts: List[Character] = Field(default_factory=list)
     """Character Prompts"""
 
     v4_prompt: Optional[V4Prompt] = Field(None, description="V4 Prompt")
@@ -136,6 +135,12 @@ class Params(BaseModel):
     uncond_scale: Optional[float] = Field(None, ge=0, le=1.5, multiple_of=0.05)
     """Undesired Content Strength"""
 
+    @model_validator(mode="after")
+    def v_character(self):
+        if len(self.characterPrompts) > 6:
+            raise ValueError("Too many character given")
+        return self
+
     @field_validator('uncond_scale')
     def v_uncond_scale(cls, v: float):
         """
@@ -684,10 +689,9 @@ def build_img2img(
 
         If you need to define more parameters, you should initialize the Param class yourself.
 
-        :param controlnet_strength:
         :param strength:
         :param image:
-        :param noise: For img2img
+        :param noise: For image to image
         :param extra_noise_seed: Get extra_noise_seed
         :param reference_information_extracted_multiple:
         :param reference_strength_multiple: