novelai-python 0.4.4__tar.gz → 0.4.5__tar.gz

{novelai_python-0.4.4 → novelai_python-0.4.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: novelai-python
-Version: 0.4.4
+Version: 0.4.5
 Summary: NovelAI Python Binding With Pydantic
 Author-Email: sudoskys <coldlando@hotmail.com>
 License: Apache License 2.0
@@ -51,11 +51,11 @@ The goal of this repository is to use Pydantic to build legitimate requests to a
 - [x] /user/information
 - [x] /ai/upscale
 - [x] /ai/generate-image/suggest-tags
+- [x] /ai/generate-voice
 - [ ] /ai/annotate-image
 - [ ] /ai/classify
 - [ ] /ai/generate-prompt
 - [ ] /ai/generate
-- [ ] /ai/generate-voice
 
 > GenerateImageInfer.calculate_cost is correct in most cases, but please request account information to get accurate
 > consumption information.

{novelai_python-0.4.4 → novelai_python-0.4.5}/README.md

@@ -20,11 +20,11 @@ The goal of this repository is to use Pydantic to build legitimate requests to a
 - [x] /user/information
 - [x] /ai/upscale
 - [x] /ai/generate-image/suggest-tags
+- [x] /ai/generate-voice
 - [ ] /ai/annotate-image
 - [ ] /ai/classify
 - [ ] /ai/generate-prompt
 - [ ] /ai/generate
-- [ ] /ai/generate-voice
 
 > GenerateImageInfer.calculate_cost is correct in most cases, but please request account information to get accurate
 > consumption information.

{novelai_python-0.4.4 → novelai_python-0.4.5}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "novelai-python"
-version = "0.4.4"
+version = "0.4.5"
 description = "NovelAI Python Binding With Pydantic"
 authors = [
     { name = "sudoskys", email = "coldlando@hotmail.com" },

{novelai_python-0.4.4 → novelai_python-0.4.5}/src/novelai_python/__init__.py

@@ -12,6 +12,7 @@ from ._exceptions import (
 )
 from .credential import JwtCredential, LoginCredential, ApiCredential
 from .sdk import GenerateImageInfer, ImageGenerateResp
+from .sdk import VoiceGenerate, VoiceResponse
 from .sdk import SuggestTags, SuggestTagsResp
 from .sdk import Information, InformationResp
 from .sdk import Login, LoginResp
@@ -22,6 +23,9 @@ __all__ = [
     "GenerateImageInfer",
     "ImageGenerateResp",
 
+    "VoiceGenerate",
+    "VoiceResponse",
+
     "Upscale",
     "UpscaleResp",
 

novelai_python-0.4.5/src/novelai_python/_response/ai/generate_voice.py

@@ -0,0 +1,7 @@
+from pydantic import BaseModel, ConfigDict
+
+
+class VoiceResponse(BaseModel):
+    meta: dict
+    audio: bytes
+    model_config = ConfigDict(extra="ignore", arbitrary_types_allowed=True)

{novelai_python-0.4.4 → novelai_python-0.4.5}/src/novelai_python/sdk/__init__.py

@@ -6,6 +6,7 @@
 
 from .ai.generate_image import GenerateImageInfer, ImageGenerateResp  # noqa 401
 from .ai.generate_image.suggest_tags import SuggestTags, SuggestTagsResp  # noqa 401
+from .ai.generate_voice import VoiceGenerate, VoiceResponse  # noqa 401
 from .ai.upscale import Upscale, UpscaleResp  # noqa 401
 from .user.information import Information, InformationResp  # noqa 401
 from .user.login import Login, LoginResp  # noqa 401

novelai_python-0.4.5/src/novelai_python/sdk/ai/_enum.py

@@ -0,0 +1,54 @@
+from typing import Optional
+
+from pydantic import BaseModel
+from enum import Enum
+
+
+class Speaker(BaseModel):
+    """
+    Speaker for /ai/generated_voice
+    """
+    sid: int = -1
+    seed: Optional[str] = None
+    name: str
+    category: str
+
+    @property
+    def version(self):
+        return "v2" if self.sid is None else "v1"
+
+
+class VoiceSpeakerV1(Enum):
+    """
+    Speaker for /ai/generated_voice
+    """
+    Cyllene = Speaker(sid=17, name="Cyllene", category="female")
+    Leucosia = Speaker(sid=95, name="Leucosia", category="female")
+    Crina = Speaker(sid=44, name="Crina", category="female")
+    Hespe = Speaker(sid=80, name="Hespe", category="female")
+    Ida = Speaker(sid=106, name="Ida", category="female")
+    Alseid = Speaker(sid=6, name="Alseid", category="male")
+    Daphnis = Speaker(sid=10, name="Daphnis", category="male")
+    Echo = Speaker(sid=16, name="Echo", category="male")
+    Thel = Speaker(sid=41, name="Thel", category="male")
+    Nomios = Speaker(sid=77, name="Nomios", category="male")
+    # SeedInput = Speaker(sid=-1, name="Seed Input", category="custom")
+
+
+class VoiceSpeakerV2(Enum):
+    """
+    Speaker for /ai/generated_voice
+    """
+    Ligeia = Speaker(name="Ligeia", category="unisex", seed="Anananan")
+    Aini = Speaker(name="Aini", category="female", seed="Aini")
+    Orea = Speaker(name="Orea", category="female", seed="Orea")
+    Claea = Speaker(name="Claea", category="female", seed="Claea")
+    Lim = Speaker(name="Lim", category="female", seed="Lim")
+    Aurae = Speaker(name="Aurae", category="female", seed="Aurae")
+    Naia = Speaker(name="Naia", category="female", seed="Naia")
+    Aulon = Speaker(name="Aulon", category="male", seed="Aulon")
+    Elei = Speaker(name="Elei", category="male", seed="Elei")
+    Ogma = Speaker(name="Ogma", category="male", seed="Ogma")
+    Raid = Speaker(name="Raid", category="male", seed="Raid")
+    Pega = Speaker(name="Pega", category="male", seed="Pega")
+    Lam = Speaker(name="Lam", category="male", seed="Lam")

{novelai_python-0.4.4 → novelai_python-0.4.5}/src/novelai_python/sdk/ai/generate_image/__init__.py

@@ -9,7 +9,7 @@ import math
 import random
 from copy import deepcopy
 from io import BytesIO
-from typing import Optional, Union
+from typing import Optional, Union, Tuple
 from urllib.parse import urlparse
 from zipfile import ZipFile
 
@@ -17,6 +17,7 @@ import curl_cffi
 import cv2
 import httpx
 import numpy as np
+from PIL import Image
 from curl_cffi.requests import AsyncSession
 from loguru import logger
 from pydantic import BaseModel, ConfigDict, PrivateAttr, field_validator, model_validator, Field
@@ -30,7 +31,6 @@ from ...._exceptions import APIError, AuthError, ConcurrentGenerationError, Sess
 from ...._response.ai.generate_image import ImageGenerateResp
 from ....credential import CredentialBase
 from ....utils import try_jsonfy
-from PIL import Image, ImageEnhance
 
 
 class GenerateImageInfer(ApiBaseModel):
@@ -159,58 +159,43 @@ class GenerateImageInfer(ApiBaseModel):
         @staticmethod
         def add_image_to_black_background(
                 image: Union[str, bytes],
-                width: int = 448,
-                height: int = 448,
-                transparency: bool = True
+                target_size: Tuple[int, int] = (448, 448),
+                transparency: bool = False
         ):
 
             """
             缩放图像到指定的黑色透明背景上，使其尽可能大且保持比例。
             :param transparency: 是否透明
             :param image: 图像
-            :param width: 宽
-            :param height: 高
+            :param target_size: 目标尺寸
             :return: 新图像
             """
 
             if isinstance(image, str):
                 image = base64.b64decode(image)
-
-            # Decode the image from the base64 string
+                # Decode the image from the base64 string
             npimg = np.frombuffer(image, np.uint8)
-
-            # Read the image using OpenCV
-            open_image = cv2.imdecode(npimg, cv2.IMREAD_UNCHANGED)
-
-            # Calculate the ratio for scaling
-            width_ratio = width / open_image.shape[1]
-            height_ratio = height / open_image.shape[0]
-            ratio = min(width_ratio, height_ratio)
-
-            # Create new image size
-            new_image_size = (int(open_image.shape[1] * ratio), int(open_image.shape[0] * ratio))
-
+            # Load the image with OpenCV
+            img = cv2.imdecode(npimg, cv2.IMREAD_UNCHANGED)
+            # Calculate the aspect ratio (keeping aspect ratio)
+            ratio = min(target_size[0] / img.shape[1], target_size[1] / img.shape[0])
+            new_image_size = (int(img.shape[1] * ratio), int(img.shape[0] * ratio))
             # Resize the image
-            open_image = cv2.resize(open_image, new_image_size, interpolation=cv2.INTER_LINEAR)
-
-            # Create black background
-            if open_image.shape[2] == 3:  # no alpha channel, add one
-                open_image = cv2.cvtColor(open_image, cv2.COLOR_BGR2BGRA)
-            if transparency:
-                bg_color = [0, 0, 0, 0]
-            else:
-                bg_color = [0, 0, 0, 255]
-            new_image = np.full((height, width, 4), bg_color, dtype=np.uint8)
-
-            # Compute coords to place image on black background
-            position = ((width - open_image.shape[1]) // 2, (height - open_image.shape[0]) // 2)
-
-            # Place the image on black background at the calculated position
-            new_image[position[1]:position[1] + open_image.shape[0],
-            position[0]:position[0] + open_image.shape[1]] = open_image
-
-            # Encode image to base64 string
-            _, buffer = cv2.imencode('.png', new_image)
+            img = cv2.resize(img, new_image_size, interpolation=cv2.INTER_LINEAR)
+            # Check if the image has alpha channel (transparency)
+            if img.shape[2] == 3:  # no alpha channel, add one
+                img = cv2.cvtColor(img, cv2.COLOR_BGR2BGRA)
+            # Create new image with black background
+            delta_w = target_size[0] - img.shape[1]
+            delta_h = target_size[1] - img.shape[0]
+            top, bottom = delta_h // 2, delta_h - (delta_h // 2)
+            left, right = delta_w // 2, delta_w - (delta_w // 2)
+            # If the transparency flag is set the background is transparent, otherwise it is black
+            color = [0, 0, 0, 0] if transparency else [0, 0, 0, 255]
+            # Add padding to the image
+            new_img = cv2.copyMakeBorder(img, top, bottom, left, right, cv2.BORDER_CONSTANT, value=color)
+            # Convert the image to base64
+            _, buffer = cv2.imencode(".png", new_img)
             return base64.b64encode(buffer).decode("utf-8")
 
         # Validators
@@ -289,7 +274,7 @@ class GenerateImageInfer(ApiBaseModel):
                 "lowres, {bad}, error, fewer, extra, missing, worst quality, jpeg artifacts, bad quality, "
                 "watermark, unfinished, displeasing, chromatic aberration, signature, extra digits, artistic error, "
                 "username, scan, [abstract], bad anatomy, bad hands, @_@, mismatched pupils, heart-shaped pupils, "
-                "glowing eyes, nsfw, lowres, bad anatomy, bad hands, text, error, missing fingers, "
+                "glowing eyes, lowres, bad anatomy, bad hands, text, error, missing fingers, "
                 "extra digit, fewer digits, cropped, worst quality, low quality, normal quality, "
                 f"jpeg artifacts, signature, watermark, username, blurry "
                 f",{self.parameters.negative_prompt}"

novelai_python-0.4.5/src/novelai_python/sdk/ai/generate_voice.py

@@ -0,0 +1,201 @@
+# -*- coding: utf-8 -*-
+# @Author  : sudoskys
+import base64
+import json
+from io import BytesIO
+from typing import Optional, Union, Literal
+from urllib.parse import urlparse
+from zipfile import ZipFile
+
+import curl_cffi
+import httpx
+import pydantic
+from curl_cffi.requests import AsyncSession
+from loguru import logger
+from pydantic import ConfigDict, PrivateAttr, Field, model_validator
+from tenacity import wait_random, retry, stop_after_attempt, retry_if_exception
+
+from ..schema import ApiBaseModel
+from ..._exceptions import APIError, AuthError, SessionHttpError
+from ..._response.ai.generate_voice import VoiceResponse
+from ...credential import CredentialBase
+from ...utils import try_jsonfy
+from ._enum import VoiceSpeakerV1, VoiceSpeakerV2
+
+
+class VoiceGenerate(ApiBaseModel):
+    """
+    Voice Generate for /ai/generate-voice
+    """
+    _endpoint: str = PrivateAttr("https://api.novelai.net")
+    text: str = Field(..., description="Text to generate voice", max_length=1000)
+    """Limit 1000 characters"""
+    voice: int = Field(default=-1, description="Seed", ge=-1)
+    """Voice engine sid"""
+    seed: Optional[str] = Field(None, description="seed_seed", max_length=50)
+    """
+    Starting with a common first name will have a relevant influence on pitch and intonation. 
+    The voice for any given seed is liable to change in the future as we continue to develop the TTS.
+    """
+    opus: bool = False
+    version: Union[Literal["v2", "v1"], str] = "v2"
+
+    model_config = ConfigDict(extra="ignore")
+
+    @model_validator(mode="after")
+    def validate_model(self):
+        # Validate seed
+        if self.voice == -1 and self.seed is None:
+            raise pydantic.ValidationError("Seed must be set when voice is -1")
+        return self
+
+    @property
+    def base_url(self):
+        return f"{self.endpoint.strip('/')}/ai/generate-voice"
+
+    @property
+    def endpoint(self):
+        return self._endpoint
+
+    @endpoint.setter
+    def endpoint(self, value):
+        self._endpoint = value
+
+    async def necessary_headers(self, request_data) -> dict:
+        """
+        :param request_data: dict
+        :return: dict
+        "Sec-Ch-Ua": '"Edge";v="123", "Not:A-Brand";v="8"',
+        "Sec-Ch-Ua-Mobile": "?0",
+        "Sec-Ch-Ua-Platform": '"Windows"',
+        "Sec-Fetch-Dest": "empty",
+        "Sec-Fetch-Mode": "cors",
+        "Sec-Fetch-Site": "same-site",
+        """
+        return {
+            "Host": urlparse(self.endpoint).netloc,
+            "Accept": "*/*",
+            "Accept-Language": "zh-CN,zh;q=0.9",
+            "Cache-Control": "no-cache",
+            "Origin": "https://novelai.net",
+            "Pragma": "no-cache",
+            "Referer": "https://novelai.net/",
+        }
+
+    @classmethod
+    def build(cls,
+              text: str,
+              voice_engine: Union[VoiceSpeakerV1, VoiceSpeakerV2, str],
+              *,
+              opus: bool = False
+              ) -> "VoiceGenerate":
+        """
+        生成图片
+        :param opus: unknown
+        :param text: str
+        :param voice_engine: VoiceSpeakerV1 or VoiceSpeakerV2 or str
+        :return: VoiceGenerate instance
+        :raises: ValueError
+        """
+        if isinstance(voice_engine, str):
+            return cls(
+                text=text,
+                voice=-1,
+                seed=voice_engine,
+                opus=opus,
+                version="v2"
+            )
+        if isinstance(voice_engine, VoiceSpeakerV2):
+            return cls(
+                text=text,
+                seed=voice_engine.value.seed,
+                voice=voice_engine.value.sid,
+                opus=opus,
+                version="v2"
+            )
+        if isinstance(voice_engine, VoiceSpeakerV1):
+            return cls(
+                text=text,
+                seed=voice_engine.value.seed,
+                voice=voice_engine.value.sid,
+                opus=opus,
+                version="v1"
+            )
+        raise ValueError("Invalid voice engine")
+
+    @retry(
+        wait=wait_random(min=1, max=3),
+        stop=stop_after_attempt(3),
+        retry=retry_if_exception(lambda e: hasattr(e, "code") and str(e.code) == "500"),
+        reraise=True
+    )
+    async def request(self,
+                      session: Union[AsyncSession, "CredentialBase"],
+                      *,
+                      override_headers: Optional[dict] = None
+                      ) -> VoiceResponse:
+        """
+        生成图片
+        :param override_headers:
+        :param session:  session
+        :return: VoiceResponse
+        """
+        # Data Build
+        request_data = self.model_dump(mode="json", exclude_none=True)
+        # Header
+        if isinstance(session, AsyncSession):
+            session.headers.update(await self.necessary_headers(request_data))
+        elif isinstance(session, CredentialBase):
+            update_header = await self.necessary_headers(request_data)
+            session = await session.get_session(update_headers=update_header)
+        if override_headers:
+            session.headers.clear()
+            session.headers.update(override_headers)
+        # Log
+        logger.debug(f"Voice request data: {request_data}")
+        # Request
+        try:
+            assert hasattr(session, "get"), "session must have get method."
+            response = await session.get(
+                self.base_url,
+                params=request_data
+            )
+            header_type = response.headers.get('Content-Type')
+            if header_type not in ['audio/mpeg', 'audio/ogg', 'audio/opus']:
+                logger.warning(
+                    f"Error with content type: {header_type} and code: {response.status_code}"
+                )
+                try:
+                    _msg = response.json()
+                except Exception as e:
+                    logger.warning(e)
+                    if not isinstance(response.content, str) and len(response.content) < 50:
+                        raise APIError(
+                            message=f"Unexpected content: {header_type} with code: {response.status_code}",
+                            request=request_data,
+                            code=response.status_code,
+                            response="UnJsoned content"
+                        )
+                    else:
+                        _msg = {"statusCode": response.status_code, "message": response.content}
+                status_code = _msg.get("statusCode", response.status_code)
+                message = _msg.get("message", "Unknown error")
+                if status_code in [400]:
+                    # Validation tts version error
+                    raise APIError(message, request=request_data, code=status_code, response=_msg)
+                raise APIError(message, request=request_data, code=status_code, response=_msg)
+            return VoiceResponse(
+                meta=request_data,
+                audio=response.content,
+            )
+        except curl_cffi.requests.errors.RequestsError as exc:
+            logger.exception(exc)
+            raise SessionHttpError("An AsyncSession RequestsError occurred, maybe SSL error. Try again later!")
+        except httpx.HTTPError as exc:
+            logger.exception(exc)
+            raise SessionHttpError("An HTTPError occurred, maybe SSL error. Try again later!")
+        except APIError as e:
+            raise e
+        except Exception as e:
+            logger.opt(exception=e).exception("An Unexpected error occurred")
+            raise e

novelai_python-0.4.5/tests/test_generate_voice.py

@@ -0,0 +1,34 @@
+import pytest
+from unittest import mock
+
+from curl_cffi.requests import AsyncSession
+
+from src.novelai_python.sdk.ai.generate_voice import VoiceGenerate, VoiceResponse
+
+
+@pytest.mark.asyncio
+async def test_request():
+    # Arrange
+    mock_response = mock.MagicMock()
+    mock_response.content = b'audio_content'
+    mock_response.headers = {'Content-Type': 'audio/mpeg'}
+    session = mock.MagicMock(spec=AsyncSession)
+    session.get = mock.AsyncMock(return_value=mock_response)
+    session.headers = {}
+
+    voice_generate = VoiceGenerate(
+        text="Hello, world!",
+        voice=-1,
+        seed="seed",
+        opus=False,
+        version="v2"
+    )
+
+    # Act
+    result = await voice_generate.request(session=session)
+
+    # Assert
+    session.get.assert_called_once()
+    assert isinstance(result, VoiceResponse)
+    assert result.audio == b'audio_content'
+