feat(api-nodes): add new Gemini model (#10789)
This commit is contained in:
Alexander Piskun
GitHub
@@ -1,22 +1,229 @@
|
||||
from typing import Optional
|
||||
from datetime import date
|
||||
from enum import Enum
|
||||
from typing import Any
|
||||
|
||||
from comfy_api_nodes.apis import GeminiGenerationConfig, GeminiContent, GeminiSafetySetting, GeminiSystemInstructionContent, GeminiTool, GeminiVideoMetadata
|
||||
from pydantic import BaseModel
|
||||
from pydantic import BaseModel, Field
|
||||
|
||||
|
||||
class GeminiSafetyCategory(str, Enum):
|
||||
HARM_CATEGORY_SEXUALLY_EXPLICIT = "HARM_CATEGORY_SEXUALLY_EXPLICIT"
|
||||
HARM_CATEGORY_HATE_SPEECH = "HARM_CATEGORY_HATE_SPEECH"
|
||||
HARM_CATEGORY_HARASSMENT = "HARM_CATEGORY_HARASSMENT"
|
||||
HARM_CATEGORY_DANGEROUS_CONTENT = "HARM_CATEGORY_DANGEROUS_CONTENT"
|
||||
|
||||
|
||||
class GeminiSafetyThreshold(str, Enum):
|
||||
OFF = "OFF"
|
||||
BLOCK_NONE = "BLOCK_NONE"
|
||||
BLOCK_LOW_AND_ABOVE = "BLOCK_LOW_AND_ABOVE"
|
||||
BLOCK_MEDIUM_AND_ABOVE = "BLOCK_MEDIUM_AND_ABOVE"
|
||||
BLOCK_ONLY_HIGH = "BLOCK_ONLY_HIGH"
|
||||
|
||||
|
||||
class GeminiSafetySetting(BaseModel):
|
||||
category: GeminiSafetyCategory
|
||||
threshold: GeminiSafetyThreshold
|
||||
|
||||
|
||||
class GeminiRole(str, Enum):
|
||||
user = "user"
|
||||
model = "model"
|
||||
|
||||
|
||||
class GeminiMimeType(str, Enum):
|
||||
application_pdf = "application/pdf"
|
||||
audio_mpeg = "audio/mpeg"
|
||||
audio_mp3 = "audio/mp3"
|
||||
audio_wav = "audio/wav"
|
||||
image_png = "image/png"
|
||||
image_jpeg = "image/jpeg"
|
||||
image_webp = "image/webp"
|
||||
text_plain = "text/plain"
|
||||
video_mov = "video/mov"
|
||||
video_mpeg = "video/mpeg"
|
||||
video_mp4 = "video/mp4"
|
||||
video_mpg = "video/mpg"
|
||||
video_avi = "video/avi"
|
||||
video_wmv = "video/wmv"
|
||||
video_mpegps = "video/mpegps"
|
||||
video_flv = "video/flv"
|
||||
|
||||
|
||||
class GeminiInlineData(BaseModel):
|
||||
data: str | None = Field(
|
||||
None,
|
||||
description="The base64 encoding of the image, PDF, or video to include inline in the prompt. "
|
||||
"When including media inline, you must also specify the media type (mimeType) of the data. Size limit: 20MB",
|
||||
)
|
||||
mimeType: GeminiMimeType | None = Field(None)
|
||||
|
||||
|
||||
class GeminiPart(BaseModel):
|
||||
inlineData: GeminiInlineData | None = Field(None)
|
||||
text: str | None = Field(None)
|
||||
|
||||
|
||||
class GeminiTextPart(BaseModel):
|
||||
text: str | None = Field(None)
|
||||
|
||||
|
||||
class GeminiContent(BaseModel):
|
||||
parts: list[GeminiPart] = Field(...)
|
||||
role: GeminiRole = Field(..., examples=["user"])
|
||||
|
||||
|
||||
class GeminiSystemInstructionContent(BaseModel):
|
||||
parts: list[GeminiTextPart] = Field(
|
||||
...,
|
||||
description="A list of ordered parts that make up a single message. "
|
||||
"Different parts may have different IANA MIME types.",
|
||||
)
|
||||
role: GeminiRole = Field(
|
||||
...,
|
||||
description="The identity of the entity that creates the message. "
|
||||
"The following values are supported: "
|
||||
"user: This indicates that the message is sent by a real person, typically a user-generated message. "
|
||||
"model: This indicates that the message is generated by the model. "
|
||||
"The model value is used to insert messages from model into the conversation during multi-turn conversations. "
|
||||
"For non-multi-turn conversations, this field can be left blank or unset.",
|
||||
)
|
||||
|
||||
|
||||
class GeminiFunctionDeclaration(BaseModel):
|
||||
description: str | None = Field(None)
|
||||
name: str = Field(...)
|
||||
parameters: dict[str, Any] = Field(..., description="JSON schema for the function parameters")
|
||||
|
||||
|
||||
class GeminiTool(BaseModel):
|
||||
functionDeclarations: list[GeminiFunctionDeclaration] | None = Field(None)
|
||||
|
||||
|
||||
class GeminiOffset(BaseModel):
|
||||
nanos: int | None = Field(None, ge=0, le=999999999)
|
||||
seconds: int | None = Field(None, ge=-315576000000, le=315576000000)
|
||||
|
||||
|
||||
class GeminiVideoMetadata(BaseModel):
|
||||
endOffset: GeminiOffset | None = Field(None)
|
||||
startOffset: GeminiOffset | None = Field(None)
|
||||
|
||||
|
||||
class GeminiGenerationConfig(BaseModel):
|
||||
maxOutputTokens: int | None = Field(None, ge=16, le=8192)
|
||||
seed: int | None = Field(None)
|
||||
stopSequences: list[str] | None = Field(None)
|
||||
temperature: float | None = Field(1, ge=0.0, le=2.0)
|
||||
topK: int | None = Field(40, ge=1)
|
||||
topP: float | None = Field(0.95, ge=0.0, le=1.0)
|
||||
|
||||
|
||||
class GeminiImageConfig(BaseModel):
|
||||
aspectRatio: Optional[str] = None
|
||||
aspectRatio: str | None = Field(None)
|
||||
resolution: str | None = Field(None)
|
||||
|
||||
|
||||
class GeminiImageGenerationConfig(GeminiGenerationConfig):
|
||||
responseModalities: Optional[list[str]] = None
|
||||
imageConfig: Optional[GeminiImageConfig] = None
|
||||
responseModalities: list[str] | None = Field(None)
|
||||
imageConfig: GeminiImageConfig | None = Field(None)
|
||||
|
||||
|
||||
class GeminiImageGenerateContentRequest(BaseModel):
|
||||
contents: list[GeminiContent]
|
||||
generationConfig: Optional[GeminiImageGenerationConfig] = None
|
||||
safetySettings: Optional[list[GeminiSafetySetting]] = None
|
||||
systemInstruction: Optional[GeminiSystemInstructionContent] = None
|
||||
tools: Optional[list[GeminiTool]] = None
|
||||
videoMetadata: Optional[GeminiVideoMetadata] = None
|
||||
contents: list[GeminiContent] = Field(...)
|
||||
generationConfig: GeminiImageGenerationConfig | None = Field(None)
|
||||
safetySettings: list[GeminiSafetySetting] | None = Field(None)
|
||||
systemInstruction: GeminiSystemInstructionContent | None = Field(None)
|
||||
tools: list[GeminiTool] | None = Field(None)
|
||||
videoMetadata: GeminiVideoMetadata | None = Field(None)
|
||||
|
||||
|
||||
class GeminiGenerateContentRequest(BaseModel):
|
||||
contents: list[GeminiContent] = Field(...)
|
||||
generationConfig: GeminiGenerationConfig | None = Field(None)
|
||||
safetySettings: list[GeminiSafetySetting] | None = Field(None)
|
||||
systemInstruction: GeminiSystemInstructionContent | None = Field(None)
|
||||
tools: list[GeminiTool] | None = Field(None)
|
||||
videoMetadata: GeminiVideoMetadata | None = Field(None)
|
||||
|
||||
|
||||
class Modality(str, Enum):
|
||||
MODALITY_UNSPECIFIED = "MODALITY_UNSPECIFIED"
|
||||
TEXT = "TEXT"
|
||||
IMAGE = "IMAGE"
|
||||
VIDEO = "VIDEO"
|
||||
AUDIO = "AUDIO"
|
||||
DOCUMENT = "DOCUMENT"
|
||||
|
||||
|
||||
class ModalityTokenCount(BaseModel):
|
||||
modality: Modality | None = None
|
||||
tokenCount: int | None = Field(None, description="Number of tokens for the given modality.")
|
||||
|
||||
|
||||
class Probability(str, Enum):
|
||||
NEGLIGIBLE = "NEGLIGIBLE"
|
||||
LOW = "LOW"
|
||||
MEDIUM = "MEDIUM"
|
||||
HIGH = "HIGH"
|
||||
UNKNOWN = "UNKNOWN"
|
||||
|
||||
|
||||
class GeminiSafetyRating(BaseModel):
|
||||
category: GeminiSafetyCategory | None = None
|
||||
probability: Probability | None = Field(
|
||||
None,
|
||||
description="The probability that the content violates the specified safety category",
|
||||
)
|
||||
|
||||
|
||||
class GeminiCitation(BaseModel):
|
||||
authors: list[str] | None = None
|
||||
endIndex: int | None = None
|
||||
license: str | None = None
|
||||
publicationDate: date | None = None
|
||||
startIndex: int | None = None
|
||||
title: str | None = None
|
||||
uri: str | None = None
|
||||
|
||||
|
||||
class GeminiCitationMetadata(BaseModel):
|
||||
citations: list[GeminiCitation] | None = None
|
||||
|
||||
|
||||
class GeminiCandidate(BaseModel):
|
||||
citationMetadata: GeminiCitationMetadata | None = None
|
||||
content: GeminiContent | None = None
|
||||
finishReason: str | None = None
|
||||
safetyRatings: list[GeminiSafetyRating] | None = None
|
||||
|
||||
|
||||
class GeminiPromptFeedback(BaseModel):
|
||||
blockReason: str | None = None
|
||||
blockReasonMessage: str | None = None
|
||||
safetyRatings: list[GeminiSafetyRating] | None = None
|
||||
|
||||
|
||||
class GeminiUsageMetadata(BaseModel):
|
||||
cachedContentTokenCount: int | None = Field(
|
||||
None,
|
||||
description="Output only. Number of tokens in the cached part in the input (the cached content).",
|
||||
)
|
||||
candidatesTokenCount: int | None = Field(None, description="Number of tokens in the response(s).")
|
||||
candidatesTokensDetails: list[ModalityTokenCount] | None = Field(
|
||||
None, description="Breakdown of candidate tokens by modality."
|
||||
)
|
||||
promptTokenCount: int | None = Field(
|
||||
None,
|
||||
description="Number of tokens in the request. When cachedContent is set, this is still the total effective prompt size meaning this includes the number of tokens in the cached content.",
|
||||
)
|
||||
promptTokensDetails: list[ModalityTokenCount] | None = Field(
|
||||
None, description="Breakdown of prompt tokens by modality."
|
||||
)
|
||||
thoughtsTokenCount: int | None = Field(None, description="Number of tokens present in thoughts output.")
|
||||
toolUsePromptTokenCount: int | None = Field(None, description="Number of tokens present in tool-use prompt(s).")
|
||||
|
||||
|
||||
class GeminiGenerateContentResponse(BaseModel):
|
||||
candidates: list[GeminiCandidate] | None = Field(None)
|
||||
promptFeedback: GeminiPromptFeedback | None = Field(None)
|
||||
usageMetadata: GeminiUsageMetadata | None = Field(None)
|
||||
|
||||
Reference in New Issue
Block a user