mirror of
https://github.com/Mintplex-Labs/anything-llm.git
synced 2024-11-19 04:30:10 +01:00
Added vertex
This commit is contained in:
parent
0bb47619dc
commit
2fccfcdd2b
@ -96,6 +96,7 @@ AnythingLLM divides your documents into objects called `workspaces`. A Workspace
|
||||
- [Text Generation Web UI](https://github.com/oobabooga/text-generation-webui)
|
||||
- [Apipie](https://apipie.ai/)
|
||||
- [xAI](https://x.ai/)
|
||||
- [Google Vertex](https://cloud.google.com/vertex-ai)
|
||||
|
||||
**Embedder models:**
|
||||
|
||||
|
@ -0,0 +1,80 @@
|
||||
export default function VertexLLMOptions({ settings }) {
|
||||
return (
|
||||
<div className="w-full flex flex-col">
|
||||
<div className="w-full flex items-center gap-[36px] mt-1.5">
|
||||
<div className="flex flex-col w-60">
|
||||
<label className="text-white text-sm font-semibold block mb-3">
|
||||
GCP Project Name
|
||||
</label>
|
||||
<input
|
||||
type="text"
|
||||
name="VertexProjectName"
|
||||
className="bg-zinc-900 text-white placeholder:text-white/20 text-sm rounded-lg focus:outline-primary-button active:outline-primary-button outline-none block w-full p-2.5"
|
||||
placeholder="your-project-name"
|
||||
defaultValue={settings?.VertexProjectName ?? ""}
|
||||
required={true}
|
||||
autoComplete="off"
|
||||
spellCheck={false}
|
||||
/>
|
||||
</div>
|
||||
{!settings?.credentialsOnly && (
|
||||
<VertexLLMModelSelection settings={settings} />
|
||||
)}
|
||||
</div>
|
||||
</div>
|
||||
);
|
||||
|
||||
function VertexLLMModelSelection({ settings }) {
|
||||
console.log(settings);
|
||||
return (
|
||||
<>
|
||||
<div className="flex flex-col w-60">
|
||||
<label className="text-white text-sm font-semibold block mb-3">
|
||||
Chat Model Selection
|
||||
</label>
|
||||
<select
|
||||
name="VertexLLMModelPref"
|
||||
defaultValue={settings?.VertexLLMModelPref || "gemini-1.5-flash"}
|
||||
required={true}
|
||||
className="bg-zinc-900 border-gray-500 text-white text-sm rounded-lg block w-full p-2.5"
|
||||
>
|
||||
<optgroup label="Stable Models">
|
||||
{[
|
||||
"gemini-1.5-flash",
|
||||
"gemini-1.5-pro",
|
||||
"gemini-1.5-flash-001",
|
||||
"gemini-1.5-pro-001",
|
||||
"gemini-1.5-flash-002",
|
||||
"gemini-1.5-pro-002",
|
||||
].map((model) => {
|
||||
return (
|
||||
<option key={model} value={model}>
|
||||
{model}
|
||||
</option>
|
||||
);
|
||||
})}
|
||||
</optgroup>
|
||||
</select>
|
||||
</div>
|
||||
<div className="flex flex-col w-60">
|
||||
<label className="text-white text-sm font-semibold block mb-3">
|
||||
Safety Setting
|
||||
</label>
|
||||
<select
|
||||
name="VertexSafetySetting"
|
||||
defaultValue={
|
||||
settings?.VertexSafetySetting || "BLOCK_MEDIUM_AND_ABOVE"
|
||||
}
|
||||
required={true}
|
||||
className="bg-zinc-900 border-gray-500 text-white text-sm rounded-lg block w-full p-2.5"
|
||||
>
|
||||
<option value="BLOCK_NONE">None</option>
|
||||
<option value="BLOCK_ONLY_HIGH">Block few</option>
|
||||
<option value="BLOCK_MEDIUM_AND_ABOVE">Block some (default)</option>
|
||||
<option value="BLOCK_LOW_AND_ABOVE">Block most</option>
|
||||
</select>
|
||||
</div>
|
||||
</>
|
||||
);
|
||||
}
|
||||
}
|
@ -11,6 +11,14 @@ export const DISABLED_PROVIDERS = [
|
||||
];
|
||||
const PROVIDER_DEFAULT_MODELS = {
|
||||
openai: [],
|
||||
vertex: [
|
||||
"gemini-1.5-flash",
|
||||
"gemini-1.5-pro",
|
||||
"gemini-1.5-flash-001",
|
||||
"gemini-1.5-pro-001",
|
||||
"gemini-1.5-flash-002",
|
||||
"gemini-1.5-pro-002",
|
||||
],
|
||||
gemini: [
|
||||
"gemini-pro",
|
||||
"gemini-1.0-pro",
|
||||
|
BIN
frontend/src/media/llmprovider/vertex.png
Normal file
BIN
frontend/src/media/llmprovider/vertex.png
Normal file
Binary file not shown.
After Width: | Height: | Size: 69 KiB |
@ -10,6 +10,7 @@ import GenericOpenAiLogo from "@/media/llmprovider/generic-openai.png";
|
||||
import AzureOpenAiLogo from "@/media/llmprovider/azure.png";
|
||||
import AnthropicLogo from "@/media/llmprovider/anthropic.png";
|
||||
import GeminiLogo from "@/media/llmprovider/gemini.png";
|
||||
import VertexLogo from "@/media/llmprovider/vertex.png";
|
||||
import OllamaLogo from "@/media/llmprovider/ollama.png";
|
||||
import LMStudioLogo from "@/media/llmprovider/lmstudio.png";
|
||||
import LocalAiLogo from "@/media/llmprovider/localai.png";
|
||||
@ -38,6 +39,7 @@ import LMStudioOptions from "@/components/LLMSelection/LMStudioOptions";
|
||||
import LocalAiOptions from "@/components/LLMSelection/LocalAiOptions";
|
||||
import NativeLLMOptions from "@/components/LLMSelection/NativeLLMOptions";
|
||||
import GeminiLLMOptions from "@/components/LLMSelection/GeminiLLMOptions";
|
||||
import VertexLLMOptions from "@/components/LLMSelection/VertexLLMOptions";
|
||||
import OllamaLLMOptions from "@/components/LLMSelection/OllamaLLMOptions";
|
||||
import TogetherAiOptions from "@/components/LLMSelection/TogetherAiOptions";
|
||||
import FireworksAiOptions from "@/components/LLMSelection/FireworksAiOptions";
|
||||
@ -92,6 +94,14 @@ export const AVAILABLE_LLM_PROVIDERS = [
|
||||
description: "Google's largest and most capable AI model",
|
||||
requiredConfig: ["GeminiLLMApiKey"],
|
||||
},
|
||||
{
|
||||
name: "Vertex",
|
||||
value: "vertex",
|
||||
logo: VertexLogo,
|
||||
options: (settings) => <VertexLLMOptions settings={settings} />,
|
||||
description: "Google's Genereative AI model platform",
|
||||
requiredConfig: ["VertexProjectName"],
|
||||
},
|
||||
{
|
||||
name: "HuggingFace",
|
||||
value: "huggingface",
|
||||
|
@ -6,6 +6,7 @@ import GenericOpenAiLogo from "@/media/llmprovider/generic-openai.png";
|
||||
import AzureOpenAiLogo from "@/media/llmprovider/azure.png";
|
||||
import AnthropicLogo from "@/media/llmprovider/anthropic.png";
|
||||
import GeminiLogo from "@/media/llmprovider/gemini.png";
|
||||
import VertexLogo from "@/media/llmprovider/vertex.png";
|
||||
import OllamaLogo from "@/media/llmprovider/ollama.png";
|
||||
import TogetherAILogo from "@/media/llmprovider/togetherai.png";
|
||||
import FireworksAILogo from "@/media/llmprovider/fireworksai.jpeg";
|
||||
@ -75,6 +76,14 @@ export const LLM_SELECTION_PRIVACY = {
|
||||
],
|
||||
logo: GeminiLogo,
|
||||
},
|
||||
vertex: {
|
||||
name: "Vertex",
|
||||
description: [
|
||||
"Your chats will not be used for training",
|
||||
"Your prompts and document text used in response creation are visible to Google",
|
||||
],
|
||||
logo: VertexLogo,
|
||||
},
|
||||
lmstudio: {
|
||||
name: "LMStudio",
|
||||
description: [
|
||||
|
@ -5,6 +5,7 @@ import GenericOpenAiLogo from "@/media/llmprovider/generic-openai.png";
|
||||
import AzureOpenAiLogo from "@/media/llmprovider/azure.png";
|
||||
import AnthropicLogo from "@/media/llmprovider/anthropic.png";
|
||||
import GeminiLogo from "@/media/llmprovider/gemini.png";
|
||||
import VertexLogo from "@/media/llmprovider/vertex.png";
|
||||
import OllamaLogo from "@/media/llmprovider/ollama.png";
|
||||
import LMStudioLogo from "@/media/llmprovider/lmstudio.png";
|
||||
import LocalAiLogo from "@/media/llmprovider/localai.png";
|
||||
@ -33,6 +34,7 @@ import LMStudioOptions from "@/components/LLMSelection/LMStudioOptions";
|
||||
import LocalAiOptions from "@/components/LLMSelection/LocalAiOptions";
|
||||
import NativeLLMOptions from "@/components/LLMSelection/NativeLLMOptions";
|
||||
import GeminiLLMOptions from "@/components/LLMSelection/GeminiLLMOptions";
|
||||
import VertexLLMOptions from "@/components/LLMSelection/VertexLLMOptions";
|
||||
import OllamaLLMOptions from "@/components/LLMSelection/OllamaLLMOptions";
|
||||
import MistralOptions from "@/components/LLMSelection/MistralOptions";
|
||||
import HuggingFaceOptions from "@/components/LLMSelection/HuggingFaceOptions";
|
||||
@ -89,6 +91,13 @@ const LLMS = [
|
||||
options: (settings) => <GeminiLLMOptions settings={settings} />,
|
||||
description: "Google's largest and most capable AI model",
|
||||
},
|
||||
{
|
||||
name: "Vertex",
|
||||
value: "vertex",
|
||||
logo: VertexLogo,
|
||||
options: (settings) => <VertexLLMOptions settings={settings} />,
|
||||
description: "Google's Genereative AI model platform",
|
||||
},
|
||||
{
|
||||
name: "HuggingFace",
|
||||
value: "huggingface",
|
||||
|
@ -430,6 +430,13 @@ const SystemSettings = {
|
||||
GeminiSafetySetting:
|
||||
process.env.GEMINI_SAFETY_SETTING || "BLOCK_MEDIUM_AND_ABOVE",
|
||||
|
||||
// Vertex
|
||||
VertexLLMModelPref:
|
||||
process.env.VERTEX_LLM_MODEL_PREF || "gemini-1.5-flash",
|
||||
VertexSafetySetting:
|
||||
process.env.VERTEX_SAFETY_SETTING || "BLOCK_MEDIUM_AND_ABOVE",
|
||||
VertexProjectName: process.env.VERTEX_PROJECT_NAME,
|
||||
|
||||
// LMStudio Keys
|
||||
LMStudioBasePath: process.env.LMSTUDIO_BASE_PATH,
|
||||
LMStudioTokenLimit: process.env.LMSTUDIO_MODEL_TOKEN_LIMIT,
|
||||
|
@ -22,6 +22,7 @@
|
||||
"@anthropic-ai/sdk": "^0.16.1",
|
||||
"@azure/openai": "1.0.0-beta.10",
|
||||
"@datastax/astra-db-ts": "^0.1.3",
|
||||
"@google-cloud/vertexai": "^1.9.0",
|
||||
"@google/generative-ai": "^0.7.1",
|
||||
"@ladjs/graceful": "^3.2.2",
|
||||
"@lancedb/lancedb": "0.5.2",
|
||||
@ -92,8 +93,8 @@
|
||||
"flow-remove-types": "^2.217.1",
|
||||
"globals": "^13.21.0",
|
||||
"hermes-eslint": "^0.15.0",
|
||||
"nodemon": "^2.0.22",
|
||||
"node-html-markdown": "^1.3.0",
|
||||
"nodemon": "^2.0.22",
|
||||
"prettier": "^3.0.3"
|
||||
}
|
||||
}
|
||||
}
|
||||
|
@ -30,6 +30,14 @@ const MODEL_MAP = {
|
||||
"gemini-1.5-flash-exp-0827": 1_048_576,
|
||||
"gemini-1.5-flash-8b-exp-0827": 1_048_576,
|
||||
},
|
||||
vertex: {
|
||||
"gemini-1.5-flash": 1_048_576,
|
||||
"gemini-1.5-pro": 2_097_152,
|
||||
"gemini-1.5-flash-001": 1_048_576,
|
||||
"gemini-1.5-pro-001": 2_097_152,
|
||||
"gemini-1.5-flash-002": 1_048_576,
|
||||
"gemini-1.5-pro-002": 2_097_152,
|
||||
},
|
||||
groq: {
|
||||
"gemma2-9b-it": 8192,
|
||||
"gemma-7b-it": 8192,
|
||||
|
296
server/utils/AiProviders/vertex/index.js
Normal file
296
server/utils/AiProviders/vertex/index.js
Normal file
@ -0,0 +1,296 @@
|
||||
const { NativeEmbedder } = require("../../EmbeddingEngines/native");
|
||||
const {
|
||||
writeResponseChunk,
|
||||
clientAbortedHandler,
|
||||
} = require("../../helpers/chat/responses");
|
||||
const { MODEL_MAP } = require("../modelMap");
|
||||
|
||||
class VertexLLM {
|
||||
constructor(embedder = null, modelPreference = null) {
|
||||
const { VertexAI } = require("@google-cloud/vertexai");
|
||||
const vertexAI = new VertexAI({
|
||||
project: process.env.VERTEX_PROJECT_NAME,
|
||||
});
|
||||
this.model =
|
||||
modelPreference ||
|
||||
process.env.VERTEX_LLM_MODEL_PREF ||
|
||||
"gemini-1.5-flash";
|
||||
this.vertex = vertexAI.getGenerativeModel({
|
||||
model: this.model,
|
||||
});
|
||||
this.limits = {
|
||||
history: this.promptWindowLimit() * 0.15,
|
||||
system: this.promptWindowLimit() * 0.15,
|
||||
user: this.promptWindowLimit() * 0.7,
|
||||
};
|
||||
|
||||
this.embedder = embedder ?? new NativeEmbedder();
|
||||
this.defaultTemp = 0.7;
|
||||
this.safetyThreshold = this.#fetchSafetyThreshold();
|
||||
}
|
||||
|
||||
#appendContext(contextTexts = []) {
|
||||
if (!contextTexts || !contextTexts.length) return "";
|
||||
return (
|
||||
"\nContext:\n" +
|
||||
contextTexts
|
||||
.map((text, i) => {
|
||||
return `[CONTEXT ${i}]:\n${text}\n[END CONTEXT ${i}]\n\n`;
|
||||
})
|
||||
.join("")
|
||||
);
|
||||
}
|
||||
|
||||
// BLOCK_NONE can be a special candidate for some fields
|
||||
// https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/configure-safety-attributes#how_to_remove_automated_response_blocking_for_select_safety_attributes
|
||||
// so if you are wondering why BLOCK_NONE still failed, the link above will explain why.
|
||||
#fetchSafetyThreshold() {
|
||||
const threshold =
|
||||
process.env.GEMINI_SAFETY_SETTING ?? "BLOCK_MEDIUM_AND_ABOVE";
|
||||
const safetyThresholds = [
|
||||
"BLOCK_NONE",
|
||||
"BLOCK_ONLY_HIGH",
|
||||
"BLOCK_MEDIUM_AND_ABOVE",
|
||||
"BLOCK_LOW_AND_ABOVE",
|
||||
];
|
||||
return safetyThresholds.includes(threshold)
|
||||
? threshold
|
||||
: "BLOCK_MEDIUM_AND_ABOVE";
|
||||
}
|
||||
|
||||
#safetySettings() {
|
||||
return [
|
||||
{
|
||||
category: "HARM_CATEGORY_HATE_SPEECH",
|
||||
threshold: this.safetyThreshold,
|
||||
},
|
||||
{
|
||||
category: "HARM_CATEGORY_SEXUALLY_EXPLICIT",
|
||||
threshold: this.safetyThreshold,
|
||||
},
|
||||
{ category: "HARM_CATEGORY_HARASSMENT", threshold: this.safetyThreshold },
|
||||
{
|
||||
category: "HARM_CATEGORY_DANGEROUS_CONTENT",
|
||||
threshold: this.safetyThreshold,
|
||||
},
|
||||
];
|
||||
}
|
||||
|
||||
streamingEnabled() {
|
||||
return "streamGetChatCompletion" in this;
|
||||
}
|
||||
|
||||
static promptWindowLimit(modelName) {
|
||||
return MODEL_MAP.vertex[modelName] ?? 30_720;
|
||||
}
|
||||
|
||||
promptWindowLimit() {
|
||||
return MODEL_MAP.vertex[this.model] ?? 30_720;
|
||||
}
|
||||
|
||||
isValidChatCompletionModel(modelName = "") {
|
||||
const validModels = [
|
||||
"gemini-1.5-flash",
|
||||
"gemini-1.5-pro",
|
||||
"gemini-1.5-flash-001",
|
||||
"gemini-1.5-pro-001",
|
||||
"gemini-1.5-flash-002",
|
||||
"gemini-1.5-pro-002",
|
||||
];
|
||||
return validModels.includes(modelName);
|
||||
}
|
||||
|
||||
/**
|
||||
* Generates appropriate content array for a message + attachments.
|
||||
* @param {{userPrompt:string, attachments: import("../../helpers").Attachment[]}}
|
||||
* @returns {string|object[]}
|
||||
*/
|
||||
#generateContent({ userPrompt, attachments = [] }) {
|
||||
if (!attachments.length) {
|
||||
return userPrompt;
|
||||
}
|
||||
|
||||
const content = [{ text: userPrompt }];
|
||||
for (let attachment of attachments) {
|
||||
content.push({
|
||||
inlineData: {
|
||||
data: attachment.contentString.split("base64,")[1],
|
||||
mimeType: attachment.mime,
|
||||
},
|
||||
});
|
||||
}
|
||||
return content.flat();
|
||||
}
|
||||
|
||||
constructPrompt({
|
||||
systemPrompt = "",
|
||||
contextTexts = [],
|
||||
chatHistory = [],
|
||||
userPrompt = "",
|
||||
attachments = [],
|
||||
}) {
|
||||
const prompt = {
|
||||
role: "system",
|
||||
content: `${systemPrompt}${this.#appendContext(contextTexts)}`,
|
||||
};
|
||||
return [
|
||||
prompt,
|
||||
{ role: "assistant", content: "Okay." },
|
||||
...chatHistory,
|
||||
{
|
||||
role: "USER_PROMPT",
|
||||
content: this.#generateContent({ userPrompt, attachments }),
|
||||
},
|
||||
];
|
||||
}
|
||||
|
||||
// This will take an OpenAi format message array and only pluck valid roles from it.
|
||||
formatMessages(messages = []) {
|
||||
// Gemini roles are either user || model.
|
||||
// and all "content" is relabeled to "parts"
|
||||
const allMessages = messages
|
||||
.map((message) => {
|
||||
if (message.role === "system")
|
||||
return { role: "user", parts: [{ text: message.content }] };
|
||||
if (message.role === "user")
|
||||
return { role: "user", parts: [{ text: message.content }] };
|
||||
if (message.role === "assistant")
|
||||
return { role: "model", parts: [{ text: message.content }] };
|
||||
return null;
|
||||
})
|
||||
.filter((msg) => !!msg);
|
||||
|
||||
// Specifically, Google cannot have the last sent message be from a user with no assistant reply
|
||||
// otherwise it will crash. So if the last item is from the user, it was not completed so pop it off
|
||||
// the history.
|
||||
if (
|
||||
allMessages.length > 0 &&
|
||||
allMessages[allMessages.length - 1].role === "user"
|
||||
)
|
||||
allMessages.pop();
|
||||
|
||||
// Validate that after every user message, there is a model message
|
||||
// sometimes when using gemini we try to compress messages in order to retain as
|
||||
// much context as possible but this may mess up the order of the messages that the gemini model expects
|
||||
// we do this check to work around the edge case where 2 user prompts may be next to each other, in the message array
|
||||
for (let i = 0; i < allMessages.length; i++) {
|
||||
if (
|
||||
allMessages[i].role === "user" &&
|
||||
i < allMessages.length - 1 &&
|
||||
allMessages[i + 1].role !== "model"
|
||||
) {
|
||||
allMessages.splice(i + 1, 0, {
|
||||
role: "model",
|
||||
parts: [{ text: "Okay." }],
|
||||
});
|
||||
}
|
||||
}
|
||||
|
||||
return allMessages;
|
||||
}
|
||||
|
||||
async getChatCompletion(messages = [], _opts = {}) {
|
||||
if (!this.isValidChatCompletionModel(this.model))
|
||||
throw new Error(
|
||||
`Gemini chat: ${this.model} is not valid for chat completion!`
|
||||
);
|
||||
|
||||
const prompt = messages.find(
|
||||
(chat) => chat.role === "USER_PROMPT"
|
||||
)?.content;
|
||||
const chatThread = this.vertex.startChat({
|
||||
history: this.formatMessages(messages),
|
||||
safetySettings: this.#safetySettings(),
|
||||
});
|
||||
const result = await chatThread.sendMessage(prompt);
|
||||
const response = result.response;
|
||||
const responseText = response.text();
|
||||
|
||||
if (!responseText) throw new Error("Gemini: No response could be parsed.");
|
||||
|
||||
return responseText;
|
||||
}
|
||||
|
||||
async streamGetChatCompletion(messages = [], _opts = {}) {
|
||||
if (!this.isValidChatCompletionModel(this.model))
|
||||
throw new Error(
|
||||
`Gemini chat: ${this.model} is not valid for chat completion!`
|
||||
);
|
||||
|
||||
const prompt = messages.find(
|
||||
(chat) => chat.role === "USER_PROMPT"
|
||||
)?.content;
|
||||
const chatThread = this.vertex.startChat({
|
||||
history: this.formatMessages(messages),
|
||||
safetySettings: this.#safetySettings(),
|
||||
});
|
||||
const responseStream = await chatThread.sendMessageStream(prompt);
|
||||
if (!responseStream.stream)
|
||||
throw new Error("Could not stream response stream from Gemini.");
|
||||
|
||||
return responseStream.stream;
|
||||
}
|
||||
|
||||
async compressMessages(promptArgs = {}, rawHistory = []) {
|
||||
const { messageArrayCompressor } = require("../../helpers/chat");
|
||||
const messageArray = this.constructPrompt(promptArgs);
|
||||
return await messageArrayCompressor(this, messageArray, rawHistory);
|
||||
}
|
||||
|
||||
async handleStream(response, stream, responseProps) {
|
||||
const { uuid = uuidv4(), sources = [] } = responseProps;
|
||||
|
||||
return new Promise(async (resolve) => {
|
||||
let fullText = "";
|
||||
|
||||
// Establish listener to early-abort a streaming response
|
||||
// in case things go sideways or the user does not like the response.
|
||||
// We preserve the generated text but continue as if chat was completed
|
||||
// to preserve previously generated content.
|
||||
const handleAbort = () => clientAbortedHandler(resolve, fullText);
|
||||
response.on("close", handleAbort);
|
||||
|
||||
for await (const chunk of stream) {
|
||||
try {
|
||||
for (const part of chunk.candidates[0].content.parts) {
|
||||
fullText += part.text;
|
||||
writeResponseChunk(response, {
|
||||
uuid,
|
||||
sources: [],
|
||||
type: "textResponseChunk",
|
||||
textResponse: part.text,
|
||||
close: false,
|
||||
error: false,
|
||||
});
|
||||
}
|
||||
} catch (e) {
|
||||
writeResponseChunk(response, {
|
||||
uuid,
|
||||
sources: [],
|
||||
type: "abort",
|
||||
textResponse: null,
|
||||
close: true,
|
||||
error: e.message,
|
||||
});
|
||||
resolve(e.message);
|
||||
return;
|
||||
}
|
||||
}
|
||||
|
||||
response.removeListener("close", handleAbort);
|
||||
resolve(fullText);
|
||||
});
|
||||
}
|
||||
|
||||
// Simple wrapper for dynamic embedder & normalize interface for all LLM implementations
|
||||
async embedTextInput(textInput) {
|
||||
return await this.embedder.embedTextInput(textInput);
|
||||
}
|
||||
async embedChunks(textChunks = []) {
|
||||
return await this.embedder.embedChunks(textChunks);
|
||||
}
|
||||
}
|
||||
|
||||
module.exports = {
|
||||
VertexLLM,
|
||||
};
|
@ -210,6 +210,8 @@ class AgentHandler {
|
||||
return process.env.KOBOLD_CPP_MODEL_PREF ?? null;
|
||||
case "gemini":
|
||||
return process.env.GEMINI_MODEL_PREF ?? "gemini-pro";
|
||||
case "vertex":
|
||||
return process.env.GEMINI_MODEL_PREF ?? "gemini-1.5-pro";
|
||||
case "localai":
|
||||
return process.env.LOCAL_AI_MODEL_PREF ?? null;
|
||||
case "openrouter":
|
||||
|
@ -108,6 +108,9 @@ function getLLMProvider({ provider = null, model = null } = {}) {
|
||||
case "gemini":
|
||||
const { GeminiLLM } = require("../AiProviders/gemini");
|
||||
return new GeminiLLM(embedder, model);
|
||||
case "vertex":
|
||||
const { VertexLLM } = require("../AiProviders/vertex");
|
||||
return new VertexLLM(embedder, model);
|
||||
case "lmstudio":
|
||||
const { LMStudioLLM } = require("../AiProviders/lmStudio");
|
||||
return new LMStudioLLM(embedder, model);
|
||||
@ -240,6 +243,9 @@ function getLLMProviderClass({ provider = null } = {}) {
|
||||
case "gemini":
|
||||
const { GeminiLLM } = require("../AiProviders/gemini");
|
||||
return GeminiLLM;
|
||||
case "vertex":
|
||||
const { VertexLLM } = require("../AiProviders/vertex");
|
||||
return VertexLLM;
|
||||
case "lmstudio":
|
||||
const { LMStudioLLM } = require("../AiProviders/lmStudio");
|
||||
return LMStudioLLM;
|
||||
|
@ -57,6 +57,19 @@ const KEY_MAPPING = {
|
||||
checks: [validGeminiSafetySetting],
|
||||
},
|
||||
|
||||
VertexLLMModelPref: {
|
||||
envKey: "VERTEX_LLM_MODEL_PREF",
|
||||
checks: [isNotEmpty, validVertexModel],
|
||||
},
|
||||
VertexProjectName: {
|
||||
envKey: "VERTEX_PROJECT_NAME",
|
||||
checks: [isNotEmpty],
|
||||
},
|
||||
VertexSafetySetting: {
|
||||
envKey: "VERTEX_SAFETY_SETTING",
|
||||
checks: [validGeminiSafetySetting],
|
||||
},
|
||||
|
||||
// LMStudio Settings
|
||||
LMStudioBasePath: {
|
||||
envKey: "LMSTUDIO_BASE_PATH",
|
||||
@ -645,6 +658,7 @@ function supportedLLM(input = "") {
|
||||
"azure",
|
||||
"anthropic",
|
||||
"gemini",
|
||||
"vertex",
|
||||
"lmstudio",
|
||||
"localai",
|
||||
"ollama",
|
||||
@ -692,6 +706,20 @@ function validGeminiModel(input = "") {
|
||||
: `Invalid Model type. Must be one of ${validModels.join(", ")}.`;
|
||||
}
|
||||
|
||||
function validVertexModel(input = "") {
|
||||
const validModels = [
|
||||
"gemini-1.5-flash",
|
||||
"gemini-1.5-pro",
|
||||
"gemini-1.5-flash-001",
|
||||
"gemini-1.5-pro-001",
|
||||
"gemini-1.5-flash-002",
|
||||
"gemini-1.5-pro-002",
|
||||
];
|
||||
return validModels.includes(input)
|
||||
? null
|
||||
: `Invalid Model type. Must be one of ${validModels.join(", ")}.`;
|
||||
}
|
||||
|
||||
function validGeminiSafetySetting(input = "") {
|
||||
const validModes = [
|
||||
"BLOCK_NONE",
|
||||
|
@ -874,6 +874,13 @@
|
||||
resolved "https://registry.yarnpkg.com/@fastify/busboy/-/busboy-2.1.1.tgz#b9da6a878a371829a0502c9b6c1c143ef6663f4d"
|
||||
integrity sha512-vBZP4NlzfOlerQTnba4aqZoMhE/a9HY7HRqoOPaETQcSQuWEIyZMHGfVu6w9wGtGK5fED5qRs2DteVCjOH60sA==
|
||||
|
||||
"@google-cloud/vertexai@^1.9.0":
|
||||
version "1.9.0"
|
||||
resolved "https://registry.yarnpkg.com/@google-cloud/vertexai/-/vertexai-1.9.0.tgz#30941faa920e1218231604285c56aa4ae172b358"
|
||||
integrity sha512-8brlcJwFXI4fPuBtsDNQqCdWZmz8gV9jeEKOU0vc5H2SjehCQpXK/NwuSEr916zbhlBHtg/sU37qQQdgvh5BRA==
|
||||
dependencies:
|
||||
google-auth-library "^9.1.0"
|
||||
|
||||
"@google/generative-ai@^0.7.1":
|
||||
version "0.7.1"
|
||||
resolved "https://registry.yarnpkg.com/@google/generative-ai/-/generative-ai-0.7.1.tgz#eb187c75080c0706245699dbc06816c830d8c6a7"
|
||||
@ -2588,7 +2595,7 @@ base-64@^0.1.0:
|
||||
resolved "https://registry.yarnpkg.com/base-64/-/base-64-0.1.0.tgz#780a99c84e7d600260361511c4877613bf24f6bb"
|
||||
integrity sha512-Y5gU45svrR5tI2Vt/X9GPd3L0HNIKzGu202EjxrXMpuc2V2CiKgemAbUUsqYmZJvPtCXoUKjNZwBJzsNScUbXA==
|
||||
|
||||
base64-js@^1.3.1, base64-js@^1.5.1:
|
||||
base64-js@^1.3.0, base64-js@^1.3.1, base64-js@^1.5.1:
|
||||
version "1.5.1"
|
||||
resolved "https://registry.yarnpkg.com/base64-js/-/base64-js-1.5.1.tgz#1b1b440160a5bf7ad40b650f095963481903930a"
|
||||
integrity sha512-AKpaYlHn8t4SVbOHCy+b5+KKgvR4vrsD8vbvrbiQJps7fKDTkjkDry6ji0rUJjC0kzbNePLwzxq8iypo41qeWA==
|
||||
@ -2606,6 +2613,11 @@ before-after-hook@^2.2.0:
|
||||
resolved "https://registry.yarnpkg.com/before-after-hook/-/before-after-hook-2.2.3.tgz#c51e809c81a4e354084422b9b26bad88249c517c"
|
||||
integrity sha512-NzUnlZexiaH/46WDhANlyR2bXRopNg4F/zuSA3OpZnllCUgRaOF2znDioDWrmbNVsuZk6l9pMquQB38cfBZwkQ==
|
||||
|
||||
bignumber.js@^9.0.0:
|
||||
version "9.1.2"
|
||||
resolved "https://registry.yarnpkg.com/bignumber.js/-/bignumber.js-9.1.2.tgz#b7c4242259c008903b13707983b5f4bbd31eda0c"
|
||||
integrity sha512-2/mKyZH9K85bzOEfhXDBFZTGd1CTs+5IHpeFQo9luiBG7hghdC851Pj2WAhb6E3R6b9tZj/XKhbg4fum+Kepug==
|
||||
|
||||
binary-extensions@^2.0.0, binary-extensions@^2.2.0:
|
||||
version "2.3.0"
|
||||
resolved "https://registry.yarnpkg.com/binary-extensions/-/binary-extensions-2.3.0.tgz#f6e14a97858d327252200242d4ccfe522c445522"
|
||||
@ -3339,7 +3351,7 @@ eastasianwidth@^0.2.0:
|
||||
resolved "https://registry.yarnpkg.com/eastasianwidth/-/eastasianwidth-0.2.0.tgz#696ce2ec0aa0e6ea93a397ffcf24aa7840c827cb"
|
||||
integrity sha512-I88TYZWc9XiYHRQ4/3c5rjjfgkjhLyW2luGIheGERbNQ6OY7yTybanSpDXZa8y7VUP9YmDcYa+eyq4ca7iLqWA==
|
||||
|
||||
ecdsa-sig-formatter@1.0.11:
|
||||
ecdsa-sig-formatter@1.0.11, ecdsa-sig-formatter@^1.0.11:
|
||||
version "1.0.11"
|
||||
resolved "https://registry.yarnpkg.com/ecdsa-sig-formatter/-/ecdsa-sig-formatter-1.0.11.tgz#ae0f0fa2d85045ef14a817daa3ce9acd0489e5bf"
|
||||
integrity sha512-nagl3RYrbNv6kQkeJIpt6NJZy8twLB/2vtz6yN9Z4vRKHN4/QZJIEbqohALSgwKdnksuY3k5Addp5lg8sVoVcQ==
|
||||
@ -3784,6 +3796,11 @@ express@^4.18.2:
|
||||
utils-merge "1.0.1"
|
||||
vary "~1.1.2"
|
||||
|
||||
extend@^3.0.2:
|
||||
version "3.0.2"
|
||||
resolved "https://registry.yarnpkg.com/extend/-/extend-3.0.2.tgz#f8b1136b4071fbd8eb140aff858b1019ec2915fa"
|
||||
integrity sha512-fjquC59cD7CyW6urNXK0FBufkZcoiGG80wTuPujX590cB5Ttln20E2UB4S/WARVqhXffZl2LNgS+gQdPIIim/g==
|
||||
|
||||
external-editor@^3.1.0:
|
||||
version "3.1.0"
|
||||
resolved "https://registry.yarnpkg.com/external-editor/-/external-editor-3.1.0.tgz#cb03f740befae03ea4d283caed2741a83f335495"
|
||||
@ -4077,6 +4094,25 @@ gauge@^4.0.3:
|
||||
strip-ansi "^6.0.1"
|
||||
wide-align "^1.1.5"
|
||||
|
||||
gaxios@^6.0.0, gaxios@^6.1.1:
|
||||
version "6.7.1"
|
||||
resolved "https://registry.yarnpkg.com/gaxios/-/gaxios-6.7.1.tgz#ebd9f7093ede3ba502685e73390248bb5b7f71fb"
|
||||
integrity sha512-LDODD4TMYx7XXdpwxAVRAIAuB0bzv0s+ywFonY46k126qzQHT9ygyoa9tncmOiQmmDrik65UYsEkv3lbfqQ3yQ==
|
||||
dependencies:
|
||||
extend "^3.0.2"
|
||||
https-proxy-agent "^7.0.1"
|
||||
is-stream "^2.0.0"
|
||||
node-fetch "^2.6.9"
|
||||
uuid "^9.0.1"
|
||||
|
||||
gcp-metadata@^6.1.0:
|
||||
version "6.1.0"
|
||||
resolved "https://registry.yarnpkg.com/gcp-metadata/-/gcp-metadata-6.1.0.tgz#9b0dd2b2445258e7597f2024332d20611cbd6b8c"
|
||||
integrity sha512-Jh/AIwwgaxan+7ZUUmRLCjtchyDiqh4KjBJ5tW3plBZb5iL/BPcso8A5DlzeD9qlw0duCamnNdpFjxwaT0KyKg==
|
||||
dependencies:
|
||||
gaxios "^6.0.0"
|
||||
json-bigint "^1.0.0"
|
||||
|
||||
generate-function@^2.3.1:
|
||||
version "2.3.1"
|
||||
resolved "https://registry.yarnpkg.com/generate-function/-/generate-function-2.3.1.tgz#f069617690c10c868e73b8465746764f97c3479f"
|
||||
@ -4165,6 +4201,18 @@ globalthis@^1.0.3:
|
||||
define-properties "^1.2.1"
|
||||
gopd "^1.0.1"
|
||||
|
||||
google-auth-library@^9.1.0:
|
||||
version "9.14.2"
|
||||
resolved "https://registry.yarnpkg.com/google-auth-library/-/google-auth-library-9.14.2.tgz#92a53ba32b3a9ff9ced8ed34129edb5a7fa7fb52"
|
||||
integrity sha512-R+FRIfk1GBo3RdlRYWPdwk8nmtVUOn6+BkDomAC46KoU8kzXzE1HLmOasSCbWUByMMAGkknVF0G5kQ69Vj7dlA==
|
||||
dependencies:
|
||||
base64-js "^1.3.0"
|
||||
ecdsa-sig-formatter "^1.0.11"
|
||||
gaxios "^6.1.1"
|
||||
gcp-metadata "^6.1.0"
|
||||
gtoken "^7.0.0"
|
||||
jws "^4.0.0"
|
||||
|
||||
gopd@^1.0.1:
|
||||
version "1.0.1"
|
||||
resolved "https://registry.yarnpkg.com/gopd/-/gopd-1.0.1.tgz#29ff76de69dac7489b7c0918a5788e56477c332c"
|
||||
@ -4197,6 +4245,14 @@ graphql@^16.7.1:
|
||||
resolved "https://registry.yarnpkg.com/graphql/-/graphql-16.9.0.tgz#1c310e63f16a49ce1fbb230bd0a000e99f6f115f"
|
||||
integrity sha512-GGTKBX4SD7Wdb8mqeDLni2oaRGYQWjWHGKPQ24ZMnUtKfcsVoiv4uX8+LJr1K6U5VW2Lu1BwJnj7uiori0YtRw==
|
||||
|
||||
gtoken@^7.0.0:
|
||||
version "7.1.0"
|
||||
resolved "https://registry.yarnpkg.com/gtoken/-/gtoken-7.1.0.tgz#d61b4ebd10132222817f7222b1e6064bd463fc26"
|
||||
integrity sha512-pCcEwRi+TKpMlxAQObHDQ56KawURgyAf6jtIY046fJ5tIv3zDe/LEIubckAO8fj6JnAxLdmWkUfNyulQ2iKdEw==
|
||||
dependencies:
|
||||
gaxios "^6.0.0"
|
||||
jws "^4.0.0"
|
||||
|
||||
guid-typescript@^1.0.9:
|
||||
version "1.0.9"
|
||||
resolved "https://registry.yarnpkg.com/guid-typescript/-/guid-typescript-1.0.9.tgz#e35f77003535b0297ea08548f5ace6adb1480ddc"
|
||||
@ -4326,6 +4382,14 @@ https-proxy-agent@^7.0.0:
|
||||
agent-base "^7.0.2"
|
||||
debug "4"
|
||||
|
||||
https-proxy-agent@^7.0.1:
|
||||
version "7.0.5"
|
||||
resolved "https://registry.yarnpkg.com/https-proxy-agent/-/https-proxy-agent-7.0.5.tgz#9e8b5013873299e11fab6fd548405da2d6c602b2"
|
||||
integrity sha512-1e4Wqeblerz+tMKPIq2EMGiiWW1dIjZOksyHWSUm1rmuvw/how9hBHZ38lAGj5ID4Ik6EdkOw7NmWPy6LAwalw==
|
||||
dependencies:
|
||||
agent-base "^7.0.2"
|
||||
debug "4"
|
||||
|
||||
human-interval@^2.0.1:
|
||||
version "2.0.1"
|
||||
resolved "https://registry.yarnpkg.com/human-interval/-/human-interval-2.0.1.tgz#655baf606c7067bb26042dcae14ec777b099af15"
|
||||
@ -4787,6 +4851,13 @@ jsbi@^4.3.0:
|
||||
resolved "https://registry.yarnpkg.com/jsbi/-/jsbi-4.3.0.tgz#b54ee074fb6fcbc00619559305c8f7e912b04741"
|
||||
integrity sha512-SnZNcinB4RIcnEyZqFPdGPVgrg2AcnykiBy0sHVJQKHYeaLUvi3Exj+iaPpLnFVkDPZIV4U0yvgC9/R4uEAZ9g==
|
||||
|
||||
json-bigint@^1.0.0:
|
||||
version "1.0.0"
|
||||
resolved "https://registry.yarnpkg.com/json-bigint/-/json-bigint-1.0.0.tgz#ae547823ac0cad8398667f8cd9ef4730f5b01ff1"
|
||||
integrity sha512-SiPv/8VpZuWbvLSMtTDU8hEfrZWg/mH/nV/b4o0CYbSxu1UIQPLdwKOCIyLQX+VIPO5vrLX3i8qtqFyhdPSUSQ==
|
||||
dependencies:
|
||||
bignumber.js "^9.0.0"
|
||||
|
||||
json-bignum@^0.0.3:
|
||||
version "0.0.3"
|
||||
resolved "https://registry.yarnpkg.com/json-bignum/-/json-bignum-0.0.3.tgz#41163b50436c773d82424dbc20ed70db7604b8d7"
|
||||
@ -5452,7 +5523,7 @@ node-domexception@1.0.0:
|
||||
resolved "https://registry.yarnpkg.com/node-domexception/-/node-domexception-1.0.0.tgz#6888db46a1f71c0b76b3f7555016b63fe64766e5"
|
||||
integrity sha512-/jKZoMpw0F8GRwl4/eLROPA3cfcXtLApP0QzLmUT/HuPCZWyB7IY9ZrMeKw2O/nFIqPQB3PVM9aYm0F312AXDQ==
|
||||
|
||||
node-fetch@2.7.0, node-fetch@^2.6.1, node-fetch@^2.6.12, node-fetch@^2.6.7:
|
||||
node-fetch@2.7.0, node-fetch@^2.6.1, node-fetch@^2.6.12, node-fetch@^2.6.7, node-fetch@^2.6.9:
|
||||
version "2.7.0"
|
||||
resolved "https://registry.yarnpkg.com/node-fetch/-/node-fetch-2.7.0.tgz#d0f0fa6e3e2dc1d27efcd8ad99d550bda94d187d"
|
||||
integrity sha512-c4FRfUm/dbcWZ7U+1Wq0AwCyFL+3nt2bEw05wfxSz+DWpWsitgmSgYmy2dQdWyKC1694ELPqMs/YzUSNozLt8A==
|
||||
|
Loading…
Reference in New Issue
Block a user