@chenchaolong/plugin-vllm 0.0.4

package/README.md

@@ -0,0 +1,68 @@
+# Xpert Plugin: vLLM
+
+## Overview
+
+`@chenchaolong/plugin-vllm` provides a model adapter for connecting vLLM inference services to the [XpertAI](https://github.com/xpert-ai/xpert) platform. The plugin communicates with vLLM clusters via an OpenAI-compatible API, enabling agents to invoke conversational models, embedding models, vision-enhanced models, and reranking models within a unified XpertAI agentic workflow.
+
+## Core Features
+
+- Provides the `VLLMPlugin` NestJS module, which automatically registers model providers, lifecycle logging, and configuration validation logic.
+- Wraps vLLM's conversational/inference capabilities as XpertAI's `LargeLanguageModel` via `VLLMLargeLanguageModel`, supporting function calling, streaming output, and agent token statistics.
+- Exposes `VLLMTextEmbeddingModel`, reusing LangChain's `OpenAIEmbeddings` to generate vector representations for knowledge base retrieval.
+- Integrates `VLLMRerankModel`, leveraging the OpenAI-compatible rerank API to improve retrieval result ranking.
+- Supports declaring capabilities such as vision, function calling, and streaming mode in plugin metadata, allowing flexible configuration of different vLLM deployments in the console.
+
+## Installation
+
+```bash
+npm install @chenchaolong/plugin-vllm
+```
+
+> **Peer Dependencies**: The host project must also provide libraries such as `@xpert-ai/plugin-sdk`, `@nestjs/common`, `@metad/contracts`, `@langchain/openai`, `lodash-es`, `chalk`, and `zod`. Please refer to `package.json` for version requirements.
+
+## Enabling in XpertAI
+
+1. Add the plugin package to your system dependencies and ensure it is resolvable by Node.js.
+2. Before starting the service, declare the plugin in your environment variables:
+   ```bash
+   PLUGINS=@chenchaolong/plugin-vllm
+   ```
+3. Add a new model provider in the XpertAI admin interface or configuration file, and select `vllm`.
+
+## Credentials & Model Configuration
+
+The form fields defined in `vllm.yaml` cover common deployment scenarios:
+
+| Field | Description |
+| --- | --- |
+| `api_key` | vLLM service access token (leave blank if the service does not require authentication). |
+| `endpoint_url` | Required. The base URL of the vLLM OpenAI-compatible API, e.g., `https://vllm.example.com/v1`. |
+| `endpoint_model_name` | Specify explicitly if the model name on the server differs from the logical model name in XpertAI. |
+| `mode` | Choose between `chat` or `completion` inference modes. |
+| `context_size` / `max_tokens_to_sample` | Control the context window and generation length. |
+| `agent_though_support`, `function_calling_type`, `stream_function_calling`, `vision_support` | Indicate whether the model supports agent thought exposure, function/tool calling, streaming function calling, and multimodal input, to inform UI capability hints. |
+| `stream_mode_delimiter` | Customize the paragraph delimiter for streaming output. |
+
+After saving the configuration, the plugin will call the `validateCredentials` method in the background, making a minimal request to the vLLM service to ensure the credentials are valid.
+
+## Model Capabilities
+
+- **Conversational Models**: Uses `ChatOAICompatReasoningModel` to proxy the vLLM OpenAI API, supporting message history, function calling, and streaming output.
+- **Embedding Models**: Relies on LangChain's `OpenAIEmbeddings` for knowledge base vectorization and retrieval-augmented generation.
+- **Reranking Models**: Wraps `OpenAICompatibleReranker` to semantically rerank recall results.
+- **Vision Models**: If the vLLM inference service supports multimodal (text+image) input, enable `vision_support` in the configuration to declare multimodal capabilities to the frontend.
+
+## Development & Debugging
+
+From the repository root, enter the `xpertai/` directory and use Nx commands to build and test:
+
+```bash
+npx nx build @chenchaolong/plugin-vllm
+npx nx test @chenchaolong/plugin-vllm
+```
+
+Build artifacts are output to `dist/` by default. Jest configuration is in `jest.config.ts` for writing and running unit tests.
+
+## License
+
+This project follows the [AGPL-3.0 License](../../../LICENSE) found at the root of the repository.

package/dist/_assets/vllm-color.svg

@@ -0,0 +1 @@
+<svg height="1em" style="flex:none;line-height:1" viewBox="0 0 24 24" width="1em" xmlns="http://www.w3.org/2000/svg"><title>vLLM</title><path d="M0 4.973h9.324V23L0 4.973z" fill="#FDB515"></path><path d="M13.986 4.351L22.378 0l-6.216 23H9.324l4.662-18.649z" fill="#30A2FF"></path></svg>

package/dist/i18n.d.ts

@@ -0,0 +1,4 @@
+import { TOptions } from 'i18next';
+export declare function translate(key: string, options?: TOptions): string;
+export declare function initI18n(pluginDir: string): void;
+//# sourceMappingURL=i18n.d.ts.map

package/dist/i18n.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"i18n.d.ts","sourceRoot":"","sources":["../src/i18n.ts"],"names":[],"mappings":"AACA,OAAO,EAAwB,QAAQ,EAAE,MAAM,SAAS,CAAC;AAGzD,wBAAgB,SAAS,CAAC,GAAG,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,QAAQ,UAIxD;AAED,wBAAgB,QAAQ,CAAC,SAAS,EAAE,MAAM,GAAG,IAAI,CAIhD"}

package/dist/i18n.js

@@ -0,0 +1,12 @@
+import { createI18nInstance, RequestContext } from '@xpert-ai/plugin-sdk';
+let i18nObject = null;
+export function translate(key, options) {
+    options = options || {};
+    options.lng = options.lng || RequestContext.getLanguageCode();
+    return i18nObject?.t(key, options) || key;
+}
+export function initI18n(pluginDir) {
+    createI18nInstance(pluginDir).then((i18n) => {
+        i18nObject = i18n;
+    });
+}

package/dist/index.d.ts

@@ -0,0 +1,6 @@
+import type { XpertPlugin } from '@xpert-ai/plugin-sdk';
+import { z } from 'zod';
+declare const ConfigSchema: z.ZodObject<{}, "strip", z.ZodTypeAny, {}, {}>;
+declare const plugin: XpertPlugin<z.infer<typeof ConfigSchema>>;
+export default plugin;
+//# sourceMappingURL=index.d.ts.map

package/dist/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAIxD,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAaxB,QAAA,MAAM,YAAY,gDAChB,CAAC;AAEH,QAAA,MAAM,MAAM,EAAE,WAAW,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,YAAY,CAAC,CA4BrD,CAAC;AAEF,eAAe,MAAM,CAAC"}

package/dist/index.js

@@ -0,0 +1,41 @@
+import { readFileSync } from 'fs';
+import { fileURLToPath } from 'url';
+import { dirname, join } from 'path';
+import { z } from 'zod';
+import { SvgIcon } from './types.js';
+import { VLLMPlugin } from './vllm.js';
+import { initI18n } from './i18n.js';
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = dirname(__filename);
+const packageJson = JSON.parse(readFileSync(join(__dirname, '../package.json'), 'utf8'));
+const ConfigSchema = z.object({});
+const plugin = {
+    meta: {
+        name: packageJson.name,
+        version: packageJson.version,
+        category: 'model',
+        icon: {
+            type: 'svg',
+            value: SvgIcon
+        },
+        displayName: 'vLLM',
+        description: 'Provide connector for vLLM models',
+        keywords: ['vLLM', 'model'],
+        author: 'XpertAI',
+    },
+    config: {
+        schema: ConfigSchema,
+    },
+    register(ctx) {
+        ctx.logger.log('register VLLM plugin');
+        initI18n(join(__dirname, '../src'));
+        return { module: VLLMPlugin, global: true };
+    },
+    async onStart(ctx) {
+        ctx.logger.log('VLLM plugin started');
+    },
+    async onStop(ctx) {
+        ctx.logger.log('VLLM plugin stopped');
+    },
+};
+export default plugin;

package/dist/llm/llm.d.ts

@@ -0,0 +1,11 @@
+import { ICopilotModel } from '@metad/contracts';
+import { ChatOAICompatReasoningModel, LargeLanguageModel, TChatModelOptions } from '@xpert-ai/plugin-sdk';
+import { VLLMProviderStrategy } from '../provider.strategy.js';
+import { VLLMModelCredentials } from '../types.js';
+export declare class VLLMLargeLanguageModel extends LargeLanguageModel {
+    #private;
+    constructor(modelProvider: VLLMProviderStrategy);
+    validateCredentials(model: string, credentials: VLLMModelCredentials): Promise<void>;
+    getChatModel(copilotModel: ICopilotModel, options?: TChatModelOptions): ChatOAICompatReasoningModel;
+}
+//# sourceMappingURL=llm.d.ts.map

package/dist/llm/llm.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"llm.d.ts","sourceRoot":"","sources":["../../src/llm/llm.ts"],"names":[],"mappings":"AACA,OAAO,EAAmB,aAAa,EAAE,MAAM,kBAAkB,CAAA;AAEjE,OAAO,EACL,2BAA2B,EAG3B,kBAAkB,EAClB,iBAAiB,EAClB,MAAM,sBAAsB,CAAA;AAE7B,OAAO,EAAE,oBAAoB,EAAE,MAAM,yBAAyB,CAAA;AAC9D,OAAO,EAAsB,oBAAoB,EAAE,MAAM,aAAa,CAAA;AAGtE,qBACa,sBAAuB,SAAQ,kBAAkB;;gBAGhD,aAAa,EAAE,oBAAoB;IAIzC,mBAAmB,CAAC,KAAK,EAAE,MAAM,EAAE,WAAW,EAAE,oBAAoB,GAAG,OAAO,CAAC,IAAI,CAAC;IAkBjF,YAAY,CAAC,YAAY,EAAE,aAAa,EAAE,OAAO,CAAC,EAAE,iBAAiB;CA4B/E"}

package/dist/llm/llm.js

@@ -0,0 +1,62 @@
+var _a;
+var VLLMLargeLanguageModel_1;
+import { __decorate, __metadata } from "tslib";
+import { ChatOpenAI } from '@langchain/openai';
+import { AiModelTypeEnum } from '@metad/contracts';
+import { Injectable, Logger } from '@nestjs/common';
+import { ChatOAICompatReasoningModel, CredentialsValidateFailedError, getErrorMessage, LargeLanguageModel } from '@xpert-ai/plugin-sdk';
+import { isNil, omitBy } from 'lodash-es';
+import { VLLMProviderStrategy } from '../provider.strategy.js';
+import { toCredentialKwargs } from '../types.js';
+import { translate } from '../i18n.js';
+let VLLMLargeLanguageModel = VLLMLargeLanguageModel_1 = class VLLMLargeLanguageModel extends LargeLanguageModel {
+    #logger = new Logger(VLLMLargeLanguageModel_1.name);
+    constructor(modelProvider) {
+        super(modelProvider, AiModelTypeEnum.LLM);
+    }
+    async validateCredentials(model, credentials) {
+        try {
+            const chatModel = new ChatOpenAI({
+                ...toCredentialKwargs(credentials, model),
+                temperature: 0,
+                maxTokens: 5
+            });
+            await chatModel.invoke([
+                {
+                    role: 'human',
+                    content: `Hi`
+                }
+            ]);
+        }
+        catch (err) {
+            throw new CredentialsValidateFailedError(getErrorMessage(err));
+        }
+    }
+    getChatModel(copilotModel, options) {
+        const { handleLLMTokens, modelProperties } = options ?? {};
+        const { copilot } = copilotModel;
+        if (!modelProperties) {
+            throw new Error(translate('Error.ModelCredentialsMissing', { model: copilotModel.model }));
+        }
+        const params = toCredentialKwargs(modelProperties, copilotModel.model);
+        const fields = omitBy({
+            ...params,
+            streaming: copilotModel.options?.['streaming'] ?? true,
+            // include token usage in the stream. this will include an additional chunk at the end of the stream with the token usage.
+            streamUsage: true
+        }, isNil);
+        return new ChatOAICompatReasoningModel({
+            ...fields,
+            verbose: options?.verbose,
+            callbacks: [
+                ...this.createHandleUsageCallbacks(copilot, params.model, modelProperties, handleLLMTokens),
+                this.createHandleLLMErrorCallbacks(fields, this.#logger)
+            ]
+        });
+    }
+};
+VLLMLargeLanguageModel = VLLMLargeLanguageModel_1 = __decorate([
+    Injectable(),
+    __metadata("design:paramtypes", [VLLMProviderStrategy])
+], VLLMLargeLanguageModel);
+export { VLLMLargeLanguageModel };

package/dist/provider.strategy.d.ts

@@ -0,0 +1,10 @@
+import { Logger } from '@nestjs/common';
+import { ModelProvider } from '@xpert-ai/plugin-sdk';
+import { VLLMModelCredentials } from './types.js';
+export declare class VLLMProviderStrategy extends ModelProvider {
+    logger: Logger;
+    validateProviderCredentials(credentials: VLLMModelCredentials): Promise<void>;
+    getBaseUrl(credentials: VLLMModelCredentials): string;
+    getAuthorization(credentials: VLLMModelCredentials): string;
+}
+//# sourceMappingURL=provider.strategy.d.ts.map

package/dist/provider.strategy.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"provider.strategy.d.ts","sourceRoot":"","sources":["../src/provider.strategy.ts"],"names":[],"mappings":"AAAA,OAAO,EAAc,MAAM,EAAE,MAAM,gBAAgB,CAAA;AACnD,OAAO,EAA2B,aAAa,EAAE,MAAM,sBAAsB,CAAA;AAC7E,OAAO,EAAQ,oBAAoB,EAAE,MAAM,YAAY,CAAA;AAEvD,qBAEa,oBAAqB,SAAQ,aAAa;IAC5C,MAAM,SAAwC;IAExC,2BAA2B,CAAC,WAAW,EAAE,oBAAoB,GAAG,OAAO,CAAC,IAAI,CAAC;IAI5F,UAAU,CAAC,WAAW,EAAE,oBAAoB,GAAG,MAAM;IAIrD,gBAAgB,CAAC,WAAW,EAAE,oBAAoB,GAAG,MAAM;CAG5D"}

package/dist/provider.strategy.js

@@ -0,0 +1,25 @@
+var VLLMProviderStrategy_1;
+import { __decorate } from "tslib";
+import { Injectable, Logger } from '@nestjs/common';
+import { AIModelProviderStrategy, ModelProvider } from '@xpert-ai/plugin-sdk';
+import { VLLM } from './types.js';
+let VLLMProviderStrategy = VLLMProviderStrategy_1 = class VLLMProviderStrategy extends ModelProvider {
+    constructor() {
+        super(...arguments);
+        this.logger = new Logger(VLLMProviderStrategy_1.name);
+    }
+    async validateProviderCredentials(credentials) {
+        // No validation needed for vLLM
+    }
+    getBaseUrl(credentials) {
+        return credentials.endpoint_url;
+    }
+    getAuthorization(credentials) {
+        return `Bearer ${credentials.api_key}`;
+    }
+};
+VLLMProviderStrategy = VLLMProviderStrategy_1 = __decorate([
+    Injectable(),
+    AIModelProviderStrategy(VLLM)
+], VLLMProviderStrategy);
+export { VLLMProviderStrategy };

package/dist/rerank/rerank.d.ts

@@ -0,0 +1,9 @@
+import { ICopilotModel } from '@metad/contracts';
+import { IRerank, RerankModel, TChatModelOptions } from '@xpert-ai/plugin-sdk';
+import { VLLMProviderStrategy } from '../provider.strategy.js';
+export declare class VLLMRerankModel extends RerankModel {
+    constructor(modelProvider: VLLMProviderStrategy);
+    validateCredentials(model: string, credentials: Record<string, any>): Promise<void>;
+    getReranker(copilotModel: ICopilotModel, options?: TChatModelOptions): Promise<IRerank>;
+}
+//# sourceMappingURL=rerank.d.ts.map

package/dist/rerank/rerank.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"rerank.d.ts","sourceRoot":"","sources":["../../src/rerank/rerank.ts"],"names":[],"mappings":"AAAA,OAAO,EAAmB,aAAa,EAAE,MAAM,kBAAkB,CAAA;AAEjE,OAAO,EAEL,OAAO,EAEP,WAAW,EACX,iBAAiB,EAElB,MAAM,sBAAsB,CAAA;AAC7B,OAAO,EAAE,oBAAoB,EAAE,MAAM,yBAAyB,CAAA;AAG9D,qBACa,eAAgB,SAAQ,WAAW;gBAClC,aAAa,EAAE,oBAAoB;IAIhC,mBAAmB,CAAC,KAAK,EAAE,MAAM,EAAE,WAAW,EAAE,MAAM,CAAC,MAAM,EAAE,GAAG,CAAC;IAcnE,WAAW,CAAC,YAAY,EAAE,aAAa,EAAE,OAAO,CAAC,EAAE,iBAAiB,GAAG,OAAO,CAAC,OAAO,CAAC;CAcvG"}

package/dist/rerank/rerank.js

@@ -0,0 +1,41 @@
+import { __decorate, __metadata } from "tslib";
+import { AiModelTypeEnum } from '@metad/contracts';
+import { Injectable } from '@nestjs/common';
+import { getErrorMessage, OpenAICompatibleReranker, RerankModel } from '@xpert-ai/plugin-sdk';
+import { VLLMProviderStrategy } from '../provider.strategy.js';
+import { translate } from '../i18n.js';
+let VLLMRerankModel = class VLLMRerankModel extends RerankModel {
+    constructor(modelProvider) {
+        super(modelProvider, AiModelTypeEnum.RERANK);
+    }
+    async validateCredentials(model, credentials) {
+        const _credentials = credentials ?? {};
+        const reranker = new OpenAICompatibleReranker({
+            endpointUrl: _credentials.endpoint_url,
+            apiKey: _credentials.api_key,
+            endpointModelName: _credentials.endpoint_model_name
+        });
+        try {
+            await reranker.rerank([], 'test', { model });
+        }
+        catch (error) {
+            throw new Error(`Reranker credentials validation failed: ${getErrorMessage(error)}`);
+        }
+    }
+    async getReranker(copilotModel, options) {
+        const credentials = options?.modelProperties;
+        if (!credentials) {
+            throw new Error(translate('Error.ModelCredentialsMissing', { model: copilotModel.model }));
+        }
+        return new OpenAICompatibleReranker({
+            endpointUrl: credentials.endpoint_url,
+            apiKey: credentials.api_key,
+            endpointModelName: credentials.endpoint_model_name || copilotModel.model
+        });
+    }
+};
+VLLMRerankModel = __decorate([
+    Injectable(),
+    __metadata("design:paramtypes", [VLLMProviderStrategy])
+], VLLMRerankModel);
+export { VLLMRerankModel };

package/dist/text-embedding/text-embedding.d.ts

@@ -0,0 +1,11 @@
+import { OpenAIEmbeddings } from '@langchain/openai';
+import { ICopilotModel } from '@metad/contracts';
+import { VLLMModelCredentials } from '../types.js';
+import { TChatModelOptions, TextEmbeddingModelManager } from '@xpert-ai/plugin-sdk';
+import { VLLMProviderStrategy } from '../provider.strategy.js';
+export declare class VLLMTextEmbeddingModel extends TextEmbeddingModelManager {
+    constructor(modelProvider: VLLMProviderStrategy);
+    getEmbeddingInstance(copilotModel: ICopilotModel, options?: TChatModelOptions): OpenAIEmbeddings;
+    validateCredentials(model: string, credentials: VLLMModelCredentials): Promise<void>;
+}
+//# sourceMappingURL=text-embedding.d.ts.map

package/dist/text-embedding/text-embedding.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"text-embedding.d.ts","sourceRoot":"","sources":["../../src/text-embedding/text-embedding.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,gBAAgB,EAAE,MAAM,mBAAmB,CAAA;AACpD,OAAO,EAAmB,aAAa,EAAE,MAAM,kBAAkB,CAAA;AAEjE,OAAO,EAAsB,oBAAoB,EAAE,MAAM,aAAa,CAAA;AACtE,OAAO,EAAmD,iBAAiB,EAAE,yBAAyB,EAAE,MAAM,sBAAsB,CAAA;AACpI,OAAO,EAAE,oBAAoB,EAAE,MAAM,yBAAyB,CAAA;AAG9D,qBACa,sBAAuB,SAAQ,yBAAyB;gBACxD,aAAa,EAAE,oBAAoB;IAI/C,oBAAoB,CAAC,YAAY,EAAE,aAAa,EAAE,OAAO,CAAC,EAAE,iBAAiB,GAAG,gBAAgB;IAqB1F,mBAAmB,CAAC,KAAK,EAAE,MAAM,EAAE,WAAW,EAAE,oBAAoB,GAAG,OAAO,CAAC,IAAI,CAAC;CAe1F"}

package/dist/text-embedding/text-embedding.js

@@ -0,0 +1,48 @@
+import { __decorate, __metadata } from "tslib";
+import { OpenAIEmbeddings } from '@langchain/openai';
+import { AiModelTypeEnum } from '@metad/contracts';
+import { Injectable } from '@nestjs/common';
+import { toCredentialKwargs } from '../types.js';
+import { CredentialsValidateFailedError, getErrorMessage, TextEmbeddingModelManager } from '@xpert-ai/plugin-sdk';
+import { VLLMProviderStrategy } from '../provider.strategy.js';
+import { translate } from '../i18n.js';
+let VLLMTextEmbeddingModel = class VLLMTextEmbeddingModel extends TextEmbeddingModelManager {
+    constructor(modelProvider) {
+        super(modelProvider, AiModelTypeEnum.TEXT_EMBEDDING);
+    }
+    getEmbeddingInstance(copilotModel, options) {
+        const { copilot } = copilotModel;
+        const { modelProvider } = copilot;
+        if (!options?.modelProperties) {
+            throw new Error(translate('Error.ModelCredentialsMissing', { model: copilotModel.model }));
+        }
+        const params = toCredentialKwargs({
+            ...(modelProvider.credentials ?? {}),
+            ...options.modelProperties,
+        }, copilotModel.model || copilotModel.copilot.copilotModel?.model);
+        return new OpenAIEmbeddings({
+            ...params,
+            // batchSize: 512, // Default value if omitted is 512. Max is 2048
+        });
+    }
+    async validateCredentials(model, credentials) {
+        try {
+            // transform credentials to kwargs for model instance
+            const params = toCredentialKwargs(credentials, model);
+            const embeddings = new OpenAIEmbeddings({
+                ...params,
+                // batchSize: 512, // Default value if omitted is 512. Max is 2048
+            });
+            // call embedding model
+            await embeddings.embedQuery('ping');
+        }
+        catch (ex) {
+            throw new CredentialsValidateFailedError(getErrorMessage(ex));
+        }
+    }
+};
+VLLMTextEmbeddingModel = __decorate([
+    Injectable(),
+    __metadata("design:paramtypes", [VLLMProviderStrategy])
+], VLLMTextEmbeddingModel);
+export { VLLMTextEmbeddingModel };

package/dist/types.d.ts

@@ -0,0 +1,41 @@
+import { ClientOptions } from "@langchain/openai";
+export declare const VLLM = "vllm";
+export declare const SvgIcon = "<svg height=\"1em\" style=\"flex:none;line-height:1\" viewBox=\"0 0 24 24\" width=\"1em\" xmlns=\"http://www.w3.org/2000/svg\"><title>vLLM</title><path d=\"M0 4.973h9.324V23L0 4.973z\" fill=\"#FDB515\"></path><path d=\"M13.986 4.351L22.378 0l-6.216 23H9.324l4.662-18.649z\" fill=\"#30A2FF\"></path></svg>";
+export type VLLMModelCredentials = {
+    api_key: string;
+    endpoint_url?: string;
+    endpoint_model_name?: string;
+    mode?: 'completion' | 'chat';
+    context_size?: number;
+    max_tokens_to_sample?: number;
+    agent_though_support?: 'supported' | 'unsupported';
+    function_calling_type?: 'function_call' | 'tool_call' | 'no_call';
+    stream_function_calling?: 'supported' | 'unsupported';
+    vision_support?: 'supported' | 'unsupported';
+    stream_mode_delimiter?: string;
+    thinking?: boolean;
+};
+export declare function toCredentialKwargs(credentials: VLLMModelCredentials, model: string): {
+    modelKwargs: {};
+    configuration: ClientOptions;
+    temperature: number;
+    maxTokens?: number;
+    maxCompletionTokens?: number;
+    topP: number;
+    frequencyPenalty: number;
+    presencePenalty: number;
+    n: number;
+    logitBias?: Record<string, number>;
+    user?: string;
+    streaming: boolean;
+    streamUsage?: boolean;
+    modelName: string;
+    model: import("@langchain/openai").OpenAIChatModelId;
+    stop?: string[];
+    stopSequences?: string[];
+    timeout?: number;
+    openAIApiKey?: string;
+    apiKey?: string;
+    verbosity?: import("@langchain/openai").OpenAIVerbosityParam;
+};
+//# sourceMappingURL=types.d.ts.map

package/dist/types.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../src/types.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAmB,MAAM,mBAAmB,CAAC;AAEnE,eAAO,MAAM,IAAI,SAAS,CAAC;AAC3B,eAAO,MAAM,OAAO,qTAAmS,CAAA;AAEvT,MAAM,MAAM,oBAAoB,GAAG;IAC/B,OAAO,EAAE,MAAM,CAAC;IAChB,YAAY,CAAC,EAAE,MAAM,CAAC;IACzB,mBAAmB,CAAC,EAAE,MAAM,CAAC;IAC7B,IAAI,CAAC,EAAE,YAAY,GAAG,MAAM,CAAC;IAC7B,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,oBAAoB,CAAC,EAAE,MAAM,CAAC;IAC9B,oBAAoB,CAAC,EAAE,WAAW,GAAG,aAAa,CAAC;IACnD,qBAAqB,CAAC,EAAE,eAAe,GAAG,WAAW,GAAG,SAAS,CAAC;IAClE,uBAAuB,CAAC,EAAE,WAAW,GAAG,aAAa,CAAC;IACtD,cAAc,CAAC,EAAE,WAAW,GAAG,aAAa,CAAC;IAC7C,qBAAqB,CAAC,EAAE,MAAM,CAAC;IAC/B,QAAQ,CAAC,EAAE,OAAO,CAAC;CACnB,CAAA;AAGD,wBAAgB,kBAAkB,CAAC,WAAW,EAAE,oBAAoB,EAAE,KAAK,EAAE,MAAM;;;;;;;;;;;;;;;;;;;;;;EAyBlF"}

package/dist/types.js

@@ -0,0 +1,25 @@
+export const VLLM = 'vllm';
+export const SvgIcon = `<svg height="1em" style="flex:none;line-height:1" viewBox="0 0 24 24" width="1em" xmlns="http://www.w3.org/2000/svg"><title>vLLM</title><path d="M0 4.973h9.324V23L0 4.973z" fill="#FDB515"></path><path d="M13.986 4.351L22.378 0l-6.216 23H9.324l4.662-18.649z" fill="#30A2FF"></path></svg>`;
+export function toCredentialKwargs(credentials, model) {
+    const credentialsKwargs = {
+        apiKey: credentials.api_key || 'no-key-required',
+        model: credentials.endpoint_model_name || model,
+    };
+    const configuration = {};
+    if (credentials.endpoint_url) {
+        const openaiApiBase = credentials.endpoint_url.replace(/\/$/, '');
+        configuration.baseURL = openaiApiBase;
+    }
+    // Handle thinking mode parameter
+    // Pass thinking parameter through modelKwargs for ChatOAICompatReasoningModel
+    const modelKwargs = {};
+    if (credentials.thinking != null) {
+        modelKwargs['chat_template_kwargs'] ??= {};
+        modelKwargs['chat_template_kwargs']['enable_thinking'] = !!credentials.thinking;
+    }
+    return {
+        ...credentialsKwargs,
+        modelKwargs,
+        configuration
+    };
+}

package/dist/vllm.d.ts

@@ -0,0 +1,13 @@
+import { IOnPluginBootstrap, IOnPluginDestroy } from '@xpert-ai/plugin-sdk';
+export declare class VLLMPlugin implements IOnPluginBootstrap, IOnPluginDestroy {
+    private logEnabled;
+    /**
+     * Called when the plugin is being initialized.
+     */
+    onPluginBootstrap(): void | Promise<void>;
+    /**
+     * Called when the plugin is being destroyed.
+     */
+    onPluginDestroy(): void | Promise<void>;
+}
+//# sourceMappingURL=vllm.d.ts.map

package/dist/vllm.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"vllm.d.ts","sourceRoot":"","sources":["../src/vllm.ts"],"names":[],"mappings":"AAAA,OAAO,EAAqB,kBAAkB,EAAE,gBAAgB,EAAE,MAAM,sBAAsB,CAAC;AAQ/F,qBAaa,UAAW,YAAW,kBAAkB,EAAE,gBAAgB;IAEtE,OAAO,CAAC,UAAU,CAAQ;IAE1B;;OAEG;IACH,iBAAiB,IAAI,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC;IAMzC;;OAEG;IACH,eAAe,IAAI,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC;CAKvC"}

package/dist/vllm.js

@@ -0,0 +1,46 @@
+var VLLMPlugin_1;
+import { __decorate } from "tslib";
+import { XpertServerPlugin } from '@xpert-ai/plugin-sdk';
+import { ConfigModule } from '@nestjs/config';
+import chalk from 'chalk';
+import { VLLMProviderStrategy } from './provider.strategy.js';
+import { VLLMLargeLanguageModel } from './llm/llm.js';
+import { VLLMRerankModel } from './rerank/rerank.js';
+import { VLLMTextEmbeddingModel } from './text-embedding/text-embedding.js';
+let VLLMPlugin = VLLMPlugin_1 = class VLLMPlugin {
+    constructor() {
+        // We disable by default additional logging for each event to avoid cluttering the logs
+        this.logEnabled = true;
+    }
+    /**
+     * Called when the plugin is being initialized.
+     */
+    onPluginBootstrap() {
+        if (this.logEnabled) {
+            console.log(chalk.green(`${VLLMPlugin_1.name} is being bootstrapped...`));
+        }
+    }
+    /**
+     * Called when the plugin is being destroyed.
+     */
+    onPluginDestroy() {
+        if (this.logEnabled) {
+            console.log(chalk.green(`${VLLMPlugin_1.name} is being destroyed...`));
+        }
+    }
+};
+VLLMPlugin = VLLMPlugin_1 = __decorate([
+    XpertServerPlugin({
+        /**
+         * An array of modules that will be imported and registered with the plugin.
+         */
+        imports: [ConfigModule],
+        providers: [
+            VLLMProviderStrategy,
+            VLLMLargeLanguageModel,
+            VLLMRerankModel,
+            VLLMTextEmbeddingModel
+        ]
+    })
+], VLLMPlugin);
+export { VLLMPlugin };

package/dist/vllm.yaml

@@ -0,0 +1,196 @@
+provider: vllm
+label:
+  en_US: Vllm
+description:
+  en_US: Models provided by vllm with guided inference supported.
+  zh_Hans: Vllm openai guided 支持
+icon_small:
+  en_US: vllm-color.svg
+background: "#E5E7EB"
+supported_model_types:
+  - llm
+  - text-embedding
+  - rerank
+configurate_methods:
+  - customizable-model
+model_credential_schema:
+  model:
+    label:
+      en_US: Model Name
+      zh_Hans: 模型名称
+    placeholder:
+      en_US: Enter your model name
+      zh_Hans: 输入模型名称
+  credential_form_schemas:
+    - variable: api_key
+      label:
+        en_US: API Key
+      type: secret-input
+      required: false
+      placeholder:
+        zh_Hans: 在此输入您的 API Key
+        en_US: Enter your API Key
+    - variable: endpoint_url
+      label:
+        zh_Hans: API endpoint URL
+        en_US: API endpoint URL
+      type: text-input
+      required: true
+      placeholder:
+        zh_Hans: Base URL, e.g. https://api.openai.com/v1
+        en_US: Base URL, e.g. https://api.openai.com/v1
+    - variable: endpoint_model_name
+      label:
+        zh_Hans: API endpoint中的模型名称
+        en_US: model name for API endpoint
+      type: text-input
+      required: false
+      placeholder:
+        zh_Hans: endpoint model name, e.g. chatgpt4.0
+        en_US: endpoint model name, e.g. chatgpt4.0
+    - variable: mode
+      show_on:
+        - variable: __model_type
+          value: llm
+      label:
+        en_US: Completion mode
+      type: select
+      required: false
+      default: chat
+      placeholder:
+        zh_Hans: 选择对话类型
+        en_US: Select completion mode
+      options:
+        - value: completion
+          label:
+            en_US: Completion
+            zh_Hans: 补全
+        - value: chat
+          label:
+            en_US: Chat
+            zh_Hans: 对话
+    - variable: context_size
+      label:
+        zh_Hans: 模型上下文长度
+        en_US: Model context size
+      required: true
+      show_on:
+        - variable: __model_type
+          value: llm
+      type: text-input
+      default: '4096'
+      placeholder:
+        zh_Hans: 在此输入您的模型上下文长度
+        en_US: Enter your Model context size
+    - variable: max_tokens_to_sample
+      label:
+        zh_Hans: 最大 token 上限
+        en_US: Upper bound for max tokens
+      show_on:
+        - variable: __model_type
+          value: llm
+      default: '4096'
+      type: text-input
+    - variable: agent_though_support
+      show_on:
+        - variable: __model_type
+          value: llm
+      label:
+        en_US: Agent Thought
+      type: select
+      required: false
+      default: not_supported
+      options:
+        - value: supported
+          label:
+            en_US: Support
+            zh_Hans: 支持
+        - value: not_supported
+          label:
+            en_US: Not Support
+            zh_Hans: 不支持
+    - variable: function_calling_type
+      show_on:
+        - variable: __model_type
+          value: llm
+      label:
+        en_US: Function calling
+      type: select
+      required: false
+      default: no_call
+      options:
+        - value: function_call
+          label:
+            en_US: Function Call
+            zh_Hans: Function Call
+        - value: tool_call
+          label:
+            en_US: Tool Call
+            zh_Hans: Tool Call
+        - value: no_call
+          label:
+            en_US: Not Support
+            zh_Hans: 不支持
+    - variable: stream_function_calling
+      show_on:
+        - variable: __model_type
+          value: llm
+      label:
+        en_US: Stream function calling
+      type: select
+      required: false
+      default: not_supported
+      options:
+        - value: supported
+          label:
+            en_US: Support
+            zh_Hans: 支持
+        - value: not_supported
+          label:
+            en_US: Not Support
+            zh_Hans: 不支持
+    - variable: vision_support
+      show_on:
+        - variable: __model_type
+          value: llm
+      label:
+        zh_Hans: Vision 支持
+        en_US: Vision Support
+      type: select
+      required: false
+      default: no_support
+      options:
+        - value: support
+          label:
+            en_US: Support
+            zh_Hans: 支持
+        - value: no_support
+          label:
+            en_US: Not Support
+            zh_Hans: 不支持
+    - variable: stream_mode_delimiter
+      label:
+        zh_Hans: 流模式返回结果的分隔符
+        en_US: Delimiter for streaming results
+      show_on:
+        - variable: __model_type
+          value: llm
+      default: '\n\n'
+      type: text-input
+    - variable: thinking
+      label:
+        zh_Hans: 思考模式
+        en_US: Thinking Mode
+      show_on:
+        - variable: __model_type
+          value: llm
+      type: boolean
+      default: false
+      help:
+        zh_Hans: 是否启用思考模式
+        en_US: Enable thinking mode
+extra:
+  python:
+    provider_source: provider/vllm.py
+    model_sources:
+      - "models/llm/llm.py"

package/package.json

@@ -0,0 +1,51 @@
+{
+  "name": "@chenchaolong/plugin-vllm",
+  "version": "0.0.4",
+  "author": {
+    "name": "XpertAI",
+    "url": "https://xpertai.cn"
+  },
+  "license": "AGPL-3.0",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/xpert-ai/xpert-plugins.git"
+  },
+  "bugs": {
+    "url": "https://github.com/xpert-ai/xpert-plugins/issues"
+  },
+  "type": "module",
+  "main": "./dist/index.js",
+  "module": "./dist/index.js",
+  "types": "./dist/index.d.ts",
+  "exports": {
+    "./package.json": "./package.json",
+    ".": {
+      "@xpert-plugins-starter/source": "./src/index.ts",
+      "types": "./dist/index.d.ts",
+      "import": "./dist/index.js",
+      "default": "./dist/index.js"
+    }
+  },
+  "files": [
+    "dist",
+    "src/i18n",
+    "!**/*.tsbuildinfo"
+  ],
+  "scripts": {
+    "prepack": "node ./scripts/copy-assets.mjs"
+  },
+  "dependencies": {
+    "tslib": "^2.3.0"
+  },
+  "peerDependencies": {
+    "@langchain/openai": "0.6.9",
+    "@metad/contracts": "^3.6.1",
+    "@nestjs/common": "^11.1.6",
+    "@nestjs/config": "^4.0.2",
+    "@xpert-ai/plugin-sdk": "^3.6.3",
+    "i18next": "25.6.0",
+    "lodash-es": "4.17.21",
+    "chalk": "4.1.2",
+    "zod": "3.25.67"
+  }
+}

package/src/i18n/en-US.json

@@ -0,0 +1,5 @@
+{
+    "Error": {
+        "ModelCredentialsMissing": "Credentials for model '{{model}}' are missing or deleted"
+    }
+}

package/src/i18n/en.json

@@ -0,0 +1,5 @@
+{
+    "Error": {
+        "ModelCredentialsMissing": "Credentials for model '{{model}}' are missing or deleted"
+    }
+}

package/src/i18n/zh-Hans.json

@@ -0,0 +1,5 @@
+{
+    "Error": {
+        "ModelCredentialsMissing": "模型‘{{model}}’的凭证丢失或已删除"
+    }
+}