ohlcv-ai 1.0.0

package/src/aliyun/model.ts

@@ -0,0 +1,475 @@
+export enum AliYunModelType {
+    QWEN_TURBO = 'qwen-turbo',
+    QWEN_PLUS = 'qwen-plus',
+    QWEN_MAX = 'qwen-max',
+    QWEN_MAX_LONGCONTEXT = 'qwen-max-longcontext',
+    QWEN_2_5B = 'qwen2.5-0.5b',
+    QWEN_2_5B_INSTRUCT = 'qwen2.5-0.5b-instruct',
+    QWEN_2_5B_7B = 'qwen2.5-7b',
+    QWEN_2_5B_7B_INSTRUCT = 'qwen2.5-7b-instruct',
+    QWEN_2_5B_14B = 'qwen2.5-14b',
+    QWEN_2_5B_14B_INSTRUCT = 'qwen2.5-14b-instruct',
+    QWEN_2_5B_32B = 'qwen2.5-32b',
+    QWEN_2_5B_32B_INSTRUCT = 'qwen2.5-32b-instruct',
+    QWEN_2_5B_72B = 'qwen2.5-72b',
+    QWEN_2_5B_72B_INSTRUCT = 'qwen2.5-72b-instruct',
+    QWEN_2_5B_CODER = 'qwen2.5-coder',
+    QWEN_2_5B_CODER_7B = 'qwen2.5-coder-7b',
+    QWEN_2_5B_CODER_14B = 'qwen2.5-coder-14b',
+    QWEN_2_5B_CODER_32B = 'qwen2.5-coder-32b',
+    QWEN_VL_LITE = 'qwen-vl-lite',
+    QWEN_VL_PLUS = 'qwen-vl-plus',
+    QWEN_VL_MAX = 'qwen-vl-max',
+    QWEN_AUDIO_TURBO = 'qwen-audio-turbo',
+    QWEN_AUDIO_CHAT = 'qwen-audio-chat',
+    QWEN_MATH_7B = 'qwen-math-7b',
+    LLAMA2_7B_CHAT_V2 = 'llama2-7b-chat-v2',
+    BAICHUAN2_7B_CHAT_V1 = 'baichuan2-7b-chat-v1',
+    QWEN_FINANCIAL = 'qwen-financial',
+    QWEN_FINANCIAL_14B = 'qwen-financial-14b',
+    QWEN_FINANCIAL_32B = 'qwen-financial-32b',
+    QWEN_MEDICAL = 'qwen-medical',
+    QWEN_MEDICAL_14B = 'qwen-medical-14b',
+    QWEN_MEDICAL_32B = 'qwen-medical-32b',
+    QWEN_OMNI = 'qwen-omni',
+    QWEN_OMNI_PRO = 'qwen-omni-pro'
+}
+
+export interface Model {
+    name: string;
+    displayName: string;
+    endpoint: string;
+    endpoints?: string[];
+    format: 'openai' | 'dashscope';
+    description?: string;
+    maxTokens?: number;
+    contextLength?: number;
+    capabilities?: string[];
+}
+
+export const ALIYUN_MODELS: Map<AliYunModelType, Model> = new Map([
+    [
+        AliYunModelType.QWEN_TURBO,
+        {
+            name: AliYunModelType.QWEN_TURBO,
+            displayName: 'Qwen-Turbo',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            endpoints: [
+                'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+                'https://dashscope.aliyuncs.com/api/v1/services/aigc/text-generation/generation'
+            ],
+            format: 'openai',
+            description: 'Lightweight version, fast response speed, suitable for general conversation scenarios',
+            maxTokens: 2000,
+            contextLength: 8000,
+            capabilities: ['text-generation', 'chat']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_PLUS,
+        {
+            name: AliYunModelType.QWEN_PLUS,
+            displayName: 'Qwen-Plus',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            endpoints: [
+                'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions'
+            ],
+            format: 'openai',
+            description: 'Enhanced version, suitable for complex tasks and long text processing',
+            maxTokens: 6000,
+            contextLength: 32000,
+            capabilities: ['text-generation', 'chat', 'reasoning']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_MAX,
+        {
+            name: AliYunModelType.QWEN_MAX,
+            displayName: 'Qwen-Max',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            endpoints: [
+                'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions'
+            ],
+            format: 'openai',
+            description: 'Maximum version, strongest capabilities, suitable for high-demand professional tasks',
+            maxTokens: 8000,
+            contextLength: 32000,
+            capabilities: ['text-generation', 'chat', 'reasoning', 'coding', 'analysis']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_MAX_LONGCONTEXT,
+        {
+            name: AliYunModelType.QWEN_MAX_LONGCONTEXT,
+            displayName: 'Qwen-Max-LongContext',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            endpoints: [
+                'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions'
+            ],
+            format: 'openai',
+            description: 'Supports 128K long context, suitable for long document processing',
+            maxTokens: 8000,
+            contextLength: 128000,
+            capabilities: ['text-generation', 'chat', 'document-analysis']
+        }
+    ],
+    // Qwen2.5 series models
+    [
+        AliYunModelType.QWEN_2_5B,
+        {
+            name: AliYunModelType.QWEN_2_5B,
+            displayName: 'Qwen2.5-0.5B',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: 'Ultra-lightweight 0.5B parameter model for edge devices',
+            maxTokens: 4000,
+            contextLength: 32000,
+            capabilities: ['text-generation', 'chat']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_2_5B_INSTRUCT,
+        {
+            name: AliYunModelType.QWEN_2_5B_INSTRUCT,
+            displayName: 'Qwen2.5-0.5B-Instruct',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: 'Instruction-tuned 0.5B model for specific tasks',
+            maxTokens: 4000,
+            contextLength: 32000,
+            capabilities: ['instruction-following', 'chat']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_2_5B_7B,
+        {
+            name: AliYunModelType.QWEN_2_5B_7B,
+            displayName: 'Qwen2.5-7B',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: '7B parameter base model, balanced performance and efficiency',
+            maxTokens: 6000,
+            contextLength: 32000,
+            capabilities: ['text-generation', 'reasoning']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_2_5B_7B_INSTRUCT,
+        {
+            name: AliYunModelType.QWEN_2_5B_7B_INSTRUCT,
+            displayName: 'Qwen2.5-7B-Instruct',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: 'Instruction-tuned 7B model for chat and tasks',
+            maxTokens: 6000,
+            contextLength: 32000,
+            capabilities: ['chat', 'instruction-following', 'coding']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_2_5B_14B,
+        {
+            name: AliYunModelType.QWEN_2_5B_14B,
+            displayName: 'Qwen2.5-14B',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: '14B parameter model with enhanced capabilities',
+            maxTokens: 8000,
+            contextLength: 32000,
+            capabilities: ['text-generation', 'analysis', 'reasoning']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_2_5B_32B,
+        {
+            name: AliYunModelType.QWEN_2_5B_32B,
+            displayName: 'Qwen2.5-32B',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: '32B parameter high-performance model',
+            maxTokens: 8000,
+            contextLength: 32000,
+            capabilities: ['text-generation', 'complex-reasoning', 'analysis']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_2_5B_72B,
+        {
+            name: AliYunModelType.QWEN_2_5B_72B,
+            displayName: 'Qwen2.5-72B',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: '72B parameter state-of-the-art model',
+            maxTokens: 8000,
+            contextLength: 32000,
+            capabilities: ['text-generation', 'expert-analysis', 'research']
+        }
+    ],
+    // Qwen2.5 Coder series
+    [
+        AliYunModelType.QWEN_2_5B_CODER,
+        {
+            name: AliYunModelType.QWEN_2_5B_CODER,
+            displayName: 'Qwen2.5-Coder',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: 'Specialized code generation model',
+            maxTokens: 8000,
+            contextLength: 32000,
+            capabilities: ['code-generation', 'code-explanation', 'debugging']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_2_5B_CODER_7B,
+        {
+            name: AliYunModelType.QWEN_2_5B_CODER_7B,
+            displayName: 'Qwen2.5-Coder-7B',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: '7B parameter code generation model',
+            maxTokens: 8000,
+            contextLength: 32000,
+            capabilities: ['code-generation', 'programming']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_2_5B_CODER_14B,
+        {
+            name: AliYunModelType.QWEN_2_5B_CODER_14B,
+            displayName: 'Qwen2.5-Coder-14B',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: '14B parameter advanced code generation model',
+            maxTokens: 8000,
+            contextLength: 32000,
+            capabilities: ['code-generation', 'code-review', 'optimization']
+        }
+    ],
+    // Vision-Language models
+    [
+        AliYunModelType.QWEN_VL_LITE,
+        {
+            name: AliYunModelType.QWEN_VL_LITE,
+            displayName: 'Qwen-VL-Lite',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: 'Lightweight vision-language model for basic image understanding',
+            maxTokens: 2000,
+            contextLength: 8000,
+            capabilities: ['image-understanding', 'visual-qa']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_VL_PLUS,
+        {
+            name: AliYunModelType.QWEN_VL_PLUS,
+            displayName: 'Qwen-VL-Plus',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: 'Vision-language model supporting image understanding',
+            maxTokens: 4000,
+            contextLength: 32000,
+            capabilities: ['image-understanding', 'document-analysis', 'visual-reasoning']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_VL_MAX,
+        {
+            name: AliYunModelType.QWEN_VL_MAX,
+            displayName: 'Qwen-VL-Max',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: 'Most powerful vision-language model',
+            maxTokens: 8000,
+            contextLength: 32000,
+            capabilities: ['image-understanding', 'video-analysis', 'multimodal-reasoning']
+        }
+    ],
+    // Audio models
+    [
+        AliYunModelType.QWEN_AUDIO_TURBO,
+        {
+            name: AliYunModelType.QWEN_AUDIO_TURBO,
+            displayName: 'Qwen-Audio-Turbo',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: 'Fast audio processing and speech-to-text model',
+            maxTokens: 2000,
+            contextLength: 8000,
+            capabilities: ['speech-recognition', 'audio-analysis']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_AUDIO_CHAT,
+        {
+            name: AliYunModelType.QWEN_AUDIO_CHAT,
+            displayName: 'Qwen-Audio-Chat',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: 'Audio conversation and processing model',
+            maxTokens: 4000,
+            contextLength: 32000,
+            capabilities: ['audio-chat', 'voice-assistant', 'speech-synthesis']
+        }
+    ],
+    // Specialized models
+    [
+        AliYunModelType.QWEN_MATH_7B,
+        {
+            name: AliYunModelType.QWEN_MATH_7B,
+            displayName: 'Qwen-Math-7B',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: 'Specialized for mathematical reasoning and problem solving',
+            maxTokens: 4000,
+            contextLength: 32000,
+            capabilities: ['mathematical-reasoning', 'problem-solving']
+        }
+    ],
+    [
+        AliYunModelType.LLAMA2_7B_CHAT_V2,
+        {
+            name: AliYunModelType.LLAMA2_7B_CHAT_V2,
+            displayName: 'LLaMA2-7B-Chat',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: 'Meta\'s LLaMA2-7B model',
+            maxTokens: 2000,
+            contextLength: 8000,
+            capabilities: ['chat', 'text-generation']
+        }
+    ],
+    [
+        AliYunModelType.BAICHUAN2_7B_CHAT_V1,
+        {
+            name: AliYunModelType.BAICHUAN2_7B_CHAT_V1,
+            displayName: 'Baichuan2-7B-Chat',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: 'Baichuan AI\'s Baichuan2-7B model',
+            maxTokens: 2000,
+            contextLength: 8000,
+            capabilities: ['chat', 'chinese-nlp']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_FINANCIAL,
+        {
+            name: AliYunModelType.QWEN_FINANCIAL,
+            displayName: 'Qwen-Financial',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: 'Specialized for financial analysis and market insights',
+            maxTokens: 6000,
+            contextLength: 32000,
+            capabilities: ['financial-analysis', 'market-prediction', 'risk-assessment']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_MEDICAL,
+        {
+            name: AliYunModelType.QWEN_MEDICAL,
+            displayName: 'Qwen-Medical',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: 'Specialized for medical consultation and health analysis',
+            maxTokens: 6000,
+            contextLength: 32000,
+            capabilities: ['medical-consultation', 'health-analysis', 'diagnostic-support']
+        }
+    ],
+    // Omni models (multimodal)
+    [
+        AliYunModelType.QWEN_OMNI,
+        {
+            name: AliYunModelType.QWEN_OMNI,
+            displayName: 'Qwen-Omni',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: 'Omnidirectional multimodal model supporting text, image, audio',
+            maxTokens: 8000,
+            contextLength: 64000,
+            capabilities: ['text-generation', 'image-understanding', 'audio-processing', 'multimodal']
+        }
+    ],
+    [
+        AliYunModelType.QWEN_OMNI_PRO,
+        {
+            name: AliYunModelType.QWEN_OMNI_PRO,
+            displayName: 'Qwen-Omni-Pro',
+            endpoint: 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions',
+            format: 'openai',
+            description: 'Professional omnidirectional multimodal model with advanced capabilities',
+            maxTokens: 16000,
+            contextLength: 128000,
+            capabilities: ['text-generation', 'multimodal', 'complex-reasoning', 'expert-analysis']
+        }
+    ]
+]);
+
+export function getModel(type: AliYunModelType): Model | undefined {
+    return ALIYUN_MODELS.get(type);
+}
+
+export function getAllModels(): Model[] {
+    return Array.from(ALIYUN_MODELS.values());
+}
+
+export function getModelByName(name: string): Model | undefined {
+    for (const model of ALIYUN_MODELS.values()) {
+        if (model.name === name) {
+            return model;
+        }
+    }
+    return undefined;
+}
+
+export function getAvailableAliYunModelTypes(): AliYunModelType[] {
+    return Array.from(ALIYUN_MODELS.keys());
+}
+
+// Helper functions for specific model categories
+export function getTextModels(): Model[] {
+    return getAllModels().filter(model => 
+        model.capabilities?.includes('text-generation') && 
+        !model.capabilities?.includes('image-understanding') &&
+        !model.capabilities?.includes('audio-processing')
+    );
+}
+
+export function getVisionModels(): Model[] {
+    return getAllModels().filter(model => 
+        model.capabilities?.includes('image-understanding')
+    );
+}
+
+export function getAudioModels(): Model[] {
+    return getAllModels().filter(model => 
+        model.capabilities?.includes('audio-processing') ||
+        model.capabilities?.includes('speech-recognition')
+    );
+}
+
+export function getCodingModels(): Model[] {
+    return getAllModels().filter(model => 
+        model.capabilities?.includes('code-generation') ||
+        model.capabilities?.includes('programming')
+    );
+}
+
+export function getSpecializedModels(): Model[] {
+    const specializedCapabilities = [
+        'financial-analysis', 
+        'medical-consultation', 
+        'mathematical-reasoning'
+    ];
+    return getAllModels().filter(model => 
+        model.capabilities?.some(cap => specializedCapabilities.includes(cap))
+    );
+}
+
+export function getMultimodalModels(): Model[] {
+    return getAllModels().filter(model => 
+        model.capabilities?.includes('multimodal')
+    );
+}