@synapseia-network/node 0.8.5

package/dist/chunk-PRVT22SM.js

@@ -0,0 +1,324 @@
+import { fileURLToPath as __synFup } from "url";import { dirname as __synDn } from "path";const __filename = __synFup(import.meta.url);const __dirname = __synDn(__filename);
+import {
+  init_logger,
+  logger_default
+} from "./chunk-V2L5SXTL.js";
+import {
+  __name
+} from "./chunk-D7ADMHK2.js";
+
+// src/modules/model/model-catalog.ts
+init_logger();
+import { Injectable } from "@nestjs/common";
+import { execSync } from "child_process";
+function _ts_decorate(decorators, target, key, desc) {
+  var c = arguments.length, r = c < 3 ? target : desc === null ? desc = Object.getOwnPropertyDescriptor(target, key) : desc, d;
+  if (typeof Reflect === "object" && typeof Reflect.decorate === "function") r = Reflect.decorate(decorators, target, key, desc);
+  else for (var i = decorators.length - 1; i >= 0; i--) if (d = decorators[i]) r = (c < 3 ? d(r) : c > 3 ? d(target, key, r) : d(target, key)) || r;
+  return c > 3 && r && Object.defineProperty(target, key, r), r;
+}
+__name(_ts_decorate, "_ts_decorate");
+var MODEL_CATALOG = [
+  // Embedding models
+  {
+    name: "locusai/all-minilm-l6-v2",
+    minVram: 0.1,
+    recommendedTier: 0,
+    category: "embedding",
+    description: "Lightweight embedding model for vector search"
+  },
+  // General models (small)
+  {
+    name: "qwen2.5-0.5b",
+    minVram: 1,
+    recommendedTier: 1,
+    category: "general",
+    description: "Tiny general-purpose model"
+  },
+  {
+    name: "gemma-3-1b-web",
+    minVram: 2,
+    recommendedTier: 1,
+    category: "general",
+    description: "Small web-optimized general model"
+  },
+  {
+    name: "phi-2",
+    minVram: 2,
+    recommendedTier: 1,
+    category: "general",
+    description: "Microsoft Phi-2 small model"
+  },
+  {
+    name: "tiny-vicuna-1b",
+    minVram: 1,
+    recommendedTier: 1,
+    category: "general",
+    description: "Tiny general-purpose model"
+  },
+  {
+    name: "home-3b-v3",
+    minVram: 2,
+    recommendedTier: 1,
+    category: "general",
+    description: "Home-3B v3 small model"
+  },
+  {
+    name: "qwen2-0.5b",
+    minVram: 1,
+    recommendedTier: 1,
+    category: "general",
+    description: "Qwen2 0.5B tiny model"
+  },
+  {
+    name: "qwen2-0.5b-instruct",
+    minVram: 1,
+    recommendedTier: 1,
+    category: "general",
+    description: "Qwen2 0.5B instruct-tuned"
+  },
+  // Code models (small)
+  {
+    name: "qwen2.5-coder-0.5b",
+    minVram: 1,
+    recommendedTier: 1,
+    category: "code",
+    description: "Tiny code model"
+  },
+  {
+    name: "qwen2.5-coder-1.5b",
+    minVram: 2,
+    recommendedTier: 1,
+    category: "code",
+    description: "Small code model"
+  },
+  // General models (medium)
+  {
+    name: "qwen2.5-coder-3b",
+    minVram: 3,
+    recommendedTier: 2,
+    category: "code",
+    description: "Medium code model"
+  },
+  {
+    name: "gemma-3-1b",
+    minVram: 2,
+    recommendedTier: 1,
+    category: "general",
+    description: "Google Gemma 3 1B"
+  },
+  {
+    name: "gemma-3-4b",
+    minVram: 4,
+    recommendedTier: 2,
+    category: "general",
+    description: "Google Gemma 3 4B"
+  },
+  // Code models (medium-large)
+  {
+    name: "qwen2.5-coder-7b",
+    minVram: 6,
+    recommendedTier: 2,
+    category: "code",
+    description: "7B code model"
+  },
+  {
+    name: "glm-4-9b",
+    minVram: 8,
+    recommendedTier: 2,
+    category: "general",
+    description: "GLM-4 9B general model"
+  },
+  {
+    name: "mistral-7b-instruct",
+    minVram: 6,
+    recommendedTier: 2,
+    category: "general",
+    description: "Mistral 7B instruct model"
+  },
+  // General models (large)
+  {
+    name: "gemma-3-12b",
+    minVram: 10,
+    recommendedTier: 3,
+    category: "general",
+    description: "Google Gemma 3 12B"
+  },
+  {
+    name: "llama-3.1-8b-instruct",
+    minVram: 10,
+    recommendedTier: 3,
+    category: "general",
+    description: "Meta Llama 3.1 8B instruct"
+  },
+  {
+    name: "llama-3.2-1b-instruct",
+    minVram: 1,
+    recommendedTier: 1,
+    category: "general",
+    description: "Meta Llama 3.2 1B instruct"
+  },
+  // Code models (large)
+  {
+    name: "qwen2.5-coder-14b",
+    minVram: 12,
+    recommendedTier: 3,
+    category: "code",
+    description: "14B code model"
+  },
+  // General models (very large)
+  {
+    name: "gpt-oss-20b",
+    minVram: 16,
+    recommendedTier: 4,
+    category: "general",
+    description: "20B general model"
+  },
+  {
+    name: "gemma-3-27b",
+    minVram: 20,
+    recommendedTier: 4,
+    category: "general",
+    description: "Google Gemma 3 27B"
+  },
+  // Code models (very large)
+  {
+    name: "qwen2.5-coder-32b",
+    minVram: 24,
+    recommendedTier: 4,
+    category: "code",
+    description: "32B code model (recommended)"
+  },
+  // General models (ultra-large)
+  {
+    name: "glm-4.7-flash",
+    minVram: 24,
+    recommendedTier: 4,
+    category: "general",
+    description: "GLM-4.7 Flash ultra model"
+  },
+  {
+    name: "qwen3-coder-30b-a3b",
+    minVram: 24,
+    recommendedTier: 4,
+    category: "code",
+    description: "Qwen3 Coder 30B A3B"
+  }
+];
+var CLOUD_MODELS = [
+  {
+    name: "gemini-2.0-flash",
+    minVram: 0,
+    recommendedTier: 0,
+    category: "general",
+    description: "Google Gemini 2.0 Flash (cloud-only)",
+    isCloud: true
+  },
+  {
+    name: "Minimax2.7",
+    minVram: 0,
+    recommendedTier: 0,
+    category: "general",
+    description: "Minimax2.7 (cloud-only)",
+    isCloud: true
+  },
+  {
+    name: "Kimi2.5",
+    minVram: 0,
+    recommendedTier: 0,
+    category: "general",
+    description: "Kimi2.5 (cloud-only)",
+    isCloud: true
+  }
+];
+var FULL_CATALOG = [
+  ...MODEL_CATALOG,
+  ...CLOUD_MODELS
+];
+var ModelCatalogHelper = class {
+  static {
+    __name(this, "ModelCatalogHelper");
+  }
+  listModels(category) {
+    if (category) return FULL_CATALOG.filter((m) => m.category === category);
+    return FULL_CATALOG;
+  }
+  getModelsForVram(vramGb) {
+    return FULL_CATALOG.filter((m) => m.minVram <= vramGb && !m.isCloud);
+  }
+  getModel(name) {
+    return FULL_CATALOG.find((m) => m.name === name);
+  }
+  async pullModel(name) {
+    try {
+      execSync("curl -s http://localhost:11434/api/tags", {
+        stdio: "pipe",
+        timeout: 1e3
+      });
+    } catch {
+      throw new Error("Ollama is not running. Start it with: ollama serve");
+    }
+    try {
+      logger_default.log(`[ModelCatalog] pulling model ${name}...`);
+      execSync(`ollama pull ${name}`, {
+        stdio: "inherit"
+      });
+      return true;
+    } catch (error) {
+      throw new Error(`Failed to pull model ${name}: ${error}`);
+    }
+  }
+  /**
+  * List locally available Ollama models.
+  * @param ollamaUrl - Optional base URL (e.g. `http://ollama:11434`). Falls back to
+  *   `OLLAMA_URL` env var, then `http://localhost:11434`. Needed when Ollama runs on
+  *   a different host (e.g. Docker Compose, where `localhost` inside the node
+  *   container is not the ollama service).
+  */
+  getLocalModels(ollamaUrl) {
+    const baseUrl = ollamaUrl ?? process.env.OLLAMA_URL ?? "http://localhost:11434";
+    try {
+      const response = execSync(`curl -s ${baseUrl}/api/tags`, {
+        encoding: "utf-8"
+      });
+      const data = JSON.parse(response);
+      return data.models.map((m) => m.name);
+    } catch {
+      return [];
+    }
+  }
+  isModelAvailable(name) {
+    return this.getLocalModels().includes(name);
+  }
+  getRecommendedModel(tier, category) {
+    let models = this.getModelsForVram(tier * 16);
+    if (category) models = models.filter((m) => m.category === category);
+    models.sort((a, b) => {
+      if (a.recommendedTier !== b.recommendedTier) return a.recommendedTier - b.recommendedTier;
+      return b.minVram - a.minVram;
+    });
+    return models[0];
+  }
+  getModelCatalog() {
+    return [
+      ...MODEL_CATALOG
+    ];
+  }
+  normalizeModelName(name) {
+    return name.replace(/^ollama\//, "").replace(/:/g, "-").toLowerCase();
+  }
+  getModelByName(name) {
+    const normalized = this.normalizeModelName(name);
+    return MODEL_CATALOG.find((m) => m.name === normalized || m.name === name) || null;
+  }
+};
+ModelCatalogHelper = _ts_decorate([
+  Injectable()
+], ModelCatalogHelper);
+
+export {
+  MODEL_CATALOG,
+  CLOUD_MODELS,
+  FULL_CATALOG,
+  ModelCatalogHelper
+};