centaurus-cli 3.0.0 → 3.1.0

package/dist/services/ollama-service.js

@@ -1,324 +1,307 @@
-/**
- * Ollama Service
- * Client for interacting with the local Ollama API
- *
- * Ollama exposes a REST API at http://localhost:11434 by default.
- * This service provides methods for:
- * - Checking if Ollama is running
- * - Fetching available local models
- * - Sending chat messages to local models
- */
-import { logError } from '../utils/logger.js';
-import { quickLog } from '../utils/conversation-logger.js';
-// Ollama API base URL (configurable via environment variable)
-const OLLAMA_BASE_URL = process.env.OLLAMA_HOST || 'http://localhost:11434';
-// Timeout for Ollama API requests (5 seconds)
-const REQUEST_TIMEOUT = 5000;
-/**
- * Model families that support tool calling
- * Based on https://ollama.com/search?c=tools
- */
-export const TOOL_CAPABLE_MODEL_FAMILIES = [
-    'llama3.1', 'llama3.2', 'llama3.3',
-    'mistral-nemo', 'mistral-small',
-    'command-r', 'command-r-plus',
-    'firefunction',
-    'qwen2.5', 'qwen2.5-coder',
-    'granite3-dense', 'granite3.1-dense',
-    'nemotron',
-    'hermes3',
-    'athene-v2',
-    'gpt-oss',
-    'exaone3.5',
-    'deepseek-r1',
-    'qwen3-coder',
-    'qwen-coder',
-    'deepseek-v3',
-    'deepseek-v2',
-    'deepseek',
+import { logError } from "../utils/logger.js";
+import { quickLog } from "../utils/conversation-logger.js";
+const OLLAMA_BASE_URL = process.env.OLLAMA_HOST || "http://localhost:11434";
+const REQUEST_TIMEOUT = 5e3;
+const TOOL_CAPABLE_MODEL_FAMILIES = [
+  "llama3.1",
+  "llama3.2",
+  "llama3.3",
+  "mistral-nemo",
+  "mistral-small",
+  "command-r",
+  "command-r-plus",
+  "firefunction",
+  "qwen2.5",
+  "qwen2.5-coder",
+  "granite3-dense",
+  "granite3.1-dense",
+  "nemotron",
+  "hermes3",
+  "athene-v2",
+  "gpt-oss",
+  "exaone3.5",
+  "deepseek-r1",
+  "qwen3-coder",
+  "qwen-coder",
+  "deepseek-v3",
+  "deepseek-v2",
+  "deepseek"
 ];
-/**
- * Ollama Service Class
- * Provides methods for interacting with the local Ollama API
- */
 class OllamaService {
-    baseUrl;
-    constructor(baseUrl = OLLAMA_BASE_URL) {
-        this.baseUrl = baseUrl;
+  baseUrl;
+  constructor(baseUrl = OLLAMA_BASE_URL) {
+    this.baseUrl = baseUrl;
+  }
+  /**
+   * Check if Ollama is running and accessible
+   * @returns OllamaStatus indicating if Ollama is available
+   */
+  async isOllamaRunning() {
+    try {
+      const controller = new AbortController();
+      const timeoutId = setTimeout(() => controller.abort(), REQUEST_TIMEOUT);
+      const response = await fetch(`${this.baseUrl}/api/version`, {
+        method: "GET",
+        signal: controller.signal
+      });
+      clearTimeout(timeoutId);
+      if (response.ok) {
+        const data = await response.json();
+        return {
+          available: true,
+          version: data.version || "unknown"
+        };
+      }
+      return {
+        available: false,
+        error: `Ollama returned status ${response.status}`
+      };
+    } catch (error) {
+      quickLog(`[${(/* @__PURE__ */ new Date()).toISOString()}] [OllamaService] isOllamaRunning error: ${error.message}
+`);
+      if (error.name === "AbortError") {
+        return {
+          available: false,
+          error: "Connection timed out. Ollama may not be running."
+        };
+      }
+      if (error.code === "ECONNREFUSED") {
+        return {
+          available: false,
+          error: "Connection refused. Ollama is not running or not installed."
+        };
+      }
+      return {
+        available: false,
+        error: error.message || "Unknown error connecting to Ollama"
+      };
     }
-    /**
-     * Check if Ollama is running and accessible
-     * @returns OllamaStatus indicating if Ollama is available
-     */
-    async isOllamaRunning() {
-        try {
-            const controller = new AbortController();
-            const timeoutId = setTimeout(() => controller.abort(), REQUEST_TIMEOUT);
-            // Ollama has a simple root endpoint that returns version info
-            const response = await fetch(`${this.baseUrl}/api/version`, {
-                method: 'GET',
-                signal: controller.signal,
-            });
-            clearTimeout(timeoutId);
-            if (response.ok) {
-                const data = await response.json();
-                return {
-                    available: true,
-                    version: data.version || 'unknown',
-                };
-            }
-            return {
-                available: false,
-                error: `Ollama returned status ${response.status}`,
-            };
-        }
-        catch (error) {
-            quickLog(`[${new Date().toISOString()}] [OllamaService] isOllamaRunning error: ${error.message}\n`);
-            if (error.name === 'AbortError') {
-                return {
-                    available: false,
-                    error: 'Connection timed out. Ollama may not be running.',
-                };
-            }
-            if (error.code === 'ECONNREFUSED') {
-                return {
-                    available: false,
-                    error: 'Connection refused. Ollama is not running or not installed.',
-                };
-            }
-            return {
-                available: false,
-                error: error.message || 'Unknown error connecting to Ollama',
-            };
-        }
+  }
+  /**
+   * Get list of locally available models from Ollama
+   * @returns Array of OllamaModel objects
+   * @throws Error if Ollama is not available or request fails
+   */
+  async getLocalModels() {
+    try {
+      const controller = new AbortController();
+      const timeoutId = setTimeout(() => controller.abort(), REQUEST_TIMEOUT);
+      const response = await fetch(`${this.baseUrl}/api/tags`, {
+        method: "GET",
+        signal: controller.signal
+      });
+      clearTimeout(timeoutId);
+      if (!response.ok) {
+        throw new Error(`Ollama API returned status ${response.status}`);
+      }
+      const data = await response.json();
+      quickLog(`[${(/* @__PURE__ */ new Date()).toISOString()}] [OllamaService] getLocalModels: Found ${data.models?.length || 0} models
+`);
+      return data.models || [];
+    } catch (error) {
+      quickLog(`[${(/* @__PURE__ */ new Date()).toISOString()}] [OllamaService] getLocalModels error: ${error.message}
+`);
+      if (error.name === "AbortError") {
+        throw new Error("Connection to Ollama timed out. Please ensure Ollama is running.");
+      }
+      if (error.code === "ECONNREFUSED") {
+        throw new Error("Cannot connect to Ollama. Please ensure Ollama is installed and running.\n\nInstall from: https://ollama.ai");
+      }
+      throw error;
     }
-    /**
-     * Get list of locally available models from Ollama
-     * @returns Array of OllamaModel objects
-     * @throws Error if Ollama is not available or request fails
-     */
-    async getLocalModels() {
-        try {
-            const controller = new AbortController();
-            const timeoutId = setTimeout(() => controller.abort(), REQUEST_TIMEOUT);
-            const response = await fetch(`${this.baseUrl}/api/tags`, {
-                method: 'GET',
-                signal: controller.signal,
-            });
-            clearTimeout(timeoutId);
-            if (!response.ok) {
-                throw new Error(`Ollama API returned status ${response.status}`);
-            }
-            const data = await response.json();
-            quickLog(`[${new Date().toISOString()}] [OllamaService] getLocalModels: Found ${data.models?.length || 0} models\n`);
-            return data.models || [];
-        }
-        catch (error) {
-            quickLog(`[${new Date().toISOString()}] [OllamaService] getLocalModels error: ${error.message}\n`);
-            if (error.name === 'AbortError') {
-                throw new Error('Connection to Ollama timed out. Please ensure Ollama is running.');
-            }
-            if (error.code === 'ECONNREFUSED') {
-                throw new Error('Cannot connect to Ollama. Please ensure Ollama is installed and running.\n\nInstall from: https://ollama.ai');
-            }
-            throw error;
-        }
+  }
+  /**
+   * Send a chat message to an Ollama model (non-streaming)
+   * @param model Model name to use (e.g., "llama3:latest")
+   * @param messages Array of chat messages
+   * @param tools Optional array of tool definitions
+   * @param options Optional generation parameters
+   * @returns OllamaChatResponse with the model's reply
+   */
+  async sendChatMessage(model, messages, tools, options) {
+    try {
+      const request = {
+        model,
+        messages,
+        stream: false,
+        tools: tools && tools.length > 0 ? tools : void 0,
+        options
+      };
+      quickLog(`[${(/* @__PURE__ */ new Date()).toISOString()}] [OllamaService] sendChatMessage: model=${model}, messages=${messages.length}, tools=${tools?.length || 0}
+`);
+      const response = await fetch(`${this.baseUrl}/api/chat`, {
+        method: "POST",
+        headers: {
+          "Content-Type": "application/json"
+        },
+        body: JSON.stringify(request)
+      });
+      if (!response.ok) {
+        const errorText = await response.text();
+        throw new Error(`Ollama API error: ${response.status} - ${errorText}`);
+      }
+      const data = await response.json();
+      if (data.message?.tool_calls?.length) {
+        quickLog(`[${(/* @__PURE__ */ new Date()).toISOString()}] [OllamaService] Response contains ${data.message.tool_calls.length} tool calls
+`);
+      }
+      return data;
+    } catch (error) {
+      logError(`Ollama sendChatMessage error: ${error.message}`);
+      throw error;
     }
-    /**
-     * Send a chat message to an Ollama model (non-streaming)
-     * @param model Model name to use (e.g., "llama3:latest")
-     * @param messages Array of chat messages
-     * @param tools Optional array of tool definitions
-     * @param options Optional generation parameters
-     * @returns OllamaChatResponse with the model's reply
-     */
-    async sendChatMessage(model, messages, tools, options) {
-        try {
-            const request = {
-                model,
-                messages,
-                stream: false,
-                tools: tools && tools.length > 0 ? tools : undefined,
-                options,
-            };
-            quickLog(`[${new Date().toISOString()}] [OllamaService] sendChatMessage: model=${model}, messages=${messages.length}, tools=${tools?.length || 0}\n`);
-            const response = await fetch(`${this.baseUrl}/api/chat`, {
-                method: 'POST',
-                headers: {
-                    'Content-Type': 'application/json',
-                },
-                body: JSON.stringify(request),
-            });
-            if (!response.ok) {
-                const errorText = await response.text();
-                throw new Error(`Ollama API error: ${response.status} - ${errorText}`);
+  }
+  /**
+   * Check if a model name indicates it supports tool calling
+   * @param modelName The model name (e.g., "llama3.2:latest")
+   * @returns True if the model likely supports tools
+   */
+  static modelSupportsTools(modelName) {
+    const lowerName = modelName.toLowerCase();
+    return TOOL_CAPABLE_MODEL_FAMILIES.some(
+      (family) => lowerName.startsWith(family) || lowerName.includes(`:${family}`)
+    );
+  }
+  /**
+   * Send a streaming chat message to an Ollama model
+   * @param model Model name to use
+   * @param messages Array of chat messages
+   * @param onChunk Callback for each streaming chunk
+   * @param options Optional generation parameters
+   * @returns Final complete response after stream ends
+   */
+  async sendChatMessageStreaming(model, messages, onChunk, options) {
+    try {
+      const request = {
+        model,
+        messages,
+        stream: true,
+        options
+      };
+      quickLog(`[${(/* @__PURE__ */ new Date()).toISOString()}] [OllamaService] Starting streaming chat with model: ${model}
+`);
+      const response = await fetch(`${this.baseUrl}/api/chat`, {
+        method: "POST",
+        headers: {
+          "Content-Type": "application/json"
+        },
+        body: JSON.stringify(request)
+      });
+      if (!response.ok) {
+        const errorText = await response.text();
+        throw new Error(`Ollama API error: ${response.status} - ${errorText}`);
+      }
+      if (!response.body) {
+        throw new Error("No response body from Ollama");
+      }
+      const reader = response.body.getReader();
+      const decoder = new TextDecoder();
+      let fullContent = "";
+      let lastChunk = null;
+      while (true) {
+        const { done, value } = await reader.read();
+        if (done) break;
+        const text = decoder.decode(value, { stream: true });
+        const lines = text.split("\n").filter((line) => line.trim());
+        for (const line of lines) {
+          try {
+            const chunk = JSON.parse(line);
+            fullContent += chunk.message?.content || "";
+            lastChunk = chunk;
+            onChunk(chunk);
+            if (chunk.done) {
+              break;
             }
-            const data = await response.json();
-            // Log if tool calls were returned
-            if (data.message?.tool_calls?.length) {
-                quickLog(`[${new Date().toISOString()}] [OllamaService] Response contains ${data.message.tool_calls.length} tool calls\n`);
-            }
-            return data;
-        }
-        catch (error) {
-            logError(`Ollama sendChatMessage error: ${error.message}`);
-            throw error;
+          } catch (parseError) {
+            quickLog(`[${(/* @__PURE__ */ new Date()).toISOString()}] [OllamaService] Skipping malformed chunk: ${line}
+`);
+          }
         }
+      }
+      const finalResponse = {
+        model,
+        created_at: lastChunk?.created_at || (/* @__PURE__ */ new Date()).toISOString(),
+        message: {
+          role: "assistant",
+          content: fullContent
+        },
+        done: true,
+        total_duration: lastChunk?.total_duration || 0,
+        load_duration: lastChunk?.load_duration || 0,
+        prompt_eval_count: lastChunk?.prompt_eval_count || 0,
+        prompt_eval_duration: lastChunk?.prompt_eval_duration || 0,
+        eval_count: lastChunk?.eval_count || 0,
+        eval_duration: lastChunk?.eval_duration || 0
+      };
+      quickLog(`[${(/* @__PURE__ */ new Date()).toISOString()}] [OllamaService] Streaming complete. Total content length: ${fullContent.length}
+`);
+      return finalResponse;
+    } catch (error) {
+      logError(`Ollama streaming error: ${error.message}`);
+      throw error;
     }
-    /**
-     * Check if a model name indicates it supports tool calling
-     * @param modelName The model name (e.g., "llama3.2:latest")
-     * @returns True if the model likely supports tools
-     */
-    static modelSupportsTools(modelName) {
-        const lowerName = modelName.toLowerCase();
-        // Check against known tool-capable model families
-        return TOOL_CAPABLE_MODEL_FAMILIES.some(family => lowerName.startsWith(family) || lowerName.includes(`:${family}`));
-    }
-    /**
-     * Send a streaming chat message to an Ollama model
-     * @param model Model name to use
-     * @param messages Array of chat messages
-     * @param onChunk Callback for each streaming chunk
-     * @param options Optional generation parameters
-     * @returns Final complete response after stream ends
-     */
-    async sendChatMessageStreaming(model, messages, onChunk, options) {
-        try {
-            const request = {
-                model,
-                messages,
-                stream: true,
-                options,
-            };
-            quickLog(`[${new Date().toISOString()}] [OllamaService] Starting streaming chat with model: ${model}\n`);
-            const response = await fetch(`${this.baseUrl}/api/chat`, {
-                method: 'POST',
-                headers: {
-                    'Content-Type': 'application/json',
-                },
-                body: JSON.stringify(request),
-            });
-            if (!response.ok) {
-                const errorText = await response.text();
-                throw new Error(`Ollama API error: ${response.status} - ${errorText}`);
-            }
-            if (!response.body) {
-                throw new Error('No response body from Ollama');
-            }
-            const reader = response.body.getReader();
-            const decoder = new TextDecoder();
-            let fullContent = '';
-            let lastChunk = null;
-            while (true) {
-                const { done, value } = await reader.read();
-                if (done)
-                    break;
-                const text = decoder.decode(value, { stream: true });
-                const lines = text.split('\n').filter(line => line.trim());
-                for (const line of lines) {
-                    try {
-                        const chunk = JSON.parse(line);
-                        fullContent += chunk.message?.content || '';
-                        lastChunk = chunk;
-                        // Invoke callback with each chunk
-                        onChunk(chunk);
-                        if (chunk.done) {
-                            break;
-                        }
-                    }
-                    catch (parseError) {
-                        // Skip malformed JSON lines
-                        quickLog(`[${new Date().toISOString()}] [OllamaService] Skipping malformed chunk: ${line}\n`);
-                    }
-                }
-            }
-            // Return a complete response object
-            const finalResponse = {
-                model,
-                created_at: lastChunk?.created_at || new Date().toISOString(),
-                message: {
-                    role: 'assistant',
-                    content: fullContent,
-                },
-                done: true,
-                total_duration: lastChunk?.total_duration || 0,
-                load_duration: lastChunk?.load_duration || 0,
-                prompt_eval_count: lastChunk?.prompt_eval_count || 0,
-                prompt_eval_duration: lastChunk?.prompt_eval_duration || 0,
-                eval_count: lastChunk?.eval_count || 0,
-                eval_duration: lastChunk?.eval_duration || 0,
-            };
-            quickLog(`[${new Date().toISOString()}] [OllamaService] Streaming complete. Total content length: ${fullContent.length}\n`);
-            return finalResponse;
-        }
-        catch (error) {
-            logError(`Ollama streaming error: ${error.message}`);
-            throw error;
-        }
+  }
+  /**
+   * Format model size for display (bytes to human-readable)
+   * @param bytes Size in bytes
+   * @returns Human-readable size string
+   */
+  static formatModelSize(bytes) {
+    if (bytes < 1024) return `${bytes} B`;
+    if (bytes < 1024 * 1024) return `${(bytes / 1024).toFixed(1)} KB`;
+    if (bytes < 1024 * 1024 * 1024) return `${(bytes / (1024 * 1024)).toFixed(1)} MB`;
+    return `${(bytes / (1024 * 1024 * 1024)).toFixed(1)} GB`;
+  }
+  /**
+   * Get a user-friendly error message for common Ollama issues
+   * @param error The error that occurred
+   * @returns User-friendly error message with helpful instructions
+   */
+  static getHelpfulErrorMessage(error) {
+    const message = error.message.toLowerCase();
+    if (message.includes("econnrefused") || message.includes("connection refused")) {
+      return `\u274C Cannot connect to Ollama
+
+Ollama is not running or not installed on this machine.
+
+To fix this:
+1. Install Ollama from: https://ollama.ai
+2. Start Ollama by running: ollama serve
+3. Pull a model: ollama pull llama3
+
+Then try again with /models local`;
     }
-    /**
-     * Format model size for display (bytes to human-readable)
-     * @param bytes Size in bytes
-     * @returns Human-readable size string
-     */
-    static formatModelSize(bytes) {
-        if (bytes < 1024)
-            return `${bytes} B`;
-        if (bytes < 1024 * 1024)
-            return `${(bytes / 1024).toFixed(1)} KB`;
-        if (bytes < 1024 * 1024 * 1024)
-            return `${(bytes / (1024 * 1024)).toFixed(1)} MB`;
-        return `${(bytes / (1024 * 1024 * 1024)).toFixed(1)} GB`;
+    if (message.includes("timeout") || message.includes("timed out")) {
+      return `\u274C Connection to Ollama timed out
+
+Ollama may be starting up or under heavy load.
+
+Please ensure Ollama is running:
+\u2022 Check if 'ollama serve' is running
+\u2022 Try again in a few seconds`;
     }
-    /**
-     * Get a user-friendly error message for common Ollama issues
-     * @param error The error that occurred
-     * @returns User-friendly error message with helpful instructions
-     */
-    static getHelpfulErrorMessage(error) {
-        const message = error.message.toLowerCase();
-        if (message.includes('econnrefused') || message.includes('connection refused')) {
-            return `❌ Cannot connect to Ollama
-
-Ollama is not running or not installed on this machine.
-
-To fix this:
-1. Install Ollama from: https://ollama.ai
-2. Start Ollama by running: ollama serve
-3. Pull a model: ollama pull llama3
-
-Then try again with /models local`;
-        }
-        if (message.includes('timeout') || message.includes('timed out')) {
-            return `❌ Connection to Ollama timed out
-
-Ollama may be starting up or under heavy load.
-
-Please ensure Ollama is running:
-• Check if 'ollama serve' is running
-• Try again in a few seconds`;
-        }
-        if (message.includes('model') && message.includes('not found')) {
-            return `❌ Model not found
-
-The requested model is not available locally.
-
-To download models, run:
-• ollama pull llama3
-• ollama pull codellama
-• ollama pull mistral
-
+    if (message.includes("model") && message.includes("not found")) {
+      return `\u274C Model not found
+
+The requested model is not available locally.
+
+To download models, run:
+\u2022 ollama pull llama3
+\u2022 ollama pull codellama
+\u2022 ollama pull mistral
+
 Then try again.`;
-        }
-        return `❌ Ollama error: ${error.message}
-
-Please ensure Ollama is installed and running.
-Install from: https://ollama.ai`;
     }
+    return `\u274C Ollama error: ${error.message}
+
+Please ensure Ollama is installed and running.
+Install from: https://ollama.ai`;
+  }
 }
-// Export singleton instance
-export const ollamaService = new OllamaService();
-// Export class for testing with different URLs
-export { OllamaService };
+const ollamaService = new OllamaService();
+export {
+  OllamaService,
+  TOOL_CAPABLE_MODEL_FAMILIES,
+  ollamaService
+};
 //# sourceMappingURL=ollama-service.js.map