osagent 0.2.78 → 0.2.79

package/cli.js

@@ -133087,7 +133087,7 @@ var init_geminiContentGenerator = __esm({
       }
       models;
       constructor(contentGeneratorConfig, config2) {
-        const version3 = "0.2.77";
+        const version3 = "0.2.79";
         const userAgent2 = `OSAgent/${version3} (${process.platform}; ${process.arch})`;
         let headers = {
           "User-Agent": userAgent2
@@ -146884,13 +146884,295 @@ var init_openaiContentGenerator2 = __esm({
   }
 });
 
+// packages/core/dist/src/core/ollamaCloudContentGenerator/nativeOllamaCloudGenerator.js
+function createNativeOllamaCloudGenerator(contentGeneratorConfig, cliConfig) {
+  return new NativeOllamaCloudGenerator(contentGeneratorConfig, cliConfig);
+}
+var OLLAMA_CLOUD_API_URL, NativeOllamaCloudGenerator;
+var init_nativeOllamaCloudGenerator = __esm({
+  "packages/core/dist/src/core/ollamaCloudContentGenerator/nativeOllamaCloudGenerator.js"() {
+    "use strict";
+    init_esbuild_shims();
+    init_request_tokenizer();
+    init_uiTelemetry();
+    init_converter();
+    OLLAMA_CLOUD_API_URL = "https://ollama.com/api/chat";
+    NativeOllamaCloudGenerator = class {
+      static {
+        __name(this, "NativeOllamaCloudGenerator");
+      }
+      apiKey;
+      model;
+      baseUrl;
+      timeout;
+      cliConfig;
+      constructor(contentGeneratorConfig, cliConfig) {
+        this.apiKey = contentGeneratorConfig.apiKey || "";
+        this.model = contentGeneratorConfig.model || "qwen3-coder:480b-cloud";
+        this.baseUrl = OLLAMA_CLOUD_API_URL;
+        this.timeout = contentGeneratorConfig.timeout || 12e4;
+        this.cliConfig = cliConfig;
+      }
+      /**
+       * Build headers for Ollama Cloud API
+       */
+      buildHeaders() {
+        const version3 = this.cliConfig.getCliVersion() || "unknown";
+        return {
+          "Authorization": `Bearer ${this.apiKey}`,
+          "Content-Type": "application/json",
+          "User-Agent": `OSAgent/${version3} (${process.platform}; ${process.arch})`
+        };
+      }
+      /**
+       * Convert Gemini-format content to Ollama messages
+       */
+      convertToOllamaMessages(contents) {
+        const messages = [];
+        for (const content of contents) {
+          const role = content.role === "model" ? "assistant" : content.role;
+          let textContent2 = "";
+          const images = [];
+          for (const part of content.parts || []) {
+            if ("text" in part && part.text) {
+              textContent2 += part.text;
+            } else if ("inlineData" in part && part.inlineData?.data) {
+              images.push(part.inlineData.data);
+            }
+          }
+          if (textContent2 || images.length > 0) {
+            const message2 = { role, content: textContent2 };
+            if (images.length > 0) {
+              message2.images = images;
+            }
+            messages.push(message2);
+          }
+        }
+        return messages;
+      }
+      /**
+       * Convert Ollama response to Gemini format
+       */
+      convertToGeminiResponse(ollamaResponse) {
+        const parts = [];
+        if (ollamaResponse.message?.content) {
+          parts.push({ text: ollamaResponse.message.content });
+        }
+        return {
+          candidates: [
+            {
+              content: {
+                role: "model",
+                parts
+              },
+              finishReason: ollamaResponse.done ? "STOP" : void 0,
+              index: 0
+            }
+          ],
+          usageMetadata: {
+            promptTokenCount: ollamaResponse.prompt_eval_count,
+            candidatesTokenCount: ollamaResponse.eval_count,
+            totalTokenCount: (ollamaResponse.prompt_eval_count || 0) + (ollamaResponse.eval_count || 0)
+          }
+        };
+      }
+      /**
+       * Generate content (non-streaming)
+       */
+      async generateContent(request4, _userPromptId) {
+        const normalizedContents = request4.contents ? toContents(request4.contents) : [];
+        const messages = this.convertToOllamaMessages(normalizedContents);
+        if (request4.config?.systemInstruction) {
+          const systemParts = request4.config.systemInstruction.parts || [];
+          const systemText = systemParts.map((p) => "text" in p ? p.text : "").join("");
+          if (systemText) {
+            messages.unshift({ role: "system", content: systemText });
+          }
+        }
+        const ollamaRequest = {
+          model: this.model,
+          messages,
+          stream: false,
+          options: {
+            temperature: request4.config?.temperature,
+            top_p: request4.config?.topP,
+            top_k: request4.config?.topK,
+            num_predict: request4.config?.maxOutputTokens
+          }
+        };
+        const controller = new AbortController();
+        const timeoutId = setTimeout(() => controller.abort(), this.timeout);
+        try {
+          const response = await fetch(this.baseUrl, {
+            method: "POST",
+            headers: this.buildHeaders(),
+            body: JSON.stringify(ollamaRequest),
+            signal: request4.config?.abortSignal || controller.signal
+          });
+          clearTimeout(timeoutId);
+          if (!response.ok) {
+            const errorText = await response.text();
+            throw new Error(`Ollama Cloud API error ${response.status}: ${errorText}`);
+          }
+          const ollamaResponse = await response.json();
+          const geminiResponse = this.convertToGeminiResponse(ollamaResponse);
+          if (geminiResponse.usageMetadata?.promptTokenCount !== void 0) {
+            uiTelemetryService.setLastPromptTokenCount(geminiResponse.usageMetadata.promptTokenCount);
+          }
+          return geminiResponse;
+        } catch (error2) {
+          clearTimeout(timeoutId);
+          if (error2 instanceof Error && error2.name === "AbortError") {
+            throw new Error("Request timed out");
+          }
+          throw error2;
+        }
+      }
+      /**
+       * Generate content with streaming
+       */
+      async generateContentStream(request4, _userPromptId) {
+        const normalizedContents = request4.contents ? toContents(request4.contents) : [];
+        const messages = this.convertToOllamaMessages(normalizedContents);
+        if (request4.config?.systemInstruction) {
+          const systemParts = request4.config.systemInstruction.parts || [];
+          const systemText = systemParts.map((p) => "text" in p ? p.text : "").join("");
+          if (systemText) {
+            messages.unshift({ role: "system", content: systemText });
+          }
+        }
+        const ollamaRequest = {
+          model: this.model,
+          messages,
+          stream: true,
+          options: {
+            temperature: request4.config?.temperature,
+            top_p: request4.config?.topP,
+            top_k: request4.config?.topK,
+            num_predict: request4.config?.maxOutputTokens
+          }
+        };
+        const response = await fetch(this.baseUrl, {
+          method: "POST",
+          headers: this.buildHeaders(),
+          body: JSON.stringify(ollamaRequest),
+          signal: request4.config?.abortSignal
+        });
+        if (!response.ok) {
+          const errorText = await response.text();
+          throw new Error(`Ollama Cloud API error ${response.status}: ${errorText}`);
+        }
+        if (!response.body) {
+          throw new Error("No response body for streaming");
+        }
+        return async function* () {
+          const reader = response.body.getReader();
+          const decoder2 = new TextDecoder();
+          let buffer = "";
+          try {
+            while (true) {
+              const { done, value } = await reader.read();
+              if (done)
+                break;
+              buffer += decoder2.decode(value, { stream: true });
+              const lines = buffer.split("\n");
+              buffer = lines.pop() || "";
+              for (const line of lines) {
+                if (!line.trim())
+                  continue;
+                try {
+                  const chunk = JSON.parse(line);
+                  const parts = [];
+                  if (chunk.message?.content) {
+                    parts.push({ text: chunk.message.content });
+                  }
+                  const geminiChunk = {
+                    candidates: [
+                      {
+                        content: {
+                          role: "model",
+                          parts
+                        },
+                        finishReason: chunk.done ? "STOP" : void 0,
+                        index: 0
+                      }
+                    ],
+                    usageMetadata: chunk.done ? {
+                      promptTokenCount: chunk.prompt_eval_count,
+                      candidatesTokenCount: chunk.eval_count,
+                      totalTokenCount: (chunk.prompt_eval_count || 0) + (chunk.eval_count || 0)
+                    } : void 0
+                  };
+                  if (chunk.done && chunk.prompt_eval_count !== void 0) {
+                    uiTelemetryService.setLastPromptTokenCount(chunk.prompt_eval_count);
+                  }
+                  yield geminiChunk;
+                } catch {
+                  continue;
+                }
+              }
+            }
+            if (buffer.trim()) {
+              try {
+                const chunk = JSON.parse(buffer);
+                const parts = [];
+                if (chunk.message?.content) {
+                  parts.push({ text: chunk.message.content });
+                }
+                yield {
+                  candidates: [
+                    {
+                      content: {
+                        role: "model",
+                        parts
+                      },
+                      finishReason: chunk.done ? "STOP" : void 0,
+                      index: 0
+                    }
+                  ]
+                };
+              } catch {
+              }
+            }
+          } finally {
+            reader.releaseLock();
+          }
+        }();
+      }
+      /**
+       * Count tokens in request
+       */
+      async countTokens(request4) {
+        try {
+          const tokenizer = getDefaultTokenizer();
+          const result = await tokenizer.calculateTokens(request4, {
+            textEncoding: "cl100k_base"
+          });
+          return { totalTokens: result.totalTokens };
+        } catch {
+          const content = JSON.stringify(request4.contents);
+          return { totalTokens: Math.ceil(content.length / 4) };
+        }
+      }
+      /**
+       * Embed content (not supported by Ollama Cloud chat API)
+       */
+      async embedContent(_request) {
+        throw new Error("Embedding not supported by Ollama Cloud chat API");
+      }
+    };
+    __name(createNativeOllamaCloudGenerator, "createNativeOllamaCloudGenerator");
+  }
+});
+
 // packages/core/dist/src/core/ollamaCloudContentGenerator/constants.js
-var OLLAMA_CLOUD_BASE_URL, DEFAULT_OLLAMA_CLOUD_MODEL, DEFAULT_TIMEOUT4, DEFAULT_MAX_RETRIES4;
+var OLLAMA_CLOUD_BASE_URL, OLLAMA_CLOUD_CHAT_URL, DEFAULT_OLLAMA_CLOUD_MODEL, DEFAULT_TIMEOUT4, DEFAULT_MAX_RETRIES4;
 var init_constants7 = __esm({
   "packages/core/dist/src/core/ollamaCloudContentGenerator/constants.js"() {
     "use strict";
     init_esbuild_shims();
-    OLLAMA_CLOUD_BASE_URL = "https://ollama.com/v1";
+    OLLAMA_CLOUD_BASE_URL = "https://ollama.com";
+    OLLAMA_CLOUD_CHAT_URL = "https://ollama.com/api/chat";
     DEFAULT_OLLAMA_CLOUD_MODEL = "qwen2.5-coder:32b";
     DEFAULT_TIMEOUT4 = 12e4;
     DEFAULT_MAX_RETRIES4 = 3;
@@ -146965,31 +147247,39 @@ __export(ollamaCloudContentGenerator_exports, {
   DEFAULT_MAX_RETRIES: () => DEFAULT_MAX_RETRIES4,
   DEFAULT_OLLAMA_CLOUD_MODEL: () => DEFAULT_OLLAMA_CLOUD_MODEL,
   DEFAULT_TIMEOUT: () => DEFAULT_TIMEOUT4,
+  NativeOllamaCloudGenerator: () => NativeOllamaCloudGenerator,
   OLLAMA_CLOUD_BASE_URL: () => OLLAMA_CLOUD_BASE_URL,
+  OLLAMA_CLOUD_CHAT_URL: () => OLLAMA_CLOUD_CHAT_URL,
   OllamaCloudProvider: () => OllamaCloudProvider,
+  createNativeOllamaCloudGenerator: () => createNativeOllamaCloudGenerator,
   createOllamaCloudContentGenerator: () => createOllamaCloudContentGenerator,
   isOllamaCloudConfig: () => isOllamaCloudConfig
 });
 function createOllamaCloudContentGenerator(contentGeneratorConfig, cliConfig, options2) {
-  const provider = new OllamaCloudProvider(contentGeneratorConfig, cliConfig);
+  const apiKey = contentGeneratorConfig.apiKey;
   if (options2?.validateCredentials) {
-    return provider.validateCredentials().then(() => {
-      return new OpenAIContentGenerator(contentGeneratorConfig, cliConfig, provider);
+    if (!apiKey) {
+      return Promise.reject(new Error("Ollama Cloud API key is required"));
+    }
+    return validateOllamaCloudCredentials(apiKey).then(() => {
+      return createNativeOllamaCloudGenerator(contentGeneratorConfig, cliConfig);
     });
   }
-  return new OpenAIContentGenerator(contentGeneratorConfig, cliConfig, provider);
+  return createNativeOllamaCloudGenerator(contentGeneratorConfig, cliConfig);
 }
 function isOllamaCloudConfig(config2) {
-  return OllamaCloudProvider.isOllamaCloudProvider(config2);
+  const baseUrl = config2.baseUrl?.toLowerCase() || "";
+  return baseUrl.includes("ollama.com");
 }
 var init_ollamaCloudContentGenerator = __esm({
   "packages/core/dist/src/core/ollamaCloudContentGenerator/index.js"() {
     "use strict";
     init_esbuild_shims();
-    init_openaiContentGenerator();
-    init_ollamaCloudProvider();
+    init_nativeOllamaCloudGenerator();
+    init_credentialValidator();
     init_ollamaCloudProvider();
     init_constants7();
+    init_nativeOllamaCloudGenerator();
     __name(createOllamaCloudContentGenerator, "createOllamaCloudContentGenerator");
     __name(isOllamaCloudConfig, "isOllamaCloudConfig");
   }
@@ -152955,8 +153245,8 @@ function createContentGeneratorConfig(config2, authType, generationConfig) {
     return {
       ...baseConfig,
       model: ollamaModel,
-      // Ollama Cloud uses OpenAI-compatible API at /v1 endpoint
-      baseUrl: "https://ollama.com/v1",
+      // Ollama Cloud uses native API at /api/chat endpoint
+      baseUrl: "https://ollama.com",
       apiKey: ollamaApiKey,
       authType: AuthType.OLLAMA_CLOUD,
       // Ollama doesn't need special cache control handling
@@ -153007,7 +153297,7 @@ function createContentGeneratorConfig(config2, authType, generationConfig) {
   };
 }
 async function createContentGenerator(config2, gcConfig, sessionId2, isInitialAuth) {
-  const version3 = "0.2.77";
+  const version3 = "0.2.79";
   const userAgent2 = `OSAgent/${version3} (${process.platform}; ${process.arch})`;
   const baseHeaders = {
     "User-Agent": userAgent2
@@ -340555,7 +340845,7 @@ __name(getPackageJson, "getPackageJson");
 // packages/cli/src/utils/version.ts
 async function getCliVersion() {
   const pkgJson = await getPackageJson();
-  return "0.2.77";
+  return "0.2.79";
 }
 __name(getCliVersion, "getCliVersion");
 
@@ -344756,8 +345046,8 @@ var formatDuration = /* @__PURE__ */ __name((milliseconds) => {
 
 // packages/cli/src/generated/git-commit.ts
 init_esbuild_shims();
-var GIT_COMMIT_INFO2 = "bc5e5a7";
-var CLI_VERSION2 = "0.2.77";
+var GIT_COMMIT_INFO2 = "7c5cf3f";
+var CLI_VERSION2 = "0.2.79";
 
 // packages/cli/src/utils/systemInfo.ts
 async function getNpmVersion() {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "osagent",
-  "version": "0.2.78",
+  "version": "0.2.79",
   "description": "OS Agent - AI-powered CLI for autonomous coding with Ollama Cloud and Qwen models",
   "repository": {
     "type": "git",