diff --git a/docs/how-tos/vs-code/datacoves-copilot/README.md b/docs/how-tos/vs-code/datacoves-copilot/README.md index a48145e..0a8797d 100644 --- a/docs/how-tos/vs-code/datacoves-copilot/README.md +++ b/docs/how-tos/vs-code/datacoves-copilot/README.md @@ -3,10 +3,10 @@ title: Datacoves Copilot sidebar_position: 85 --- # AI LLMs for Datacoves Copilot -Datacoves can integrate seamlessly with your existing ChatGPT or Azure Open AI LLMs. These how tos will go over configuration and usage of AI within Datacoves. +Datacoves can integrate seamlessly with your existing ChatGPT or Azure OpenAI LLMs. These how tos will go over configuration and usage of AI within Datacoves. ### Prereqs -- Have an existing LLM such as ChatGPT or [Azure Open AI](https://learn.microsoft.com/en-us/azure/ai-services/openai/assistants-quickstart?tabs=command-line%2Ckeyless%2Ctypescript-keyless&pivots=ai-foundry-portal) +- Have an existing LLM such as ChatGPT or [Azure OpenAI](https://learn.microsoft.com/en-us/azure/ai-services/openai/assistants-quickstart?tabs=command-line%2Ckeyless%2Ctypescript-keyless&pivots=ai-foundry-portal) - Have access to API and Endpoint url credentials. - Have `Admin` access to configure credentials in Datacoves diff --git a/docs/how-tos/vs-code/datacoves-copilot/v2.md b/docs/how-tos/vs-code/datacoves-copilot/v2.md index 7709f6a..c923db2 100644 --- a/docs/how-tos/vs-code/datacoves-copilot/v2.md +++ b/docs/how-tos/vs-code/datacoves-copilot/v2.md @@ -11,7 +11,8 @@ This section describes how to configure and use Datacoves Copilot v2, which come - DeepSeek - Google Gemini - OpenAI -- OpenAI Compatible (i.e. Azure) +- Azure OpenAI +- OpenAI Compatible - Open Router - xAI (Grok) @@ -24,6 +25,7 @@ import TabItem from '@theme/TabItem'; {label: 'Config', value: 'config'}, {label: 'Anthropic', value: 'anthropic'}, {label: 'OpenAI', value: 'openai'}, +{label: 'Azure OpenAI', value: 'azure'}, {label: 'OpenAI Compatible', value: 'openaicompatible'}, {label: 'Google Gemini', value: 'gemini'}, {label: 'More Providers', value: 'additional'}, @@ -219,6 +221,109 @@ Optimized GPT-4 models: Refer to the [OpenAI Models documentation](https://platform.openai.com/docs/models) for the most up-to-date list of models and capabilities. + + + +## Azure OpenAI LLM Provider + +Datacoves Copilot supports Azure OpenAI models through the OpenAI API compatible interface. + +Website: https://azure.microsoft.com/en-us/products/ai-services/openai-service + +### Secret value format + +```json +{ + "default": { + "apiProvider": "openai", + "openAiApiKey": "", + "openAiBaseUrl": "https://.cognitiveservices.azure.com/openai/deployments//chat/completions?api-version=", + "openAiModelId": "", + "openAiUseAzure": true, + "id": "default" + } +} +``` + +### Getting Azure OpenAI Credentials + +1. **Create Azure OpenAI Resource**: Go to Azure Portal and create an Azure OpenAI service resource +2. **Deploy a Model**: In Azure AI Foundry, deploy a model (e.g., gpt-4o, gpt-4.1, gpt-5) +3. **Get Endpoint**: Copy your endpoint URL from the resource overview +4. **Get API Key**: Navigate to "Keys and Endpoint" section and copy one of the API keys +5. **Get Deployment Name**: Use the deployment name you created (not the model name) + +### Endpoint URL Format + +Datacoves Copilot uses the **Chat Completions API**, so your `openAiBaseUrl` must use the chat completions path, including the deployment name and `api-version` parameter: + +``` +https://.cognitiveservices.azure.com/openai/deployments//chat/completions?api-version= +``` + +For models that support **both** the Responses API and the Chat Completions API (for example GPT-4.1, GPT-5.x, o3, o4-mini), **always** use the Chat Completions URL above in Datacoves Copilot and **do not** use the default `/openai/v1/responses` URL shown in some Azure examples. + +### Supported Azure OpenAI Models (`openAiModelId`) + +Use your **deployment name** from Azure AI Foundry as the `openAiModelId`. This must match the deployment name exactly (for example `gpt-5.1` if your deployment is named `gpt-5.1`), not just the base model family name. + +#### GPT-5 Series (Latest) + +**Models with Chat Completions API support:** +- gpt-5.2 (2025-12-11) - Flagship model, 400K context +- gpt-5.2-chat (2025-12-11) - Chat optimized +- gpt-5.1 (2025-11-13) - Advanced reasoning, 400K context +- gpt-5.1-chat (2025-11-13) - Chat optimized reasoning +- gpt-5 (2025-08-07) - Advanced reasoning, 400K context +- gpt-5-mini (2025-08-07) - Cost-efficient, 400K context +- gpt-5-nano (2025-08-07) - Fast, cost-efficient, 400K context +- gpt-5-chat (2025-08-07, 2025-10-03) - Conversational, 128K context +- gpt-oss-120b - Open-weight reasoning model +- gpt-oss-20b - Open-weight reasoning model + +**Note:** The following GPT-5 models use Responses API only and are **not supported** by Datacoves Copilot: +- gpt-5-codex, gpt-5-pro, gpt-5.1-codex, gpt-5.1-codex-mini, gpt-5.1-codex-max + +#### GPT-4.1 Series + +- gpt-4.1 (2025-04-14) - Advanced multimodal, 1M context +- gpt-4.1-mini (2025-04-14) - Balanced performance, 1M context +- gpt-4.1-nano (2025-04-14) - Lightweight, 1M context + +#### GPT-4o Series + +- gpt-4o (2024-11-20) - Optimized GPT-4, 128K context +- gpt-4o (2024-08-06) - Optimized GPT-4, 128K context +- gpt-4o (2024-05-13) - Original GPT-4o, 128K context +- gpt-4o-mini (2024-07-18) - Fast, cost-efficient, 128K context + +#### GPT-4 Series + +- gpt-4 (turbo-2024-04-09) - GPT-4 Turbo with Vision, 128K context + +#### o-Series Reasoning Models + +- o3 (2025-04-16) - Reasoning model, 200K context +- o4-mini (2025-04-16) - Mini reasoning, 200K context +- o3-mini (2025-01-31) - Compact reasoning, 200K context +- o1 (2024-12-17) - Reasoning model, 200K context +- o1-mini (2024-09-12) - Smaller reasoning, 128K context +- codex-mini (2025-05-16) - Coding specialized, 200K context + +#### GPT-3.5 Series + +- gpt-35-turbo (0125) - Chat optimized, 16K context +- gpt-35-turbo (1106) - Chat optimized, 16K context +- gpt-35-turbo-instruct (0914) - Completions API only + +**Important Notes:** +- Datacoves Copilot uses the Chat Completions API endpoint +- Responses API is not currently supported for Azure OpenAI +- Use your **deployment name** in Azure as both the URL path segment and `openAiModelId` +- The `api-version` parameter is required in the endpoint URL + +Refer to [Azure OpenAI documentation](https://learn.microsoft.com/en-us/azure/ai-foundry/openai/concepts/models) for the most current model availability and regional deployment options. + @@ -229,9 +334,11 @@ Refer to the [OpenAI Models documentation](https://platform.openai.com/docs/mode Datacoves Copilot supports a wide range of AI model providers that offer APIs compatible with the OpenAI API standard. This means you can use models from providers other than OpenAI, while still using a familiar API interface. This includes providers like: - Local models running through tools like Ollama and LM Studio (covered in separate sections). -- Cloud providers like Azure, Perplexity, Together AI, Anyscale, and others. +- Cloud providers like Perplexity, Together AI, Anyscale, and others. - Any other provider offering an OpenAI-compatible API endpoint. +**Note:** For Azure OpenAI, see the dedicated [Azure OpenAI tab](#azure-openai) for specific setup instructions. + ### Secret value format ```json @@ -256,7 +363,7 @@ Where: 2. `openAiApiKey`: This is the secret key you obtain from the provider. 3. `openAiModelId`: This is the model name of the specific model, each provider will expose a different set of models, please check provider's documentation. -If you're using `Azure`, please set `"openAiUseAzure": true`, optionally you could specify a specific API version using `"azureApiVersion": ""`. + #### Fine tune model usage