AI Assistant

Click New Setting. Enter the parameters and click Save to complete. All users can create and manage their own settings.

Setting Type

• System

  Admin users can set the Setting Type as System, which shares the setting across all users.

• Private

  All users can create Private settings for themselves and set any setting as default. The default setting will be automatically activated when the chat starts.

Note that the Agent supports streaming APIs with chunks formatted as follows:

interface ChatCompletionChunk {
  id: string;
  choices: Array<{
    delta: {
      content?: string | null;
      // The reasoning content must be placed within this field.
      reasoning_content?: string | null;
      // Some platforms, like OpenRouter, use "reasoning" instead.
      reasoning?: string | null
      function_call?: {
        arguments?: string;
        name?: string;
      };
      refusal?: string | null;
      role?: 'system' | 'user' | 'assistant';

    };
    finish_reason: 'stop' | 'length' | 'tool_calls' | 'content_filter' | 'function_call' | null;
    index: number;
  }>;
  model: string;
  usage?: {
    completion_tokens: number;
    prompt_tokens: number;
    total_tokens: number;
  } | null;
}

Click the top-left corner of the chat box to open the dropdown menu and view available settings. The active setting is marked with a green dot. Click any setting to switch quickly.

Input Description

Switch to Factor mode via the dropdown in the top-right corner of the chat window. Enter the description of the desired factor. For example:

“How to calculate Alpha41 from the WorldQuant factors using DolphinDB? Use the formula: (((high * low)^0.5) - vwap).”

Generate Factor Code

After submission, the AI Assistant will process the question and return the code as a complete script.

Select and Test Run

Check the factor codes and select the one that meets the requirements. Click Next Step.

Click Test Run or Test Run All Factors to execute the factor code.

Create Factor Module

Click Create Factor Based on Script after the test run has succeeded to create a new factor module containing the selected factors.

Factor Module Editing and Test & Evaluation Configuration

The system will jump to Factor Module editing page. You can set the compute and analysis configurations in the chat window.

​​Execution and Results​

Click Run to execute factor calculation and evaluation after configuration. A pop-up window in the upper-right corner will display the current status. Click ​​View​​ to check the detailed information, including performance, run result​​, and visualization report​​.

You can try the following operations:

• Batch Generation: Generate multiple factors at once.
• Derivation: Give new instructions based on the existing factors. For example: “Derive 4 factors from this factor”. The AI will return 4 new factors. You can select from them for further use.