Metad AI Copilot

English | 中文

@metad/copilot is a general-purpose abstract core logic package designed specifically for developing AI Copilot applications. It provides a flexible set of tools and features that enable you to quickly build and integrate Copilot functionality without having to focus on specific user interface frameworks.

Installation

Install @metad/copilot via npm:

npm install @metad/copilot

Usage in Angular

@metad/ocap-angular is a UI component library designed for the Angular framework. Among them, @metad/ocap-angular/copilot is the Copilot chat UI component, built on the foundation of @metad/copilot, aiming to assist users in seamlessly integrating and building Copilot chat functionality in Angular applications.

UI Installation

Install @metad/ocap-angular via npm:

npm install @metad/ocap-angular

Configuration

Before using the Copilot component, you need to provide configuration parameters. You can do this by using the provideClientCopilot function and passing in the AI API parameters.

import { provideClientCopilot } from '@metad/ocap-angular/copilot'
import { provideMarkdown } from 'ngx-markdown'

export const appConfig: ApplicationConfig = {
  providers: [
    provideClientCopilot(async () =>
      ({
        enabled: true,
        chatUrl: '',
        modelsUrl: '',
        apiKey: 'sk-xxxxxxx',
        provider: "openai",
      })
    ),
    provideMarkdown()
  ],
};

Usage in Components

Introduce the @metad/ocap-angular/copilot module into your Angular application:

import {
  NgmCopilotChatComponent,
  injectCopilotCommand,
  injectMakeCopilotActionable,
} from '@metad/ocap-angular/copilot';

@Component({
  selector: 'app-root',
  standalone: true,
  imports: [CommonModule, RouterOutlet, NgmCopilotChatComponent],
})
export class AppComponent { }

Use the Copilot Chat component in your component template:

<ngm-copilot-chat></ngm-copilot-chat>

This way, the basic chat functionality of Copilot Chat can be used.

Custom Commands

Two functions are also provided for customizing commands to perform specific operations:

• injectCopilotCommand Inject custom commands
• injectMakeCopilotActionable Inject the callable function of the custom command

injectCopilotCommand

| Property | Example | Description | |------|------|------| | name | 'form' | The unique identifier of the custom command, which is used to identify the command in front of the Copilot dialog prompt | | description | 'Descripe how to fill the form' | Describes how to use this command, prompting the user how to write the prompt | | examples | ['A', 'B'] | Examples of how to write the prompt for this command, which helps the user to quickly enter the prompt | | actions | [] | The operation function implementation available for this command, if not specified, all operation functions in the current context will be used |

injectMakeCopilotActionable

| Property | Example | Description | |------|------|------| | name | 'fill_form' | The unique identifier of the operation function, which is used by the AI to identify the call to this function | | description | 'Fill the form' | Describes the function | | argumentAnnotations | [] | The definition of the input parameters of the function | | implementation | async function | The implementation logic of the function, returning no value or returning a string will end the command session, returning Message will continue the command session |

argumentAnnotations:

| Property | Example | Description | |------|------|------| | name | 'form' | The name of the input parameter | | type | 'string' | The type of the parameter | | description | 'The form to fill' | The description of the parameter | | required | true | Whether the parameter is required | | properties | [] | The property definition of the input parameter of the function |

For the properties attribute, you can define it directly or use the zod library for definition, for example:

import { z, ZodType, ZodTypeDef } from 'zod'
import zodToJsonSchema from 'zod-to-json-schema'

{
    properties: (<{ properties: any }>zodToJsonSchema(z.object({
                title: z.string().describe('The title of form'),
                desc: z.string().description('My Milestone Work Objectives'),
                standard: z.string().description('My Satisfaction Measurement Standard'),
              }))).properties
}

Examples

Customize a command that can provide suggestion links:

#askCommand = injectCopilotCommand({
    name: 'help',
    description: 'Help',
    examples: [
      'Show helps for angular',
      'Show helps for react',
      'Show helps for vue',
    ],
    actions: [
      injectMakeCopilotActionable({
        name: 'show-help',
        description: 'Show helps',
        argumentAnnotations: [
          {
            name: 'helps',
            description: 'Helps for the commands',
            type: 'array',
            items: {
              type: 'object',
              properties: {
                title: {
                  type: 'string',
                  description: 'The title of help link',
                },
                link: {
                  type: 'string',
                  description: 'The url of help link',
                },
              },
            } as any,
            required: true,
          },
        ],
        implementation: async (helps: Link[]) => {
          this.helps.set(helps)
          return `Command finish!`
        },
      }),
    ],
  });

The Copilot framework provides a demo project that you can refer to for its implementation:

https://stackblitz.com/~/github.com/tiven-w/metad-copilot-demo

Contributing

If you find any issues or have suggestions for improvement, please feel free to raise issues or submit pull requests. We welcome and encourage contributions from the community.

License

@metad/copilot is released under the MIT license. For more information, please refer to the LICENSE file.

Contact Us

If you have any questions or need further assistance, please feel free to contact us via email at: mailto:[email protected]

Thank you for using @metad/copilot! We look forward to seeing your Copilot application succeed.