@coveo/platform-client

Official Coveo Cloud APIs JavaScript Client

The main goal of this package is to provide an easy to configure and straightforward way of querying Coveo Cloud APIs using JavaScript.

Getting started

Build

npm i
npm run build

Install

npm install @coveo/platform-client
# Typescript types are included in the package

Import

Note, this project is pure ESM, require won't work and CommonJS support is limited.

// using default import
import PlatformClient from '@coveo/platform-client';

// using named import
import {PlatformClient} from '@coveo/platform-client';

// using dynamic import, works in CJS ⚠ requires an async context ⚠
const {PlatformClient} = await import('@coveo/platform-client');

Configure

const platform = new PlatformClient({
    /* configuration options */
});

Use

platform.resource.action({
    /* action dependant options */
});

Every action returns a Promise object.

Example

const platform = new PlatformClient({
    organizationId: 'some-coveo-platform-organization-id',
    accessToken: () => 'your-coveo-platform-access-token-or-api-key',
});

// List the organization catalogs
const catalogs = await platform.catalog.list({page: 0, pageSize: 10});
doSometing(catalogs);

Compatibility

The platform-client package is built on top of the fetch API and is not entirely supported by all JavaScript runtime environments. Consequently, for Node, we recommend using Node 18 or undici to polyfill fetch globally

Node.js example

if (!global['fetch']) {
    global['fetch'] = require('undici').fetch;
}
const PlatformClient = require('@coveo/platform-client').default;

const coveoPlatform = new PlatformClient({
    /* options */
});

coveoPlatform.source
    .list()
    .then((res) => {
        console.log(JSON.stringify(res));
    })
    .catch((e) => {
        console.error(e);
    });

Global Configuration

It is also possible to provide global request settings for all requests made by the client; Simply provide the configuration in the globalRequestSettings property of the options. This can be useful for applying necessary headers or configurations for network requests. For example, you might include a proxy authorization header for requests routed through a proxy server using the Proxy-Authorization header:

Node.js example

if (!global['fetch']) {
    global['fetch'] = require('undici').fetch;
}
const PlatformClient = require('@coveo/platform-client').default;

const coveoPlatform = new PlatformClient({
    globalRequestSettings: {
        headers: {
            'Proxy-Authorization': 'Basic YWxhZGRpbjpvcGVuc2VzYW1l',
        },
    },
});

coveoPlatform.source
    .list()
    .then((res) => {
        console.log(JSON.stringify(res));
    })
    .catch((e) => {
        console.error(e);
    });

Documentation

This project is built using TypeScript and automatically generates relevant type declarations. Most IDEs with TypeScript integration will display those type declarations as autocompletions, so that you typically will not need to refer to external documentation. Hence, the decision has been made not to document any option, resource, or action, except for the main configuration options. For in-depth documentation on the APIs exposed by this package, please consult our official documentation portal.

Configuration Options

Option	Required	Default Value	Description
accessToken	yes	undefined	The access token or API key.
organizationId	yes	undefined	The unique identifier of the target organization.
environment	optional	'production'	The target environment. If one of following: 'development', 'stg' 'production', 'hipaa'; automatically targets the associated host.
host	optional	'https://platform.cloud.coveo.com'	The target host. Useful to target local hosts when testing.
serverlessHost	optional	'https://api.cloud.coveo.com'	The target host for serverless APIs.
responseHandlers	optional	[]	Custom server response handlers. See error handling section for detailed explanation.
region	optional	Region.US	The target region.

Error handling

Each request made by the platform-client, once resolved or rejected, gets processed by one (and only one) of the response handlers. Some very basic response handlers are used by default, but you can override their behavior by specifying your own in the responseHandlers configuration option. The order in which they are specified defines their priority. Meaning that the first handler of the array that can process the response is used to do so.

A response handler is defined as such:

interface ResponseHandler {
    canProcess(response: Response): boolean; // whether the handler should be used to process the response
    process<T>(response: Response): Promise<T>; // defines how the handler processes the response
}

Example

const MySuccessResponseHandler: ResponseHandler = {
    canProcess: (response: Response): boolean => response.ok;
    process: async <T>(response: Response): Promise<T> => {
        const data = await response.json();
        console.log(data);
        return data;
    };
}

Ad hoc requests

Sometimes, a specific platform-client configuration is required for a one-time task. In such use cases, we can leverage the withFeatures() method on the platform-client instance. This will prevent having to create a new PlatformClient instance for just one usage.

A feature is defined as a function that returns a set of new configuration options to be used for the next request.

type Feature = (currentOptions: PlatformClientOptions) => PlatformClientOptions;

Where currentOptions are the options currently used by the client instance to make requests.

Example

const europeRegion: Feature = (currentOptions) => ({
    ...currentOptions,
    region: Region.EU,
});

const europeOrganizations = await platform.withFeatures(europeRegion).organization.list();

Multiple features can be used at once, each feature will override the options returned by the previous one.

const notifyOnSuccess = (message) => (currentOptions) => ({
    ...currentOptions,
    responseHandlers: [
        {
            canProcess: (response) => response.ok,
            process: async (response) => {
                const result = await response.json();
                showSuccessToast(message);
                return result;
            },
        },
        ...currentOptions.responseHandlers,
    ],
});

const notifyOnError = (message) => (currentOptions) => ({
    ...currentOptions,
    responseHandlers: [
        ...currentOptions.responseHandlers,
        {
            canProcess: () => true,
            process: async (response) => {
                const error = await response.json();
                showErrorToast(message);
                throw error;
            },
        },
    ],
});

platform.withFeatures(notifyOnSuccess('It worked!'), notifyOnError('It failed!')).field.delete('unwanted-field-id');

Extending the client

In cases where you want to expand the standard capabilities of the platform client with internal or experimental resources, you can do so within the scope of your own project by extending the PlatformClient class.

import {API, PlatformClient, PlatformClientOptions, Resource} from '@coveo/platform-client';

class Something extends Resource {
    static baseUrl = `/rest/organizations/${API.orgPlaceholder}/something`;

    list() {
        this.api.get(Something.baseUrl);
    }
}

const experimentalResources: Array<{key: string; resource: typeof Resource}> = [
    {key: 'something', resource: Something},
];

class ExperimentalPlatformClient extends PlatformClient {
    something: Something;

    constructor(options: PlatformClientOptions) {
        super(options);

        experimentalResources.forEach(({key, resource}) => {
            this[key] = new resource(this.API, this.ServerlessAPI);
        });
    }
}

// usage
ExperimentalPlatformClient.something.list();

Contributing

See CONTRIBUTING.md