CodeWhisperer Server

Guides

Proxy

To use the CodeWhisperer server behind a proxy, import the CodeWhispererServerTokenProxy as the server and set up the environment variable HTTPS_PROXY or https_proxy to the proxy URL you are using. You can pass the environment variable to the process or just set it up globally on your system.

export HTTPS_PROXY=https://proxy.example.com:5678
export https_proxy=https://proxy.example.com:5678

or

export HTTPS_PROXY=http://username:[email protected]:5678
export https_proxy=http://username:[email protected]:5678

Bundling with webworker runtime

Amazon Q Servers implemented in this package may have dependencies on NodeJS native modules, not available in non-Node.js environments.

To make cross-platform bundle, use bundler that allow overriding modules with compatible alternatives (e.g. Webpack).

NodeJS modules used in this package

• CodeWhispererServer:
  • path
• SecurityScanServer:
  • path, os

To override modules use next alternatives:

• path - https://www.npmjs.com/package/path-browserify
• os - https://www.npmjs.com/package/os-browserify

Fetching Amazon Q configuration(s)

The following Amazon Q configurations can be fetched with the aws/getConfigurationFromServer request exposed by QConfigurationServer:

• customizations
• developer profiles

The request expects a section parameter, recognizing the following options:

• aws.q
  • aws.q.customizations
  • aws.q.developerProfiles

Example:

await languageClient.sendRequest(getConfigurationFromServerRequestType.method, {
            section: 'aws.q',
        })
// result:
{
  'customizations': [customization1, customization2, ...],
  'developerProfiles': [profile1, profile2, ...] // (if enabled)
}

Granular requests such as aws.q.customizations will only return that particular configuration.

Example:

await languageClient.sendRequest(getConfigurationFromServerRequestType.method, {
            section: 'aws.q.customizations',
        })
// result:
[customization1, customization2, ...]

Developer Profiles

By default, developer profiles are not fetched. To enable the fetching, the client needs to signal support for them at initialization in the InitializeParams.

Example:

const params: InitializeParams = {
  // ...
  aws: {
    // ...
    awsClientCapabilities: {
      q: {
        developerProfiles: true
      }
    }
  }
}

Supported workspace configurations

The server supports the following workspace configurations:

• aws.q.customization (type: string | undefined, default: undefined)
  • The customization ARN to be used in API requests, an empty string will be interpreted as undefined.
• aws.q.optOutTelemetry (type: OPTOUT | OPTIN, default: OPTIN)
  • This flag controls whether to opt-in or opt-out to telemetry.
• aws.q.inlineSuggestions.extraContext (type: string | undefined, default: undefined)
  • The extra context to be included for suggestions, an empty string will be interpreted as undefined. See below for more context.
• aws.q.inlineChat.extraContext (type: string | undefined, default: undefined)
  • The extra context to be included for inline chat, an empty string will be interpreted as undefined. See below for more context.
• aws.codeWhisperer.includeSuggestionsWithCodeReferences: (type: boolean, default: false)
  • This flag controls whether to include references with code suggestions.
• aws.codeWhisperer.shareCodeWhispererContentWithAWS: (type: boolean, default: false)
  • This flag controls whether to share Q content with AWS.

The client can signal updates to the workspace configuration with the DidChangeConfiguration notification.

Extra context for Q Inline Suggestions

In cases when the client runs in a specific environment that requires customized suggestions, the server supports a aws.q.inlineSuggestions.extraContext workspace configuration. This extra context will be passed to the left file content of the request in the beginning of the file.

Extra context for Q Inline Chat

In cases when the client runs in a specific environment that requires customized inline chat responses, the server supports a aws.q.inlineChat.extraContext workspace configuration. This extra context will be prepended to the document text of the request, similar to how inline suggestions work.