How should the header be configured in the generative AI custom LLM Transformer?
Options
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
2 weeks ago
I configured a generic large language model (LLM) connector, but for the Authorization part in the header, I have to either set it as a fixed value or use a script to query. I believe that after the authentication has been configured, there should no longer be a need to query the apikey in the AI custom LLM Transformer. Could you please tell me exactly how to configure it so that I don't need to use a script to obtain the API key?
my generative AI custom LLM Transformer request transformer :
(function(inputs) {
/* write code here to construct the request body and any custom headers needed using the inputs object.
inputs structure: {
prompt_data: object,
request_data: object
} */
var requestData = inputs.request_data;
var promptData = inputs.prompt_data;
var prompt = promptData.prompt;
var model = promptData.model;
// construct body using the inputs
var body = {
"model": model,
"messages": [{
"role": "user",
"content": prompt
}],
"stream": false
};
var apikey = "Bearer ";
var provider = new sn_cc.ConnectionInfoProvider();
// Replace 'MyExternalServiceAPI' with the actual ID of your Connection & Credential Alias
var connectionInfo = provider.getConnectionInfo('XXXXXX');
if (connectionInfo) {
// Access the API key securely
apiKey = apikey + connectionInfo.getCredentialAttribute('api_key');
var headers = {
"Authorization": apiKey,
"Content-Type": "application/json"
};
return {
body: body,
headers: headers
};
} else {
gs.error('Connection & Credential Alias not found or invalid.');
}
})(inputs);
0 REPLIES 0