A library for using the OpenAI GPT-3 API in Node. It tries to resemble OpenAI's provided Python bindings as closely as possible but with the additional asynchronous ways of Node.
I am not affiliated with OpenAI.
All methods require that the api_key is set to something.
organization is optional for your organization ID, but if it's set to something, it'll be included in the header.
npm i openai-node
Completion takes the normal GPT-3 parameters. The only one that is actually required in this library is prompt, the rest have the default values shown below.
var openai = require("openai-node");
openai.api_key = "YOUR API KEY"; // required
openai.organization = "YOUR ORGANIZATION ID"; // optional
//Completion
openai.Completion.create({
model: "davinci",
prompt: "Original: She no went to the market.\nStandard American English:",
temperature: 1,
max_tokens: 64,
top_p: 1,
frequency_penalty: 0,
presence_penalty: 1,
n: 1,
stream: false,
logprobs: null,
echo: false,
best_of: 1,
stop: null,
}).then((response) => {
console.log(response);
//EXAMPLE OUTPUT: She didn't go to the market.
});Search takes the normal GPT-3 parameters. documents and query must not be empty. Davinci is chosen by default.
var openai = require("openai-node");
openai.api_key = "YOUR API KEY"; // required
openai.organization = "YOUR ORGANIZATION ID"; // optional
//Search
openai
.Engine("davinci")
.search({
documents: ["whitehouse", "school", "hospital"],
query: "president",
})
.then((response) => {
console.log(response);
});
/*
OUTPUT:
{
object: 'list',
data: [
{ object: 'search_result', document: 0, score: 563.445 },
{ object: 'search_result', document: 1, score: 155.26 },
{ object: 'search_result', document: 2, score: 241.302 }
],
model: 'davinci:2020-05-03'
}*/Engine List returns a list off all the engines GPT-3 has available. (api_key is still required)
openai.Engine.list().then((res) => console.log(res));Returns engine information. (api_key is still required)
openai.Engine.retrieve("ada").then((res) => console.log(res));
/*
OUTPUT:
{
id: 'ada',
object: 'engine',
created: null,
max_replicas: null,
owner: 'openai',
permissions: null,
ready: true,
ready_replicas: null,
replicas: null
}
*/Lists all available files that belong to your organization.
openai.File.list().then((res) => {
console.log(res);
});Sends a new file to the endpoint and returns back it's ID for use with Search and Classification. Learn more about files at OpenAI's documention. File must be a .jsonl (JSON Lines).
const fs = require("fs");
openai.File.create({
purpose: "answers",
file: fs.createReadStream("./traning.jsonl"),
}).then((res) => {
console.log(res);
/* OUTPUT:
{
"id": "file-Twfy8T7YyIZR6jxPpIipbpxh",
"object": "file",
"bytes": 140,
"created_at": 1616240608,
"filename": "training.jsonl",
"purpose": "answers"
}
*/
});Removes file from your organization. OpenAI's Python library calls this function delete but that's a reserved word in JS.
openai
.File("file-Twfy8T7YyIZR6jxPpIipbpxh")
.remove()
.then((res) => {
console.log(res);
});
/* OUTPUT:
{
"id": "file-Twfy8T7YyIZR6jxPpIipbpxh",
"object": "file",
"deleted": true
}
*/Creates a new classification request. All parameters match OpenAI's. You can see them here
openai.Classification.create({
search_model: "ada",
model: "curie",
file: "file-XsOVjn4yoY2BRweTtcZB1P1B"
query: "It is a raining day :(",
labels: ["Positive", "Negative", "Neutral"],
}).then(res => {
console.log(res);
})Creates a new answer request. All parameters match OpenAI's. You can see them here
openai.Answer.create({
search_model: "ada",
mode: "curie",
question: "which puppy is happy?",
documents: ["Puppy A is happy.", "Puppy B is sad."],
examples_context: "In 2017, U.S. life expectancy was 78.6 years.",
examples: [["What is human life expectancy in the United States?", "78 years."]],
max_tokens: 5,
stop: ["\n", "<|endoftext|>"
}).then(res => {
console.log(res);
})