Eric Ciarla
76 lines
2.3 KiB
TypeScript
76 lines
2.3 KiB
TypeScript
import OpenAI from "openai/index.mjs";
|
|
import { encoding_for_model } from "@dqbd/tiktoken";
|
|
|
|
/**
|
|
* Function to generate a roast for a website based on its screenshot and markdown content.
|
|
* @param roastPrompt - Initial prompt text for the roast.
|
|
* @param screenshotUrl - URL of the screenshot of the website.
|
|
* @param content - Raw markdown content of the website.
|
|
*/
|
|
export async function roastPrompt(roastPrompt: string, screenshotUrl: string, content: string) {
|
|
try {
|
|
// Initialize OpenAI with the API key from environment variables
|
|
const openai = new OpenAI({
|
|
apiKey: process.env.OPEN_AI_KEY
|
|
});
|
|
|
|
|
|
let contentTruncated = await truncateContentToFit(content, 30000);
|
|
|
|
const response = await openai.chat.completions.create({
|
|
model: "gpt-4o",
|
|
messages: [
|
|
{
|
|
role: "user",
|
|
content: [
|
|
{ type: "text", text: roastPrompt },
|
|
{ type: "image_url", image_url: { url: screenshotUrl, detail: "low" } },
|
|
{ type: "text", text: contentTruncated }
|
|
],
|
|
},
|
|
],
|
|
});
|
|
// Return the first choice's message instead of logging it
|
|
return response.choices[0].message.content;
|
|
} catch (error) {
|
|
console.error("Error generating roast:", error);
|
|
// Assert error as an instance of Error to access message property
|
|
return `Error generating roast: ${(error as Error).message}`;
|
|
}
|
|
}
|
|
|
|
|
|
|
|
export function numTokensFromString(message: string, model: string): number {
|
|
const encoder = encoding_for_model(model as any);
|
|
|
|
// Encode the message into tokens
|
|
const tokens = encoder.encode(message);
|
|
|
|
// Free the encoder resources after use
|
|
encoder.free();
|
|
|
|
// Return the number of tokens
|
|
return tokens.length;
|
|
}
|
|
|
|
|
|
|
|
async function truncateContentToFit(content: string, maxTokens: number): Promise<string> {
|
|
const modifier = 4;
|
|
|
|
let contentTotruncate = content;
|
|
const numTokens = numTokensFromString(contentTotruncate, "gpt-4");
|
|
|
|
if (numTokens > maxTokens) {
|
|
// trim the document to the maximum number of tokens, tokens != characters
|
|
contentTotruncate = content.slice(0, (maxTokens * modifier));
|
|
|
|
}
|
|
return contentTotruncate
|
|
}
|
|
|
|
|
|
|
|
|