2023-11-23 16:48:59 +00:00
|
|
|
const https = require('https'); // Import for webscraping (fetchContentFromURL(url) function
|
2023-11-24 17:40:49 +00:00
|
|
|
import { OpenAIApi, Configuration } from 'openai';
|
2023-11-24 17:52:38 +00:00
|
|
|
import { fetch } from 'node-fetch';
|
2023-11-23 16:32:58 +00:00
|
|
|
|
2023-11-24 20:38:02 +00:00
|
|
|
// Function to fetch content from URL using a web scraping service
|
2023-11-22 11:41:01 +00:00
|
|
|
async function fetchContentFromURL(url) {
|
2023-11-24 20:38:02 +00:00
|
|
|
try {
|
|
|
|
const response = await fetch(url);
|
|
|
|
if (!response.ok) {
|
|
|
|
throw new Error(`HTTP error! status: ${response.status}`);
|
|
|
|
}
|
|
|
|
return await response.text();
|
|
|
|
} catch (error) {
|
|
|
|
console.error(`Could not fetch content from URL: ${error}`);
|
|
|
|
throw error;
|
|
|
|
}
|
2023-11-22 12:33:41 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
function simplifyContent(content) {
|
2023-11-24 14:34:13 +00:00
|
|
|
// Remove HTML tags
|
2023-11-24 14:05:17 +00:00
|
|
|
let simplifiedContent = content.replace(/<[^>]*>/g, '');
|
2023-11-24 14:34:13 +00:00
|
|
|
// Remove CSS within style tags
|
2023-11-24 14:05:17 +00:00
|
|
|
simplifiedContent = simplifiedContent.replace(/<style[^>]*>.*<\/style>/gms, '');
|
2023-11-24 14:34:13 +00:00
|
|
|
// Remove inline CSS and JavaScript within script tags
|
|
|
|
simplifiedContent = simplifiedContent.replace(/<script[^>]*>.*<\/script>/gms, '');
|
|
|
|
// Remove special characters and HTML entities
|
|
|
|
simplifiedContent = simplifiedContent.replace(/[^\w\s]/gi, '').replace(/&[a-z]+;/gi, '');
|
|
|
|
// Remove URLs
|
2023-11-24 14:26:49 +00:00
|
|
|
simplifiedContent = simplifiedContent.replace(/https?:\/\/[^\s]+/gi, '');
|
2023-11-24 14:34:13 +00:00
|
|
|
// Normalize whitespace
|
2023-11-24 14:05:17 +00:00
|
|
|
simplifiedContent = simplifiedContent.replace(/\s+/g, ' ').trim();
|
2023-11-24 14:34:13 +00:00
|
|
|
// Basic language simplification
|
2023-11-24 14:26:49 +00:00
|
|
|
simplifiedContent = simplifiedContent.toLowerCase();
|
2023-11-24 15:39:31 +00:00
|
|
|
// // Simple summarization: taking the first few sentences
|
|
|
|
// const sentences = simplifiedContent.split('. ');
|
|
|
|
// const summarizedContent = sentences.slice(0, Math.min(5, sentences.length)).join('. ');
|
2023-11-24 15:40:13 +00:00
|
|
|
return simplifiedContent;
|
2023-11-22 12:33:41 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
// Placeholder function to perform GPT analysis for media type and topics using Mistral-7b via OpenRouter
|
|
|
|
async function performGPTAnalysis(content) {
|
2023-11-22 11:41:01 +00:00
|
|
|
// Implement logic to send content to Mistral-7b via OpenRouter for GPT analysis
|
|
|
|
// Send content and receive GPT analysis response
|
|
|
|
// Placeholder code
|
|
|
|
const inferredMediaType = "article";
|
|
|
|
const extractedTopics = ["topic1", "topic2"];
|
|
|
|
return { inferredMediaType, extractedTopics };
|
2023-11-22 12:33:41 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
// Placeholder function to map inferred values to predefined formats and topics
|
|
|
|
function mapInferredValues(mediaType, topics) {
|
2023-11-22 11:41:01 +00:00
|
|
|
// Implement logic to map inferred media type and topics to predefined formats and topics
|
|
|
|
// Match inferred values with predefined taxonomy
|
|
|
|
// Placeholder code
|
|
|
|
const predefinedMediaType = "Article";
|
|
|
|
const predefinedTopics = ["Topic 1", "Topic 2"];
|
|
|
|
return { predefinedMediaType, predefinedTopics };
|
2023-11-22 12:33:41 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
// Placeholder function to format the response
|
|
|
|
function formatResponse(predefinedMediaType, predefinedTopics) {
|
2023-11-22 11:41:01 +00:00
|
|
|
// Implement logic to format the extracted metadata into the desired response structure
|
|
|
|
// Construct the response object
|
|
|
|
// Placeholder code
|
|
|
|
const response = {
|
2023-11-22 20:44:21 +00:00
|
|
|
format: predefinedMediaType,
|
2023-11-23 15:42:06 +00:00
|
|
|
topics: predefinedTopics,
|
2023-11-22 11:41:01 +00:00
|
|
|
};
|
|
|
|
return response;
|
2023-11-22 12:33:41 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
export async function handler(event) {
|
2023-11-22 11:41:01 +00:00
|
|
|
try {
|
2023-11-22 12:33:41 +00:00
|
|
|
// Extract URL and API Key from the request body
|
|
|
|
const { url, apiKey } = JSON.parse(event.body);
|
|
|
|
|
|
|
|
// Validate if URL and API Key are present
|
|
|
|
if (!url || !apiKey) {
|
|
|
|
return {
|
|
|
|
statusCode: 400,
|
|
|
|
body: JSON.stringify({ error: 'URL and API Key are required' }),
|
|
|
|
};
|
|
|
|
}
|
|
|
|
|
|
|
|
// Step 1: Fetch content from the URL using a web scraping service
|
|
|
|
const fetchedContent = await fetchContentFromURL(url);
|
|
|
|
|
|
|
|
// Step 2: Simplify the fetched content for GPT analysis
|
|
|
|
const simplifiedContent = simplifyContent(fetchedContent);
|
|
|
|
|
|
|
|
// Step 3: Perform GPT analysis for media type and topics
|
|
|
|
const { inferredMediaType, extractedTopics } = await performGPTAnalysis(simplifiedContent);
|
|
|
|
|
|
|
|
// Step 4: Map inferred values to predefined formats and topics
|
|
|
|
const { predefinedMediaType, predefinedTopics } = mapInferredValues(inferredMediaType, extractedTopics);
|
|
|
|
|
|
|
|
// Step 5: Format the response
|
|
|
|
const formattedResponse = formatResponse(predefinedMediaType, predefinedTopics);
|
|
|
|
|
|
|
|
// Return the formatted response
|
|
|
|
return {
|
|
|
|
statusCode: 200,
|
2023-11-24 20:35:00 +00:00
|
|
|
body: JSON.stringify(fetchedContent),
|
2023-11-22 12:33:41 +00:00
|
|
|
};
|
2023-11-22 11:41:01 +00:00
|
|
|
} catch (error) {
|
2023-11-22 12:33:41 +00:00
|
|
|
return {
|
|
|
|
statusCode: 500,
|
|
|
|
body: JSON.stringify({ error: 'Something went wrong' }),
|
|
|
|
};
|
2023-11-22 11:41:01 +00:00
|
|
|
}
|
2023-11-23 16:26:25 +00:00
|
|
|
}
|