116 lines
3.8 KiB
JavaScript
116 lines
3.8 KiB
JavaScript
const aiMarkDrawing = async (req, res) => {
|
|
|
|
|
|
const express = require('express');
|
|
const fetch = require('node-fetch'); // For making HTTP requests
|
|
const path = require('path');
|
|
const fs = require('fs');
|
|
|
|
// Create an Express app
|
|
const app = express();
|
|
const port = 3000;
|
|
|
|
// Middleware to handle JSON and URL-encoded form data
|
|
app.use(express.json());
|
|
app.use(express.urlencoded({ extended: true }));
|
|
|
|
// Replace with your OpenAI API key
|
|
const OPENAI_API_KEY = 'your-openai-api-key';
|
|
|
|
// Route to handle image upload and forward it to OpenAI API
|
|
app.post('/upload-image', async (req, res) => {
|
|
try {
|
|
// Check if the file is included in the request
|
|
if (!req.headers['content-type']?.includes('multipart/form-data')) {
|
|
return res.status(400).send({ message: 'No file uploaded' });
|
|
}
|
|
|
|
// Extract the raw body data (binary) from the request
|
|
let imageBuffer = Buffer.alloc(0);
|
|
|
|
// Collect the image data from the stream (raw body)
|
|
req.on('data', chunk => {
|
|
imageBuffer = Buffer.concat([imageBuffer, chunk]);
|
|
});
|
|
|
|
req.on('end', async () => {
|
|
// Forward the image to OpenAI's API for scoring
|
|
const score = await sendImageToOpenAI(imageBuffer);
|
|
|
|
// Respond with the score
|
|
res.json({ score: score });
|
|
});
|
|
|
|
} catch (error) {
|
|
console.error('Error processing the image:', error);
|
|
res.status(500).send({ message: 'Internal Server Error' });
|
|
}
|
|
});
|
|
|
|
// Function to send image to OpenAI API and get the response
|
|
async function sendImageToOpenAI(imageBuffer) {
|
|
try {
|
|
// Convert binary image data to base64 format
|
|
const imageBase64 = imageBuffer.toString('base64');
|
|
|
|
// Make a request to the OpenAI API using fetch
|
|
const response = await fetch('https://api.openai.com/v1/chat/completions', {
|
|
method: 'POST',
|
|
headers: {
|
|
'Authorization': `Bearer ${OPENAI_API_KEY}`,
|
|
'Content-Type': 'application/json',
|
|
},
|
|
body: JSON.stringify({
|
|
model: "gpt-4-vision", // Use the appropriate model
|
|
messages: [
|
|
{ role: "system", content: "You are a grading assistant for 2nd grade students." },
|
|
{ role: "user", content: "Evaluate this student's drawing or coloring." }
|
|
],
|
|
functions: {
|
|
function_call: {
|
|
name: "vision_function",
|
|
args: {
|
|
image: imageBase64 // Send the image in base64 format
|
|
}
|
|
}
|
|
}
|
|
})
|
|
});
|
|
|
|
// Parse the response
|
|
const data = await response.json();
|
|
|
|
// Extract the score or other relevant information from the response
|
|
const aiResponse = data.choices[0].message.content;
|
|
|
|
// You can parse this response and extract the score, for example:
|
|
const score = extractScoreFromResponse(aiResponse);
|
|
return score;
|
|
|
|
} catch (error) {
|
|
console.error('Error interacting with OpenAI API:', error);
|
|
throw error;
|
|
}
|
|
}
|
|
|
|
// Example function to extract a score from OpenAI's response (you need to define this based on the actual response structure)
|
|
function extractScoreFromResponse(responseText) {
|
|
// You can use a regex or logic to extract the score from OpenAI's response
|
|
const scoreMatch = responseText.match(/score:\s*(\d+)/i);
|
|
if (scoreMatch) {
|
|
return parseInt(scoreMatch[1], 10);
|
|
} else {
|
|
// If no score is found, return a default value
|
|
return 0;
|
|
}
|
|
}
|
|
|
|
// Start the server
|
|
app.listen(port, () => {
|
|
console.log(`Server running at http://localhost:${port}`);
|
|
});
|
|
|
|
|
|
}
|
|
|
|
module.exports = aiMarkDrawing; |