production.php

<?php

require 'vendor/autoload.php';
require 'mindee.php';

use Smalot\PdfParser\Parser;
use Dotenv\Dotenv;
use OpenAI\Client;

// Initialize Dotenv and load .env file
$dotenv = Dotenv::createImmutable(__DIR__);
$dotenv->load();

// Retrieve the OPENAI_API_KEY from environment
$openaiKey = $_ENV['OPENAI_API_KEY'];
$mindeeKey = $_ENV['MINDEE_API_KEY'];

header('Content-type: application/json');

function ExtractTextFromCopyablePDF($inputFile)
{
    // Initialize the parser
    $parser = new Parser();

    // Parse the PDF file
    $pdf = $parser->parseFile($inputFile);

    // Extract the text from the document
    $text = $pdf->getText();

    // Return the extracted text
    return $text;
}


function OpenAIFunctionCalling($extractedText, $openaiKey, $outputFile, $eurovalue = null)
{
	error_log("OpenAIFunctionCalling())");
    $client = OpenAI::client($openaiKey);
    $prompt = "Extract the emitter and recipient company information, invoice number and invoice date are date and order number, currency, and line items.\nSepatrate vat country code if present at the beginning and vat number.\n\nFrom the following invoice text:\n\n" . $extractedText;

	$description = 'Extracts emitter and recipient company information, invoice number, date, currency, and line items from an invoice. Sepatrate vat country code if present at the beginning and vat number. Add to the row description the service period, if any. Split emitter address into address, city, zip code and country. Transform country name in international country code. Convert any utf8 charachet to plan and convert any currency symbol to currency abbreviation (EUR, USD, ...). Transform all dated to forma yyyy-mm-dd. BluCloud if present it the the recipient of the invoice.';
    $functions = [
        [
            'name' => 'extract_invoice_data',
            'description' => $description,
            'parameters' => [
                'type' => 'object',
                'properties' => [
                    'emitter' => [
                        'type' => 'object',
                        'properties' => [
                            'name' => ['type' => 'string'],
                            'address' => ['type' => 'string'],
							'city' => ['type' => 'string'],
							'zip_code' => ['type' => 'string'],
							'country' => ['type' => 'string'],
							'vat_country' => ['type' => 'string'],
                            'vat_number' => ['type' => 'string'],
                            'contact' => [
                                'type' => 'object',
                                'properties' => [
                                    'phone' => ['type' => 'string'],
                                    'email' => ['type' => 'string'],
                                ],
                            ],
                        ],
                    ],
                    'recipient' => [
                        'type' => 'object',
                        'properties' => [
                            'name' => ['type' => 'string'],
                            'address' => ['type' => 'string'],
							'vat_country' => ['type' => 'string'],
                            'vat_number' => ['type' => 'string'],
                            'contact' => [
                                'type' => 'object',
                                'properties' => [
                                    'phone' => ['type' => 'string'],
                                    'email' => ['type' => 'string'],
                                ],
                            ],
                        ],
                    ],
                    'invoice' => [
                        'type' => 'object',
                        'properties' => [
                            'number' => ['type' => 'string'],
                            'date' => ['type' => 'string', 'format' => 'date', 'description' => 'Invoice date in YYYY-MM-DD format',],
                            'currency' => ['type' => 'string'],
                        ],
                    ],
                    'invoice_lines' => [
                        'type' => 'array',
                        'items' => [
                            'type' => 'object',
                            'properties' => [
                                'quantity' => ['type' => 'integer'],
                                'description' => ['type' => 'string'],
                                'unit_price' => ['type' => 'number'],
								'vat_rate' => ['type' => 'number'],
								'vat_amount' => ['type' => 'number'],
                                'total' => ['type' => 'number'],
                                'period' => [
                                    'type' => 'object',
                                    'properties' => [
                                        'from' => ['type' => 'string', 'format' => 'date-time'],
                                        'to' => ['type' => 'string', 'format' => 'date-time'],
                                    ],
                                ],
                            ],
                        ],
                    ],
                    'totals' => [
                        'type' => 'object',
                        'properties' => [
                            'subtotal' => ['type' => 'number'],
                            'vat' => ['type' => 'number'],
                            'total' => ['type' => 'number'],
                        ],
                    ],
                    'notes' => ['type' => 'string'],
                ],
            ],
        ],
    ];

    $response = $client->chat()->create([
        'model' => 'gpt-4',
        'messages' => [
            ['role' => 'system', 'content' => 'You are a helpful assistant.'],
            ['role' => 'user', 'content' => $prompt],
        ],
        'functions' => $functions,
        'function_call' => ['name' => 'extract_invoice_data'],
    ]);

    $function_call = $response['choices'][0]['message']['function_call'];
    $invoice_data = json_decode($function_call['arguments'], true);

    if (isset($invoice_data['invoice']['date'])) {
        try {
            $date = new DateTime($invoice_data['invoice']['date']);
            // Reformat the date to 'Y-m-d' (YYYY-MM-DD)
            $invoice_data['invoice']['date'] = $date->format('Y-m-d');
        } catch (Exception $e) {
            // Handle the exception if the date is invalid
            echo 'Invalid date format: ', $e->getMessage();
        }
    }

    file_put_contents($outputFile, json_encode($invoice_data, JSON_PRETTY_PRINT));

    //echo 'Generated JSON file successfully.';
	
	echo json_encode($invoice_data, JSON_PRETTY_PRINT);
	die();
}

$filecontent = $_POST['file'];
$eurovalue = null;
if (isset($_POST['eurovalue'])){
	$eurovalue = $_POST['eurovalue'];
}

$inputFile = "example.pdf";
file_put_contents($inputFile, $filecontent);
$outputFile = "example.json";

$extractedText = ExtractTextFromCopyablePDF($inputFile);
//echo "{$extractedText}\n";
//echo strlen($extractedText) . "\n";

if (strlen($extractedText) > 50) {
    //echo "Given PDF file is copyable file.\n";
    OpenAIFunctionCalling($extractedText, $openaiKey, $outputFile, $eurovalue);
} else {
    //echo "Given PDF file is non-copyable file.\n";
    $extractedText = CustomizedMindeeAPI($mindeeKey, $inputFile);
    //echo "{$extractedText}\n";
    //echo strlen($extractedText) . "\n";
    OpenAIFunctionCalling($extractedText, $openaiKey, $outputFile, $eurovalue);
}