index.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'];

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)
{
    $client = OpenAI::client($openaiKey);

    $prompt = "Extract the emitter and recipient company information, invoice number, date, currency, and line items from the following invoice text:\n\n" . $extractedText;

    $functions = [
        [
            'name' => 'extract_invoice_data',
            'description' => 'Extracts emitter and recipient company information, invoice number, date, currency, and line items from an invoice.',
            'parameters' => [
                'type' => 'object',
                'properties' => [
                    'emitter' => [
                        'type' => 'object',
                        'properties' => [
                            'name' => ['type' => 'string'],
                            'address' => ['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_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'],
                                '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'],
    ]);

    if (isset($response['usage'])) {
        $totalTokens = $response['usage']['total_tokens'];
        $promptTokens = $response['usage']['prompt_tokens'];
        $completionTokens = $response['usage']['completion_tokens'];

        echo "Total Tokens: $totalTokens\n";
        echo "Prompt Tokens: $promptTokens\n";
        echo "Completion Tokens: $completionTokens\n";

        // Save token usage to a file
        $fileContent = "Total Tokens: $totalTokens\n";
        $fileContent .= "Prompt Tokens: $promptTokens\n";
        $fileContent .= "Completion Tokens: $completionTokens\n";

        file_put_contents("usedToken.txt", $fileContent);

        echo "Token usage has been saved to usedToken.txt.\n";
    } else {
        echo "Usage data not available.";
    }

    $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.';
}

$inputFile = "2.pdf";
$outputFile = "2.json";

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

if (strlen($extractedText) > 50) {
    echo "Given PDF file is copyable file.\n";

    file_put_contents("extractedText.txt", $extractedText);
    OpenAIFunctionCalling($extractedText, $openaiKey, $outputFile);
} else {
    echo "Given PDF file is non-copyable file.\n";

    $extractedText = CustomizedMindeeAPI($mindeeKey, $inputFile);
    file_put_contents("extractedText.txt", $extractedText);
    echo "{$extractedText}\n";
    echo strlen($extractedText) . "\n";

    OpenAIFunctionCalling($extractedText, $openaiKey, $outputFile);
}