netgescon-day0/app/Services/Consumi/GasLucePdfTextParser.php

<?php
namespace App\Services\Consumi;

use Carbon\Carbon;

class GasLucePdfTextParser
{
    /**
     * @return array{
     *   codici: array{cliente: string|null, contratto: string|null, pdr: string|null, pod: string|null},
     *   pagamento: array{scadenza: string|null},
     *   consumi: array<int, array{dal: string|null, al: string|null, valore: float|null, unita: string, tipologia_lettura: string}>
     * }
     */
    public function parse(string $text): array
    {
        $text = str_replace("\r\n", "\n", $text);
        $text = str_replace("\r", "\n", $text);

        // Cliente
        $cliente = $this->firstMatch($text, '/(?:Codice\s+Cliente|Cod\.\s*Cliente|Cliente\s*n\.|Client\s*code)\s*:?\s*([A-Z0-9_]{4,})/i');
        
        // Contratto
        $contratto = $this->firstMatch($text, '/(?:Contratto|Cod\.\s*Contratto|Codice\s+Contratto|Contract\s*n\.)\s*:?\s*([A-Z0-9_]{4,})/i');
        
        // PDR / POD
        $pdr = $this->firstMatch($text, '/\bPDR\b\s*:?\s*([0-9]{14})\b/i');
        $pod = $this->firstMatch($text, '/\bPOD\b\s*:?\s*([A-Z0-9]{14,15})\b/i');

        // Scadenza
        $scadenza = $this->toIsoDate($this->firstMatch($text, '/(?:Scadenza|Entro\s+il|Data\s+scadenza)\s*:?\s*(\d{2}\/\d{2}\/\d{4})/i'));

        // Cerca consumi del tipo: "Consumo Fatturato APRILE 2026 910 Smc"
        // O genericamente "Consumo... [valore] Smc/kWh"
        $consumi = [];
        if (preg_match_all('/(?:Consumo\s+Fatturato|Consumo|Consumo\s+rilevato|Consumo\s+stimato)\s+(?:([A-Z]+)\s+(\d{4})\s+)?(\d+(?:[\.,]\d+)?)\s*(Smc|kWh)/i', $text, $matches, PREG_SET_ORDER)) {
            foreach ($matches as $m) {
                $monthName = !empty($m[1]) ? strtoupper(trim($m[1])) : null;
                $yearVal = !empty($m[2]) ? (int) $m[2] : null;
                $valore = $this->toFloatIt($m[3]);
                $unita = trim($m[4]);

                $dal = null;
                $al = null;
                if ($monthName && $yearVal) {
                    $months = [
                        'GENNAIO' => 1, 'FEBBRAIO' => 2, 'MARZO' => 3, 'APRILE' => 4,
                        'MAGGIO' => 5, 'GIUGNO' => 6, 'LUGLIO' => 7, 'AGOSTO' => 8,
                        'SETTEMBRE' => 9, 'OTTOBRE' => 10, 'NOVEMBRE' => 11, 'DICEMBRE' => 12,
                        'JANUARY' => 1, 'FEBRUARY' => 2, 'MARCH' => 3, 'APRIL' => 4,
                        'MAY' => 5, 'JUNE' => 6, 'JULY' => 7, 'AUGUST' => 8,
                        'SEPTEMBER' => 9, 'OCTOBER' => 10, 'NOVEMBER' => 11, 'DECEMBER' => 12
                    ];
                    $mNum = $months[$monthName] ?? null;
                    if ($mNum) {
                        $dal = Carbon::create($yearVal, $mNum, 1)->startOfMonth()->format('Y-m-d');
                        $al = Carbon::create($yearVal, $mNum, 1)->endOfMonth()->format('Y-m-d');
                    }
                }

                $consumi[] = [
                    'dal' => $dal,
                    'al' => $al,
                    'valore' => $valore,
                    'unita' => $unita,
                    'tipologia_lettura' => 'fatturata',
                ];
            }
        }

        // Se non troviamo date dai consumi ma c'è un intervallo di date nel testo, tipo "dal 01/04/2026 al 30/04/2026"
        if (count($consumi) === 0 && preg_match('/dal\s+(\d{2}\/\d{2}\/\d{4})\s+al\s+(\d{2}\/\d{2}\/\d{4})/i', $text, $mDates)) {
            $dal = $this->toIsoDate($mDates[1]);
            $al = $this->toIsoDate($mDates[2]);
            
            // Prova a estrarre il valore di consumo vicino a Smc o kWh
            $val = null;
            $uni = 'Smc';
            if (preg_match('/(\d+(?:[\.,]\d+)?)\s*(Smc|kWh)/i', $text, $mVal)) {
                $val = $this->toFloatIt($mVal[1]);
                $uni = $mVal[2];
            }

            if ($val !== null) {
                $consumi[] = [
                    'dal' => $dal,
                    'al' => $al,
                    'valore' => $val,
                    'unita' => $uni,
                    'tipologia_lettura' => 'rilevata',
                ];
            }
        }

        return [
            'codici' => [
                'cliente' => $cliente,
                'contratto' => $contratto,
                'pdr' => $pdr,
                'pod' => $pod,
            ],
            'pagamento' => [
                'scadenza' => $scadenza,
            ],
            'consumi' => $consumi,
        ];
    }

    private function firstMatch(string $text, string $pattern): ?string
    {
        if (preg_match($pattern, $text, $matches)) {
            return trim($matches[1]);
        }
        return null;
    }

    private function toFloatIt(mixed $val): ?float
    {
        if ($val === null) return null;
        $val = str_replace('.', '', $val);
        $val = str_replace(',', '.', $val);
        return is_numeric($val) ? (float) $val : null;
    }

    private function toIsoDate(?string $value): ?string
    {
        if (! $value) {
            return null;
        }
        try {
            return Carbon::createFromFormat('d/m/Y', $value)->format('Y-m-d');
        } catch (\Throwable) {
            return null;
        }
    }
}