netgescon-day0/app/Services/Contabilita/BpmCsvParser.php

<?php

namespace App\Services\Contabilita;

use Carbon\Carbon;

class BpmCsvParser
{
    /**
     * Parser per export CSV Banco BPM con header tipo:
     * "Ragione Sociale","Data contabile","Data valuta","Banca","Rapporto","Importo","Divisa","Descrizione","Categoria/sottocategoria","Hashtag"
     *
     * @return array{
     *   rows: array<int, array{data: Carbon, valuta: Carbon|null, descrizione: string, importo: float, causale: string|null, raw_line: string}>,
     *   meta: array{header_row:int, delimiter:string, parsed_rows:int, skipped_rows:int}
     * }
     */
    public function parse(string $content): array
    {
        $content = str_replace(["\r\n", "\r"], "\n", $content);
        $lines = explode("\n", $content);

        $delimiter = ',';
        $headerIndex = null;

        $scanMax = min(count($lines), 100);
        for ($i = 0; $i < $scanMax; $i++) {
            $line = trim((string) $lines[$i]);
            if ($line === '') {
                continue;
            }

            if (
                stripos($line, 'Data contabile') !== false
                && stripos($line, 'Importo') !== false
                && stripos($line, 'Descrizione') !== false
            ) {
                $headerIndex = $i;
                if (substr_count($line, ';') > substr_count($line, ',')) {
                    $delimiter = ';';
                }
                break;
            }
        }

        if ($headerIndex === null) {
            throw new \RuntimeException('Formato CSV Banco BPM non riconosciuto: header con "Data contabile", "Importo" e "Descrizione" non trovato');
        }

        $header = str_getcsv((string) $lines[$headerIndex], $delimiter);
        $header = array_map(fn($h) => $this->normHeader((string) $h), $header);

        $hmap = [];
        foreach ($header as $idx => $h) {
            if ($h === '') {
                continue;
            }
            if (!array_key_exists($h, $hmap)) {
                $hmap[$h] = (int) $idx;
            }
        }

        $idxData = $hmap['datacontabile'] ?? ($hmap['data'] ?? null);
        $idxValuta = $hmap['datavaluta'] ?? ($hmap['valuta'] ?? null);
        $idxDescr = $hmap['descrizione'] ?? null;
        $idxImporto = $hmap['importo'] ?? null;
        $idxCat = $hmap['categoriasottocategoria'] ?? ($hmap['categoria'] ?? null);

        if ($idxData === null || $idxDescr === null || $idxImporto === null) {
            throw new \RuntimeException('Formato CSV Banco BPM non riconosciuto: colonne richieste mancanti');
        }

        $rows = [];
        $skipped = 0;

        for ($i = $headerIndex + 1; $i < count($lines); $i++) {
            $line = trim((string) $lines[$i]);
            if ($line === '') {
                continue;
            }

            $cols = str_getcsv($line, $delimiter);
            if (count($cols) <= max($idxData, $idxDescr, $idxImporto)) {
                $skipped++;
                continue;
            }

            $dateRaw = trim((string) ($cols[$idxData] ?? ''));
            $valutaRaw = $idxValuta !== null ? trim((string) ($cols[$idxValuta] ?? '')) : '';
            $descr = trim((string) ($cols[$idxDescr] ?? ''));
            $importoRaw = trim((string) ($cols[$idxImporto] ?? ''));
            $categoria = $idxCat !== null ? trim((string) ($cols[$idxCat] ?? '')) : null;

            if ($dateRaw === '' || $importoRaw === '') {
                $skipped++;
                continue;
            }

            $date = $this->parseDate($dateRaw);
            if (! $date) {
                $skipped++;
                continue;
            }

            $valuta = $valutaRaw !== '' ? $this->parseDate($valutaRaw) : null;
            $importo = $this->parseAmount($importoRaw);
            if ($importo === null) {
                $skipped++;
                continue;
            }

            $causale = null;
            if ($categoria !== null && $categoria !== '') {
                $causale = mb_substr($categoria, 0, 20);
            }

            $rows[] = [
                'data'        => $date,
                'valuta'      => $valuta,
                'descrizione' => $descr,
                'importo'     => $importo,
                'causale'     => $causale,
                'raw_line'    => $line,
            ];
        }

        return [
            'rows' => $rows,
            'meta' => [
                'header_row'   => $headerIndex,
                'delimiter'    => $delimiter,
                'parsed_rows'  => count($rows),
                'skipped_rows' => $skipped,
            ],
        ];
    }

    private function normHeader(string $value): string
    {
        $v = mb_strtolower(trim($value));
        return (string) preg_replace('/[^a-z0-9]+/i', '', $v);
    }

    private function parseDate(string $raw): ?Carbon
    {
        $raw = trim($raw);
        if ($raw === '') {
            return null;
        }

        foreach (['d/m/Y', 'Y-m-d', 'd-m-Y', 'd.m.Y'] as $fmt) {
            try {
                $dt = Carbon::createFromFormat($fmt, $raw);
                if ($dt && $dt->year >= 1990 && $dt->year <= 2100) {
                    return $dt->startOfDay();
                }
            } catch (\Throwable) {}
        }

        return null;
    }

    private function parseAmount(string $raw): ?float
    {
        $raw = trim($raw);
        if ($raw === '') {
            return null;
        }

        $raw = str_replace('€', '', $raw);
        $raw = trim($raw);

        if (str_contains($raw, ',') && str_contains($raw, '.')) {
            $raw = str_replace('.', '', $raw);
            $raw = str_replace(',', '.', $raw);
        } elseif (str_contains($raw, ',')) {
            $raw = str_replace(',', '.', $raw);
        }

        $raw = preg_replace('/[^\d\.\+\-]/', '', $raw);
        if ($raw === '' || !is_numeric($raw)) {
            return null;
        }

        return round((float) $raw, 2);
    }
}