[FEAT] CertiSource atti-service.php integration: structured data, PAT auth, ATECO fix

- VisuraService::fetchFromCertiSource: new atti-service.php API (POST richiesta → polling stato → GET dati)
- Structured data mapping: sedi/ateco_codes/cariche/addetti → formato interno
- mapAtecoToNis2Sector: allineato ENUM DB (digital_infra, water, waste, public_admin, ecc.)
- config.php: CERTISOURCE_API_URL, CERTISOURCE_API_KEY, CERTISOURCE_POLL_MAX/SEC
- PHP 8.4: curl_close → unset,  usato in logAiInteraction

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
DevEnv nis2-agile 2026-03-17 13:24:03 +01:00
parent 0e2774d1a6
commit 75a678f60e
2 changed files with 339 additions and 172 deletions

View File

@ -82,6 +82,14 @@ define('ANTHROPIC_API_KEY', Env::get('ANTHROPIC_API_KEY', ''));
define('ANTHROPIC_MODEL', Env::get('ANTHROPIC_MODEL', 'claude-sonnet-4-5-20250929'));
define('ANTHROPIC_MAX_TOKENS', Env::int('ANTHROPIC_MAX_TOKENS', 4096));
// ═══════════════════════════════════════════════════════════════════════════
// CERTISOURCE (atti-service.php)
// ═══════════════════════════════════════════════════════════════════════════
define('CERTISOURCE_API_URL', Env::get('CERTISOURCE_API_URL', 'https://certisource.it/atti-service.php'));
define('CERTISOURCE_API_KEY', Env::get('CERTISOURCE_API_KEY', '')); // cs_pat_...
define('CERTISOURCE_POLL_MAX', Env::int('CERTISOURCE_POLL_MAX', 30)); // max tentativi polling
define('CERTISOURCE_POLL_SEC', Env::int('CERTISOURCE_POLL_SEC', 3)); // secondi tra poll
// ═══════════════════════════════════════════════════════════════════════════
// FEEDBACK & SEGNALAZIONI
// ═══════════════════════════════════════════════════════════════════════════

View File

@ -2,14 +2,21 @@
/**
* NIS2 Agile - Visura Service
*
* Estrae dati aziendali da visura camerale PDF tramite AI
* e recupera dati da CertiSource.
* Estrae dati aziendali da visura camerale PDF tramite AI (upload)
* oppure tramite CertiSource atti-service.php (ricerca per P.IVA).
*
* CertiSource API: https://certisource.it/atti-service.php
* Auth: Authorization: Bearer CERTISOURCE_API_KEY (cs_pat_...)
*/
class VisuraService
{
// ─────────────────────────────────────────────────────────────────────────
// METODI PUBBLICI
// ─────────────────────────────────────────────────────────────────────────
/**
* Extract company data from a PDF visura camerale using Claude AI
* Estrai dati aziendali da un PDF visura caricato dall'utente (via Claude AI).
*/
public function extractFromPdf(string $filePath): array
{
@ -21,18 +28,15 @@ class VisuraService
throw new RuntimeException('Chiave API Anthropic non configurata');
}
// Read PDF and base64 encode it
$pdfContent = file_get_contents($filePath);
$base64Pdf = base64_encode($pdfContent);
$base64Pdf = base64_encode(file_get_contents($filePath));
// Call Claude API with the PDF
$response = $this->callClaudeApi([
[
'type' => 'document',
'type' => 'document',
'source' => [
'type' => 'base64',
'type' => 'base64',
'media_type' => 'application/pdf',
'data' => $base64Pdf,
'data' => $base64Pdf,
],
],
[
@ -42,170 +46,352 @@ class VisuraService
]);
if (!$response) {
throw new RuntimeException('Nessuna risposta dall\'AI');
throw new RuntimeException("Nessuna risposta dall'AI");
}
// Parse JSON response
$jsonStr = trim($response);
// Remove potential markdown code blocks
$jsonStr = preg_replace('/^```(?:json)?\s*/i', '', $jsonStr);
$jsonStr = preg_replace('/^```(?:json)?\s*/i', '', trim($response));
$jsonStr = preg_replace('/\s*```$/', '', $jsonStr);
$data = json_decode($jsonStr, true);
$data = json_decode($jsonStr, true);
if (json_last_error() !== JSON_ERROR_NONE) {
error_log('[VISURA_PARSE_ERROR] Could not parse AI response: ' . $jsonStr);
error_log('[VISURA_PARSE_ERROR] ' . $jsonStr);
throw new RuntimeException('Impossibile interpretare i dati estratti dalla visura');
}
// Map to suggested NIS2 sector based on ATECO code
$data['suggested_sector'] = $this->mapAtecoToNis2Sector($data['ateco_code'] ?? '', $data['ateco_description'] ?? '');
$data['suggested_sector'] = $this->mapAtecoToNis2Sector(
$data['ateco_code'] ?? '',
$data['ateco_description'] ?? ''
);
$data['source'] = 'pdf_upload';
// Log AI interaction
$this->logAiInteraction('visura_extraction', 'Estrazione dati da visura camerale PDF');
return $data;
}
/**
* Fetch company data from CertiSource API
* Recupera dati aziendali da CertiSource atti-service.php tramite P.IVA.
*
* Flusso:
* POST ?action=richiesta ottieni request ID
* GET ?action=stato polling fino a completed
* GET ?action=dati structured_data JSON
*/
public function fetchFromCertiSource(string $vatNumber): array
{
// CertiSource is on the same server - call its API internally
$certisourceUrl = $this->getCertiSourceBaseUrl() . '/api/company/enrich';
$apiKey = defined('CERTISOURCE_API_KEY') ? CERTISOURCE_API_KEY : '';
if ($apiKey === '') {
error_log('[CERTISOURCE] API key non configurata (CERTISOURCE_API_KEY)');
throw new RuntimeException('Servizio visure non disponibile. Inserire i dati manualmente.');
}
$ch = curl_init();
curl_setopt_array($ch, [
CURLOPT_URL => $certisourceUrl . '?vat=' . urlencode($vatNumber),
CURLOPT_RETURNTRANSFER => true,
CURLOPT_TIMEOUT => 30,
CURLOPT_HTTPHEADER => [
'Content-Type: application/json',
'Accept: application/json',
'X-Internal-Service: nis2-agile',
],
// Same server, skip SSL verification for internal calls
CURLOPT_SSL_VERIFYPEER => false,
// ── Step 1: richiesta visura ─────────────────────────────────────────
$richiesta = $this->callCertiSource('POST', 'richiesta', [
'vat' => $vatNumber,
'type' => 'visura_camerale',
]);
$response = curl_exec($ch);
if (empty($richiesta['success'])) {
$msg = $richiesta['error'] ?? $richiesta['message'] ?? 'Errore richiesta visura';
throw new RuntimeException('CertiSource: ' . $msg);
}
$requestId = (int) $richiesta['id'];
$status = $richiesta['status'] ?? 'processing';
// ── Step 2: polling se non già completato (cache hit) ────────────────
if ($status !== 'completed') {
$pollMax = defined('CERTISOURCE_POLL_MAX') ? (int) CERTISOURCE_POLL_MAX : 30;
$pollSec = defined('CERTISOURCE_POLL_SEC') ? (int) CERTISOURCE_POLL_SEC : 3;
for ($i = 0; $i < $pollMax; $i++) {
sleep($pollSec);
$stato = $this->callCertiSource('GET', 'stato', null, ['id' => $requestId]);
if (!empty($stato['completed'])) {
$status = 'completed';
break;
}
if (($stato['status'] ?? '') === 'failed') {
throw new RuntimeException('CertiSource: elaborazione visura fallita');
}
}
if ($status !== 'completed') {
throw new RuntimeException('Timeout: visura camerale non disponibile. Riprova tra qualche minuto.');
}
}
// ── Step 3: dati strutturati ─────────────────────────────────────────
$datiRes = $this->callCertiSource('GET', 'dati', null, ['id' => $requestId]);
if (empty($datiRes['success']) || empty($datiRes['structured_data'])) {
throw new RuntimeException('CertiSource: dati strutturati non disponibili');
}
return $this->mapStructuredData($datiRes['structured_data'], $vatNumber, $requestId);
}
// ─────────────────────────────────────────────────────────────────────────
// METODI PRIVATI — CertiSource
// ─────────────────────────────────────────────────────────────────────────
/**
* Chiama l'API CertiSource atti-service.php.
*
* @param string $method GET|POST
* @param string $action action=richiesta|stato|dati|...
* @param array|null $body JSON body per POST
* @param array $params Query params extra (es. ['id'=>42])
*/
private function callCertiSource(string $method, string $action, ?array $body = null, array $params = []): array
{
$apiKey = defined('CERTISOURCE_API_KEY') ? CERTISOURCE_API_KEY : '';
$baseUrl = defined('CERTISOURCE_API_URL') ? CERTISOURCE_API_URL : 'https://certisource.it/atti-service.php';
$qs = http_build_query(array_merge(['action' => $action], $params));
$url = $baseUrl . '?' . $qs;
$ch = curl_init($url);
$headers = [
'Authorization: Bearer ' . $apiKey,
'Content-Type: application/json',
'Accept: application/json',
];
curl_setopt_array($ch, [
CURLOPT_RETURNTRANSFER => true,
CURLOPT_TIMEOUT => 15,
CURLOPT_HTTPHEADER => $headers,
]);
if ($method === 'POST' && $body !== null) {
curl_setopt($ch, CURLOPT_POST, true);
curl_setopt($ch, CURLOPT_POSTFIELDS, json_encode($body));
}
$raw = curl_exec($ch);
$httpCode = curl_getinfo($ch, CURLINFO_HTTP_CODE);
$error = curl_error($ch);
curl_close($ch);
$curlErr = curl_error($ch);
unset($ch);
if ($error) {
error_log("[CERTISOURCE_CURL_ERROR] $error");
throw new RuntimeException('Impossibile contattare CertiSource: ' . $error);
if ($curlErr) {
throw new RuntimeException('CertiSource connessione: ' . $curlErr);
}
if ($httpCode !== 200) {
error_log("[CERTISOURCE_HTTP_ERROR] HTTP $httpCode: $response");
throw new RuntimeException('CertiSource ha restituito un errore (HTTP ' . $httpCode . ')');
if ($httpCode === 402) {
throw new RuntimeException('CertiSource: credito insufficiente');
}
$result = json_decode($response, true);
if (!$result) {
throw new RuntimeException('Risposta CertiSource non valida');
if ($httpCode >= 500) {
throw new RuntimeException('CertiSource: errore server (HTTP ' . $httpCode . ')');
}
// Map CertiSource response to our format
// CertiSource typically returns data in its own format, normalize it
$companyData = $result['data'] ?? $result;
$result = json_decode($raw ?: '{}', true);
if (!is_array($result)) {
throw new RuntimeException('CertiSource: risposta non valida');
}
return $result;
}
/**
* Mappa structured_data CertiSource nel formato interno NIS2 Agile.
*/
private function mapStructuredData(array $sd, string $vatNumber, int $requestId): array
{
$sedeLegale = $this->extractSedeLegale($sd['sedi'] ?? []);
$ateco = $this->extractPrimaryAteco($sd['ateco_codes'] ?? []);
$atecoCode = $ateco['code'] ?? null;
$atecoDesc = $ateco['description'] ?? '';
return [
'company_name' => $companyData['ragione_sociale'] ?? $companyData['denominazione'] ?? $companyData['company_name'] ?? null,
'vat_number' => $companyData['partita_iva'] ?? $companyData['vat_number'] ?? $vatNumber,
'fiscal_code' => $companyData['codice_fiscale'] ?? $companyData['fiscal_code'] ?? null,
'legal_form' => $companyData['forma_giuridica'] ?? $companyData['legal_form'] ?? null,
'address' => $companyData['indirizzo'] ?? $companyData['address'] ?? null,
'city' => $companyData['comune'] ?? $companyData['city'] ?? null,
'province' => $companyData['provincia'] ?? $companyData['province'] ?? null,
'zip_code' => $companyData['cap'] ?? $companyData['zip_code'] ?? null,
'pec' => $companyData['pec'] ?? null,
'phone' => $companyData['telefono'] ?? $companyData['phone'] ?? null,
'ateco_code' => $companyData['codice_ateco'] ?? $companyData['ateco_code'] ?? null,
'ateco_description' => $companyData['descrizione_ateco'] ?? $companyData['ateco_description'] ?? null,
'suggested_sector' => $this->mapAtecoToNis2Sector(
$companyData['codice_ateco'] ?? '',
$companyData['descrizione_ateco'] ?? ''
),
'source' => 'certisource',
'company_name' => $sd['ragione_sociale'] ?? null,
'vat_number' => $sd['partita_iva'] ?? $vatNumber,
'fiscal_code' => $sd['codice_fiscale'] ?? null,
'legal_form' => $sd['forma_giuridica'] ?? null,
'address' => $sedeLegale['via'] ?? null,
'city' => $sedeLegale['comune'] ?? null,
'province' => $sedeLegale['provincia'] ?? null,
'zip_code' => $sedeLegale['cap'] ?? null,
'pec' => $sd['pec'] ?? $sedeLegale['pec'] ?? null,
'phone' => null,
'ateco_code' => $atecoCode,
'ateco_description' => $atecoDesc ?: null,
'incorporation_date' => null,
'share_capital' => isset($sd['capitale_sociale']) ? (float) $sd['capitale_sociale'] : null,
'employees_range' => $this->computeEmployeesRange($sd['addetti'] ?? null),
'legal_representative' => $this->findLegalRepresentative($sd['cariche'] ?? []),
'suggested_sector' => $this->mapAtecoToNis2Sector($atecoCode ?? '', $atecoDesc),
'source' => 'certisource_atti',
'certisource_id' => $requestId,
];
}
/**
* Map ATECO code to NIS2 sector
* Trova la sede legale nell'array sedi.
*/
private function extractSedeLegale(array $sedi): array
{
foreach ($sedi as $sede) {
if (($sede['tipo'] ?? '') === 'sede_legale') {
return $sede;
}
}
return $sedi[0] ?? [];
}
/**
* Trova il codice ATECO primario.
*/
private function extractPrimaryAteco(array $atecos): array
{
foreach ($atecos as $a) {
if (($a['type'] ?? '') === 'primary') {
return $a;
}
}
return $atecos[0] ?? [];
}
/**
* Calcola il range dipendenti dai dati INPS (ultimo trimestre disponibile).
*/
private function computeEmployeesRange(?array $addetti): ?string
{
if (empty($addetti['trimestri'])) {
return null;
}
$last = end($addetti['trimestri']);
$n = (int) ($last['dipendenti'] ?? 0);
if ($n === 0) return null;
if ($n < 10) return '1-9';
if ($n < 50) return '10-49';
if ($n < 250) return '50-249';
return '250+';
}
/**
* Trova il legale rappresentante tra le cariche.
*/
private function findLegalRepresentative(array $cariche): ?string
{
foreach ($cariche as $c) {
if (!empty($c['rappresentante_legale']) || str_contains(strtolower($c['ruolo'] ?? ''), 'legale')) {
return $c['nome'] ?? null;
}
}
// Fallback: primo amministratore
foreach ($cariche as $c) {
if (str_contains(strtolower($c['ruolo'] ?? ''), 'ammin')) {
return $c['nome'] ?? null;
}
}
return null;
}
// ─────────────────────────────────────────────────────────────────────────
// ATECO → SETTORE NIS2 (valori ENUM DB)
// ─────────────────────────────────────────────────────────────────────────
/**
* Mappa codice ATECO + descrizione al settore NIS2 (valori ENUM organizations.sector).
*
* Valori validi: energy, transport, banking, health, water, digital_infra,
* public_admin, manufacturing, postal, chemical, food, waste,
* ict_services, digital_providers, space, research, other
*/
private function mapAtecoToNis2Sector(string $atecoCode, string $atecoDesc): ?string
{
$code = substr($atecoCode, 0, 2); // Use first 2 digits
$descLower = strtolower($atecoDesc);
$code2 = substr(preg_replace('/[^0-9]/', '', $atecoCode), 0, 2);
$desc = strtolower($atecoDesc);
// ATECO to NIS2 mapping (approximate)
$mapping = [
'35' => 'energy_electricity', // Electricity, gas, steam
'49' => 'transport_road', // Land transport
'50' => 'transport_water', // Water transport
'51' => 'transport_air', // Air transport
'64' => 'banking', // Financial services
'65' => 'banking', // Insurance
'66' => 'financial_markets', // Financial auxiliaries
'86' => 'health', // Health
'36' => 'drinking_water', // Water supply
'37' => 'waste_water', // Sewerage
'38' => 'waste_management', // Waste management
'61' => 'digital_infrastructure', // Telecommunications
'62' => 'ict_service_management', // IT services
'63' => 'digital_providers', // Information services
'84' => 'public_administration', // Public admin
'53' => 'postal_courier', // Postal services
'20' => 'chemicals', // Chemicals manufacturing
'10' => 'food', // Food manufacturing
'11' => 'food', // Beverages
'21' => 'manufacturing_medical', // Pharma/medical
'26' => 'manufacturing_computers', // Electronics
'27' => 'manufacturing_electrical', // Electrical equipment
'28' => 'manufacturing_machinery', // Machinery
'29' => 'manufacturing_vehicles', // Motor vehicles
'30' => 'manufacturing_transport', // Other transport
'72' => 'research', // Scientific research
$byCode = [
'35' => 'energy', // Fornitura energia elettrica, gas
'06' => 'energy', // Estrazione petrolio e gas
'19' => 'energy', // Raffinazione petrolio
'49' => 'transport', // Trasporto terrestre
'50' => 'transport', // Trasporto marittimo
'51' => 'transport', // Trasporto aereo
'52' => 'transport', // Magazzinaggio e attività connesse
'64' => 'banking', // Servizi finanziari
'65' => 'banking', // Assicurazioni
'66' => 'banking', // Attività ausiliarie servizi finanziari
'86' => 'health', // Assistenza sanitaria
'87' => 'health', // Strutture di assistenza
'36' => 'water', // Raccolta e distribuzione acqua
'37' => 'waste', // Raccolta e trattamento acque reflue
'38' => 'waste', // Raccolta e smaltimento rifiuti
'61' => 'digital_infra', // Telecomunicazioni
'62' => 'ict_services', // Produzione software e IT
'63' => 'digital_providers',// Servizi di informazione
'84' => 'public_admin', // Pubblica amministrazione
'53' => 'postal', // Servizi postali e corrieri
'20' => 'chemical', // Fabbricazione prodotti chimici
'10' => 'food', // Industria alimentare
'11' => 'food', // Industria delle bevande
'21' => 'manufacturing', // Fabbricazione farmaci
'26' => 'manufacturing', // Fabbricazione computer e elettronica
'27' => 'manufacturing', // Fabbricazione apparecchiature elettriche
'28' => 'manufacturing', // Fabbricazione macchinari
'29' => 'manufacturing', // Fabbricazione autoveicoli
'30' => 'manufacturing', // Fabbricazione altri mezzi di trasporto
'72' => 'research', // Ricerca e sviluppo
'85' => 'research', // Istruzione
];
if (isset($mapping[$code])) {
return $mapping[$code];
if (isset($byCode[$code2])) {
return $byCode[$code2];
}
// Try to match by description keywords
$keywords = [
'energia' => 'energy_electricity',
'elettric' => 'energy_electricity',
'gas' => 'energy_gas',
'petroli' => 'energy_oil',
'trasport' => 'transport_road',
'ferrov' => 'transport_rail',
'maritt' => 'transport_water',
'aere' => 'transport_air',
'banc' => 'banking',
'finanz' => 'financial_markets',
'sanit' => 'health',
'osped' => 'health',
'farm' => 'manufacturing_medical',
'acqua' => 'drinking_water',
'rifiut' => 'waste_management',
'telecom' => 'digital_infrastructure',
'informatica' => 'ict_service_management',
'software' => 'ict_service_management',
'digital' => 'digital_providers',
'postale' => 'postal_courier',
'corriere' => 'postal_courier',
'chimic' => 'chemicals',
'alimentar' => 'food',
'ricerca' => 'research',
// Fallback per keyword nella descrizione
$byKeyword = [
'energia' => 'energy',
'elettric' => 'energy',
'gas' => 'energy',
'petroli' => 'energy',
'idrogeno' => 'energy',
'trasport' => 'transport',
'ferrov' => 'transport',
'maritt' => 'transport',
'aereo' => 'transport',
'aere' => 'transport',
'logistic' => 'transport',
'banc' => 'banking',
'finanz' => 'banking',
'assicur' => 'banking',
'sanit' => 'health',
'osped' => 'health',
'clinic' => 'health',
'medic' => 'health',
'acqua' => 'water',
'idric' => 'water',
'rifiut' => 'waste',
'smaltim' => 'waste',
'telecom' => 'digital_infra',
'teleco' => 'digital_infra',
'fibra' => 'digital_infra',
'internet' => 'digital_infra',
'informatica' => 'ict_services',
'software' => 'ict_services',
'cloud' => 'ict_services',
'digital' => 'digital_providers',
'piattaform' => 'digital_providers',
'postale' => 'postal',
'corriere' => 'postal',
'spediz' => 'postal',
'chimic' => 'chemical',
'farmac' => 'manufacturing',
'alimentar' => 'food',
'bevand' => 'food',
'ricerca' => 'research',
'spazial' => 'space',
'aerospaz' => 'space',
'manifattur' => 'manufacturing',
];
foreach ($keywords as $kw => $sector) {
if (str_contains($descLower, $kw)) {
foreach ($byKeyword as $kw => $sector) {
if (str_contains($desc, $kw)) {
return $sector;
}
}
@ -213,20 +399,16 @@ class VisuraService
return null;
}
/**
* Call Claude API
*/
// ─────────────────────────────────────────────────────────────────────────
// METODI PRIVATI — Claude AI
// ─────────────────────────────────────────────────────────────────────────
private function callClaudeApi(array $content): ?string
{
$payload = [
'model' => ANTHROPIC_MODEL,
'max_tokens' => ANTHROPIC_MAX_TOKENS,
'messages' => [
[
'role' => 'user',
'content' => $content,
],
],
'messages' => [['role' => 'user', 'content' => $content]],
];
$ch = curl_init('https://api.anthropic.com/v1/messages');
@ -244,58 +426,36 @@ class VisuraService
$response = curl_exec($ch);
$httpCode = curl_getinfo($ch, CURLINFO_HTTP_CODE);
$error = curl_error($ch);
curl_close($ch);
if ($error) {
throw new RuntimeException('Claude API error: ' . $error);
}
$error = curl_error($ch);
unset($ch);
if ($error) throw new RuntimeException('Claude API error: ' . $error);
if ($httpCode !== 200) {
error_log("[CLAUDE_API_ERROR] HTTP $httpCode: $response");
error_log('[CLAUDE_API_ERROR] HTTP ' . $httpCode . ': ' . $response);
throw new RuntimeException('Claude API returned HTTP ' . $httpCode);
}
$result = json_decode($response, true);
return $result['content'][0]['text'] ?? null;
}
/**
* Get CertiSource API base URL (same server)
*/
private function getCertiSourceBaseUrl(): string
{
// Both apps are on the same server, use internal URL
if (defined('CERTISOURCE_API_URL')) {
return CERTISOURCE_API_URL;
}
// Default: same server via localhost
return 'https://certisource.it/certisource';
}
/**
* Log AI interaction to database
*/
private function logAiInteraction(string $type, string $summary): void
{
try {
// Get current user from JWT if available
$userId = null;
$token = $_SERVER['HTTP_AUTHORIZATION'] ?? '';
if (preg_match('/Bearer\s+(.+)$/i', $token, $matches)) {
$parts = explode('.', $matches[1]);
$token = $_SERVER['HTTP_AUTHORIZATION'] ?? '';
if (preg_match('/Bearer\s+(.+)$/i', $token, $m)) {
$parts = explode('.', $m[1]);
if (count($parts) === 3) {
$payload = json_decode(base64_decode(strtr($parts[1], '-_', '+/')), true);
$userId = $payload['sub'] ?? null;
$pl = json_decode(base64_decode(strtr($parts[1], '-_', '+/')), true);
$userId = $pl['sub'] ?? null;
}
}
if ($userId) {
Database::insert('ai_interactions', [
'organization_id' => 0, // Not yet created during onboarding
'organization_id' => 0,
'user_id' => $userId,
'interaction_type' => 'qa',
'interaction_type' => $type,
'prompt_summary' => $summary,
'response_summary' => 'Dati estratti',
'tokens_used' => 0,
@ -303,7 +463,6 @@ class VisuraService
]);
}
} catch (Throwable $e) {
// Silently fail - logging should not break the flow
error_log('[AI_LOG_ERROR] ' . $e->getMessage());
}
}