feat: add getDocumentBytes endpoint for OCR + SignatureRequest context

- New POST /api/v1/SmartAssistant/action/getDocumentBytes returns
  {success, fileName, mimeType, sizeBytes, base64} for use by
  shira-hermes Claude Vision OCR fallback.
- DocumentAnalyzer: new getDocumentBytes() method + guessMimeType helper.
- CaseContextBuilder: expose signatureRequests in case context (for
  DigitalSignature integration — gracefully empty if entity absent).
- README: sync version badge.
- Version bump 2.7.3 → 2.8.0 (new public endpoint).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
2026-04-19 11:45:05 +00:00
parent 405ceb7684
commit f400f2fec7
7 changed files with 442 additions and 23 deletions
+1 -1
View File
@@ -2,5 +2,5 @@
"currentTag": "master", "currentTag": "master",
"lastSwitched": "2026-04-14T06:05:06.044Z", "lastSwitched": "2026-04-14T06:05:06.044Z",
"branchTagMapping": {}, "branchTagMapping": {},
"migrationNoticeShown": false "migrationNoticeShown": true
} }
+23 -6
View File
@@ -2,21 +2,38 @@
"master": { "master": {
"tasks": [ "tasks": [
{ {
"id": 1, "id": "1",
"title": "fix: broaden task detection to include Contact-linked tasks", "title": "fix: broaden task detection to include Contact-linked tasks",
"description": "AlertCalculator, CaseContextBuilder, and OfficeContextBuilder only queried tasks with parentType='Case', missing tasks linked to case contacts (parentType='Contact'). This caused false 'no preparation task' alerts in Shira's daily standup.", "description": "AlertCalculator, CaseContextBuilder, and OfficeContextBuilder only queried tasks with parentType='Case', missing tasks linked to case contacts (parentType='Contact'). This caused false 'no preparation task' alerts in Shira's daily standup.",
"status": "in-progress", "status": "done",
"priority": "high", "priority": "high",
"dependencies": [], "dependencies": [],
"details": "Root cause confirmed via production API: preparation tasks had parentType='Contact' because Shira created them linked to the contact entity. Fix broadens all task queries to include both Case and Contact parent types, plus NhActivity-linked tasks.", "details": "Root cause confirmed via production API: preparation tasks had parentType='Contact' because Shira created them linked to the contact entity. Fix broadens all task queries to include both Case and Contact parent types, plus NhActivity-linked tasks.",
"testStrategy": "Deploy to staging, trigger standup for a case with Contact-linked tasks, verify no false alert.", "testStrategy": "Deploy to staging, trigger standup for a case with Contact-linked tasks, verify no false alert.",
"subtasks": [] "subtasks": [],
"updatedAt": "2026-04-14T06:08:19.494Z"
},
{
"id": "2",
"title": "Network-storage diagnostics + folder discovery",
"description": "Expose diagnostics when a case folder can't be listed and add endpoints (findCaseFolder, setCaseFolderPath) so Shira can locate and persist the correct network-storage path.",
"details": "",
"testStrategy": "",
"status": "in-progress",
"dependencies": [],
"priority": "medium",
"subtasks": [],
"updatedAt": "2026-04-19T06:11:17.422Z"
} }
], ],
"metadata": { "metadata": {
"created": "2026-04-14T06:05:06.042Z", "version": "1.0.0",
"updated": "2026-04-14T09:15:00.000Z", "lastModified": "2026-04-19T06:11:17.423Z",
"description": "Tasks for master context" "taskCount": 2,
"completedCount": 1,
"tags": [
"master"
]
} }
} }
} }
+1 -1
View File
@@ -1,6 +1,6 @@
# SmartAssistant - עוזר חכם # SmartAssistant - עוזר חכם
**גרסה:** 2.7.0 | **מחבר:** klear | **EspoCRM:** >= 8.0.0 **גרסה:** 2.7.3 | **מחבר:** klear | **EspoCRM:** >= 8.0.0
## תיאור ## תיאור
עוזר AI מאוחד למשרד עורכי דין. מספק ממשק צ'אט צף עם שני מצבי עבודה: **מצב משרד** (סקירה כללית, התראות, סטטיסטיקות) ו**מצב תיק** (סיוע מעמיק בתיק ספציפי). כולל מערכת זיכרון תיק מובנית, ביצוע פעולות באישור המשתמש, ואינטגרציה עם Stream של EspoCRM. עוזר AI מאוחד למשרד עורכי דין. מספק ממשק צ'אט צף עם שני מצבי עבודה: **מצב משרד** (סקירה כללית, התראות, סטטיסטיקות) ו**מצב תיק** (סיוע מעמיק בתיק ספציפי). כולל מערכת זיכרון תיק מובנית, ביצוע פעולות באישור המשתמש, ואינטגרציה עם Stream של EspoCRM.
@@ -217,6 +217,20 @@ class SmartAssistant
]; ];
} }
public function postActionGetDocumentBytes(Request $request, Response $response): array
{
$this->checkAccess();
$data = $request->getParsedBody();
$filePath = $data->filePath ?? null;
if (empty($filePath)) {
throw new BadRequest('filePath is required.');
}
$analyzer = $this->injectableFactory->create(DocumentAnalyzer::class);
return $analyzer->getDocumentBytes($filePath);
}
public function postActionReadMultipleDocuments(Request $request, Response $response): array public function postActionReadMultipleDocuments(Request $request, Response $response): array
{ {
$this->checkAccess(); $this->checkAccess();
@@ -279,6 +293,75 @@ class SmartAssistant
]; ];
} }
public function postActionListDocuments(Request $request, Response $response): array
{
$this->checkAccess();
$data = $request->getParsedBody();
$caseId = $data->caseId ?? null;
if (empty($caseId)) {
throw new BadRequest('caseId is required.');
}
$analyzer = $this->injectableFactory->create(DocumentAnalyzer::class);
return $analyzer->listDocumentsRecursive($caseId);
}
public function postActionBrowseFolder(Request $request, Response $response): array
{
$this->checkAccess();
$data = $request->getParsedBody();
$path = $data->path ?? null;
if ($path === null) {
throw new BadRequest('path is required.');
}
$analyzer = $this->injectableFactory->create(DocumentAnalyzer::class);
return [
'path' => $path,
'entries' => $analyzer->browseFolder((string) $path),
];
}
public function postActionFindCaseFolder(Request $request, Response $response): array
{
$this->checkAccess();
$data = $request->getParsedBody();
$query = $data->query ?? null;
if (empty($query)) {
throw new BadRequest('query is required.');
}
$limit = isset($data->limit) ? max(1, min((int) $data->limit, 50)) : 20;
$analyzer = $this->injectableFactory->create(DocumentAnalyzer::class);
return [
'query' => $query,
'matches' => $analyzer->findCaseFolder((string) $query, $limit),
];
}
public function postActionSetCaseFolderPath(Request $request, Response $response): array
{
$this->checkAccess();
if (!$this->acl->checkScope('Case', 'edit')) {
throw new Forbidden('No edit access to Case.');
}
$data = $request->getParsedBody();
$caseId = $data->caseId ?? null;
$path = $data->path ?? null;
if (empty($caseId) || empty($path)) {
throw new BadRequest('caseId and path are required.');
}
$analyzer = $this->injectableFactory->create(DocumentAnalyzer::class);
return $analyzer->setCaseFolderPath((string) $caseId, (string) $path);
}
public function postActionGenerateFromTemplate(Request $request, Response $response): array public function postActionGenerateFromTemplate(Request $request, Response $response): array
{ {
$this->checkAccess(); $this->checkAccess();
@@ -33,6 +33,7 @@ class CaseContextBuilder
'recentNotes' => $this->getRecentNotes($caseId), 'recentNotes' => $this->getRecentNotes($caseId),
'availableTemplates' => $this->getAvailableTemplates(), 'availableTemplates' => $this->getAvailableTemplates(),
'documents' => $this->getDocumentListing($caseId), 'documents' => $this->getDocumentListing($caseId),
'signatureRequests' => $this->getSignatureRequests($caseId),
'currentUser' => $this->getUserData($userId), 'currentUser' => $this->getUserData($userId),
'validStatuses' => ActionExecutor::VALID_STATUSES, 'validStatuses' => ActionExecutor::VALID_STATUSES,
]; ];
@@ -202,13 +203,51 @@ class CaseContextBuilder
return ['id' => $user->get('id'), 'name' => $user->get('name')]; return ['id' => $user->get('id'), 'name' => $user->get('name')];
} }
private function getSignatureRequests(string $caseId): array
{
try {
$collection = $this->entityManager->getRDBRepository('SignatureRequest')
->where([
'caseId' => $caseId,
'status!=' => ['Completed', 'Voided'],
])
->order('createdAt', 'DESC')
->limit(0, 10)
->find();
$requests = [];
foreach ($collection as $sr) {
$requests[] = [
'id' => $sr->get('id'),
'name' => $sr->get('name'),
'status' => $sr->get('status'),
'sentAt' => $sr->get('sentAt'),
];
}
return $requests;
} catch (\Exception $e) {
// SignatureRequest entity might not exist if DigitalSignature not installed
return [];
}
}
private function getDocumentListing(string $caseId): array private function getDocumentListing(string $caseId): array
{ {
try { try {
$analyzer = $this->injectableFactory->create(DocumentAnalyzer::class); $analyzer = $this->injectableFactory->create(DocumentAnalyzer::class);
$result = $analyzer->listDocumentsRecursive($caseId); $result = $analyzer->listDocumentsRecursive($caseId);
$simplified = ['totalFiles' => $result['totalFiles'], 'folders' => [], 'rootFiles' => []]; $simplified = [
'totalFiles' => $result['totalFiles'],
'folders' => [],
'rootFiles' => [],
'resolvedPath' => $result['resolvedPath'] ?? null,
'pathSource' => $result['pathSource'] ?? null,
'folderExists' => $result['folderExists'] ?? null,
'diagnostic' => $result['reason'] ?? null,
];
foreach ($result['folders'] as $folder) { foreach ($result['folders'] as $folder) {
$files = []; $files = [];
@@ -41,11 +41,67 @@ class DocumentAnalyzer
public function listDocumentsRecursive(string $caseId): array public function listDocumentsRecursive(string $caseId): array
{ {
$casePath = $this->getCaseFolderPath($caseId); $description = $this->describeCaseFolderPath($caseId);
if (!$casePath) return ['casePath' => null, 'totalFiles' => 0, 'folders' => [], 'rootFiles' => []]; $casePath = $description['path'];
if (!$casePath) {
return [
'casePath' => null,
'totalFiles' => 0,
'folders' => [],
'rootFiles' => [],
'resolvedPath' => null,
'pathSource' => $description['source'],
'folderExists' => false,
'reason' => $description['error'] ?? 'Case has no networkStorageFolderPath and no default could be computed.',
];
}
$client = $this->getStorageClient(); $client = $this->getStorageClient();
try {
$folderExists = $client->exists($casePath);
} catch (\Exception $e) {
return [
'casePath' => $casePath,
'totalFiles' => 0,
'folders' => [],
'rootFiles' => [],
'resolvedPath' => $casePath,
'pathSource' => $description['source'],
'folderExists' => false,
'reason' => "Storage client failed while checking '{$casePath}': " . $e->getMessage(),
];
}
if (!$folderExists) {
return [
'casePath' => $casePath,
'totalFiles' => 0,
'folders' => [],
'rootFiles' => [],
'resolvedPath' => $casePath,
'pathSource' => $description['source'],
'folderExists' => false,
'reason' => "Folder '{$casePath}' (source: {$description['source']}) does not exist on storage.",
];
}
try {
$topLevel = $client->listFolder($casePath); $topLevel = $client->listFolder($casePath);
} catch (\Exception $e) {
return [
'casePath' => $casePath,
'totalFiles' => 0,
'folders' => [],
'rootFiles' => [],
'resolvedPath' => $casePath,
'pathSource' => $description['source'],
'folderExists' => true,
'reason' => "Failed to list folder '{$casePath}': " . $e->getMessage(),
];
}
$folders = []; $rootFiles = []; $totalFiles = 0; $folders = []; $rootFiles = []; $totalFiles = 0;
foreach ($topLevel as $item) { foreach ($topLevel as $item) {
@@ -70,7 +126,16 @@ class DocumentAnalyzer
} }
} }
return ['casePath' => $casePath, 'totalFiles' => $totalFiles, 'folders' => $folders, 'rootFiles' => $rootFiles]; return [
'casePath' => $casePath,
'totalFiles' => $totalFiles,
'folders' => $folders,
'rootFiles' => $rootFiles,
'resolvedPath' => $casePath,
'pathSource' => $description['source'],
'folderExists' => true,
'reason' => $totalFiles === 0 ? "Folder '{$casePath}' exists but is empty." : null,
];
} }
public function extractTextContent(string $filePath): ?string public function extractTextContent(string $filePath): ?string
@@ -115,6 +180,85 @@ class DocumentAnalyzer
return $text; return $text;
} }
/**
* Fetch raw file bytes + metadata for OCR/vision fallback.
*
* @return array{success: bool, fileName: string, mimeType: string, sizeBytes: int, base64: ?string, error: ?string}
*/
public function getDocumentBytes(string $filePath): array
{
$fileName = basename($filePath);
$client = $this->getStorageClient();
try {
$content = $client->downloadFile($filePath);
} catch (\Exception $e) {
return [
'success' => false,
'fileName' => $fileName,
'mimeType' => '',
'sizeBytes' => 0,
'base64' => null,
'error' => 'Failed to download: ' . $e->getMessage(),
];
}
$sizeBytes = strlen($content);
if ($sizeBytes > self::MAX_FILE_SIZE) {
return [
'success' => false,
'fileName' => $fileName,
'mimeType' => '',
'sizeBytes' => $sizeBytes,
'base64' => null,
'error' => 'File too large (max 10MB).',
];
}
$mimeType = $this->guessMimeType($fileName, $content);
return [
'success' => true,
'fileName' => $fileName,
'mimeType' => $mimeType,
'sizeBytes' => $sizeBytes,
'base64' => base64_encode($content),
'error' => null,
];
}
private function guessMimeType(string $fileName, string $content): string
{
$ext = strtolower(pathinfo($fileName, PATHINFO_EXTENSION));
$map = [
'pdf' => 'application/pdf',
'docx' => 'application/vnd.openxmlformats-officedocument.wordprocessingml.document',
'doc' => 'application/msword',
'txt' => 'text/plain',
'csv' => 'text/csv',
'log' => 'text/plain',
'png' => 'image/png',
'jpg' => 'image/jpeg',
'jpeg' => 'image/jpeg',
'gif' => 'image/gif',
'bmp' => 'image/bmp',
'tiff' => 'image/tiff',
'tif' => 'image/tiff',
'webp' => 'image/webp',
];
if (isset($map[$ext])) {
return $map[$ext];
}
// Fallback: finfo
if (function_exists('finfo_buffer')) {
$finfo = finfo_open(FILEINFO_MIME_TYPE);
$mime = finfo_buffer($finfo, $content);
finfo_close($finfo);
if ($mime) return $mime;
}
return 'application/octet-stream';
}
public function extractMultipleDocuments(array $filePaths, int $maxPerFile = 50000): array public function extractMultipleDocuments(array $filePaths, int $maxPerFile = 50000): array
{ {
$results = []; $results = [];
@@ -192,22 +336,158 @@ class DocumentAnalyzer
} }
public function getCaseFolderPath(string $caseId): ?string public function getCaseFolderPath(string $caseId): ?string
{
return $this->describeCaseFolderPath($caseId)['path'];
}
/**
* Resolve the case folder path with diagnostic metadata.
*
* @return array{path: ?string, source: string, error: ?string, caseExists: bool}
* source: 'stored' | 'computed' | 'none'
*/
public function describeCaseFolderPath(string $caseId): array
{ {
$case = $this->entityManager->getEntityById('Case', $caseId); $case = $this->entityManager->getEntityById('Case', $caseId);
if (!$case) return null; if (!$case) {
return ['path' => null, 'source' => 'none', 'error' => "Case '{$caseId}' not found.", 'caseExists' => false];
}
// Try stored path first
$storedPath = $case->get('networkStorageFolderPath') ?: null; $storedPath = $case->get('networkStorageFolderPath') ?: null;
if ($storedPath) return $storedPath; if ($storedPath) {
return ['path' => $storedPath, 'source' => 'stored', 'error' => null, 'caseExists' => true];
}
// Fall back to computed path from NetworkDocumentService
try { try {
$nds = $this->getNetworkDocumentService(); $nds = $this->getNetworkDocumentService();
return $nds->getEntityFolderPath('Case', $caseId); $computed = $nds->getEntityFolderPath('Case', $caseId);
} catch (\Exception $e) { if ($computed) {
$this->log->warning("SmartAssistant: Failed to get case folder path: " . $e->getMessage()); return ['path' => $computed, 'source' => 'computed', 'error' => null, 'caseExists' => true];
return null;
} }
return ['path' => null, 'source' => 'none', 'error' => 'No stored path, and default path computation returned null.', 'caseExists' => true];
} catch (\Exception $e) {
$this->log->warning("SmartAssistant: Failed to compute case folder path: " . $e->getMessage());
return ['path' => null, 'source' => 'none', 'error' => $e->getMessage(), 'caseExists' => true];
}
}
/**
* List the immediate contents of an arbitrary folder path on network storage.
* Manual override for when auto-resolution of a case folder fails.
*
* @return array<int, array{name: string, path: string, type: string, size: ?int, mimeType: ?string, modified: ?string}>
*/
public function browseFolder(string $path): array
{
$client = $this->getStorageClient();
$entries = [];
foreach ($client->listFolder($path) as $item) {
$isFolder = ($item['type'] ?? null) === 'folder' || !empty($item['isFolder']);
$entries[] = [
'name' => $item['name'] ?? '',
'path' => $item['path'] ?? '',
'type' => $isFolder ? 'folder' : 'file',
'size' => $item['size'] ?? null,
'mimeType' => $item['mimeType'] ?? null,
'modified' => $item['modified'] ?? null,
];
}
return $entries;
}
/**
* Search the storage for folders matching a query (case-insensitive substring).
* Scans root and one level deep. Returns sorted matches.
*
* @return array<int, array{path: string, name: string, level: int, score: int}>
*/
public function findCaseFolder(string $query, int $limit = 20): array
{
$normalized = trim($query);
if ($normalized === '') return [];
$client = $this->getStorageClient();
$needle = mb_strtolower($normalized);
$matches = [];
try {
$roots = $client->listFolder('');
} catch (\Exception $e) {
$this->log->warning("SmartAssistant: findCaseFolder root listing failed: " . $e->getMessage());
return [];
}
foreach ($roots as $item) {
$isFolder = ($item['type'] ?? null) === 'folder' || !empty($item['isFolder']);
if (!$isFolder) continue;
$name = (string) ($item['name'] ?? '');
$path = (string) ($item['path'] ?? $name);
$score = $this->fuzzyScore($name, $needle);
if ($score > 0) {
$matches[] = ['path' => $path, 'name' => $name, 'level' => 0, 'score' => $score];
}
// One level deeper, but only inspect folder names — don't list files to keep it cheap.
try {
foreach ($client->listFolder($path) as $sub) {
$subIsFolder = ($sub['type'] ?? null) === 'folder' || !empty($sub['isFolder']);
if (!$subIsFolder) continue;
$subName = (string) ($sub['name'] ?? '');
$subPath = (string) ($sub['path'] ?? ($path . '/' . $subName));
$subScore = $this->fuzzyScore($subName, $needle);
if ($subScore > 0) {
$matches[] = ['path' => $subPath, 'name' => $subName, 'level' => 1, 'score' => $subScore];
}
}
} catch (\Exception $e) {
// Non-fatal: skip unreadable subfolders silently.
}
}
usort($matches, fn($a, $b) => $b['score'] <=> $a['score']);
return array_slice($matches, 0, $limit);
}
/**
* Persist the resolved folder path on the Case so future calls skip discovery.
*/
public function setCaseFolderPath(string $caseId, string $path): array
{
$case = $this->entityManager->getEntityById('Case', $caseId);
if (!$case) {
throw new Error("Case '{$caseId}' not found.");
}
$client = $this->getStorageClient();
if (!$client->exists($path)) {
throw new Error("Path '{$path}' does not exist on storage.");
}
$case->set('networkStorageFolderPath', $path);
$this->entityManager->saveEntity($case);
return ['success' => true, 'caseId' => $caseId, 'path' => $path];
}
private function fuzzyScore(string $haystack, string $needleLower): int
{
$hay = mb_strtolower($haystack);
if ($hay === $needleLower) return 100;
if (mb_strpos($hay, $needleLower) !== false) return 80;
// Token overlap (whitespace/dash/underscore).
$hayTokens = preg_split('/[\s\-_]+/u', $hay) ?: [];
$needleTokens = preg_split('/[\s\-_]+/u', $needleLower) ?: [];
$hits = 0;
foreach ($needleTokens as $nt) {
if ($nt === '') continue;
foreach ($hayTokens as $ht) {
if ($ht !== '' && mb_strpos($ht, $nt) !== false) { $hits++; break; }
}
}
return $hits > 0 ? 40 + $hits * 5 : 0;
} }
private function getMaxTextLength(): int private function getMaxTextLength(): int
+2 -2
View File
@@ -3,11 +3,11 @@
"module": "SmartAssistant", "module": "SmartAssistant",
"description": "Unified AI Assistant for Legal CRM — floating chat with case memory, office alerts, and AI Gateway integration", "description": "Unified AI Assistant for Legal CRM — floating chat with case memory, office alerts, and AI Gateway integration",
"author": "klear", "author": "klear",
"version": "2.7.3", "version": "2.8.0",
"acceptableVersions": [ "acceptableVersions": [
">=8.0.0" ">=8.0.0"
], ],
"releaseDate": "2026-04-14", "releaseDate": "2026-04-19",
"php": [ "php": [
">=8.1" ">=8.1"
] ]