csc-cia-stne 0.1.27__py3-none-any.whl → 0.1.29__py3-none-any.whl

csc_cia_stne/__init__.py

@@ -16,6 +16,7 @@ from .web import web_screen
 from .wacess import Waccess
 from .gcp_bucket import GCPBucket
 from .ftp import FTP
+from .gcp_document_ai import GCPDocumentAIClient
 
 # Define os itens disponíveis para importação
 __all__ = [
@@ -30,6 +31,7 @@ __all__ = [
     "Provio",
     "Email", 
     "GoogleDrive",
+    "GCPDocumentAIClient"
     "Slack",
     "web_screen",
     "Waccess",

csc_cia_stne/gcp_document_ai.py

@@ -0,0 +1,100 @@
+from typing import Optional, Dict, Union
+from google.cloud import documentai_v1beta3 as documentai
+from google.oauth2 import service_account
+
+class GCPDocumentAIClient:
+    def __init__(self, credential_json: Optional[dict] = None, processor_id: Optional[str] = None) -> None:
+        """
+        Inicializa o cliente do Google Cloud Document AI.
+
+        Args:
+            credential_json (Optional[dict]): Dicionário contendo as credenciais do Google Cloud.
+            processor_id (Optional[str]): ID do processador do Document AI.
+        
+        Attributes:
+            credential_json (dict): Credenciais do Google Cloud
+            project_id (str): ID do projeto extraído das credenciais
+            location (str): Localização do processador (fixo: "us")
+            processor_id (str): ID do processador do Document AI
+            client (documentai.DocumentProcessorServiceClient): Cliente do Document AI
+            is_connected (bool): Status da conexão
+            error (str|None): Mensagem de erro se houver falha na inicialização
+        """
+        self.credential_json: dict = credential_json
+        self.project_id: str = self.credential_json.get("project_id")
+        self.location: str = "us"
+        self.processor_id: str = processor_id
+
+        try:
+            self.client: documentai.DocumentProcessorServiceClient = self._get_document_ai_client(self.credential_json)
+            self.is_connected: bool = True
+            self.error = None
+
+        except Exception as e:
+            error_msg = f"Erro ao inicializar o cliente do Document AI: {e}"
+            self.is_connected = False
+            self.error = error_msg
+
+    def _get_document_ai_client(self, credential_json: dict) -> documentai.DocumentProcessorServiceClient:
+        """
+        Cria e retorna o cliente do Document AI.
+
+        Args:
+            credential_json (dict): Dicionário contendo as credenciais do Google Cloud.
+
+        Returns:
+            documentai.DocumentProcessorServiceClient: Cliente autenticado do Document AI.
+        
+        Raises:
+            Exception: Se houver erro na autenticação ou inicialização do cliente.
+        """
+        try:
+            credential = service_account.Credentials.from_service_account_info(credential_json)
+        except Exception as e:
+            error_msg = f"Erro ao criar credenciais do Document AI: {e}"
+            raise Exception(error_msg)
+
+        return documentai.DocumentProcessorServiceClient(credentials=credential)
+
+    
+    def ler_documento(self, file_bytes: bytes, mime_type: str) -> Dict[str, Union[bool, str, documentai.Document, None]]:
+        """
+        Processa um documento PDF usando o Google Cloud Document AI para extrair texto.
+
+        Args:
+            file_bytes (bytes): Bytes do arquivo PDF a ser processado.
+            mime_type (str): Tipo MIME do arquivo (ex.: "application/pdf").
+
+        Returns:
+            Dict[str, Union[bool, str, documentai.Document, None]]: Resultado do processamento
+                - success (bool): True se o processamento foi bem-sucedido
+                - error (str|None): Mensagem de erro se houver falha
+                - data (documentai.Document|None): Documento processado pelo Document AI
+        
+        Example:
+            >>> client = GCPDocumentAIClient(creds, processor_id)
+            >>> with open("documento.pdf", "rb") as f:
+            ...     resultado = client.ler_documento(f.read(), "application/pdf")
+            >>> if resultado["success"]:
+            ...     texto = resultado["data"].text
+            ...     print(f"Texto extraído: {texto[:100]}...")
+        
+        Note:
+            - Utiliza o processador configurado no __init__
+            - Processa o documento completo enviado em file_bytes
+            - Para limitar páginas, use extrair_x_paginas_pdf antes desta função
+        """
+        try:
+            name = self.client.processor_path(self.project_id, self.location, self.processor_id)
+            request = documentai.ProcessRequest(
+                name=name,
+                raw_document=documentai.RawDocument(content=file_bytes, mime_type=mime_type)
+            )
+            
+            result = self.client.process_document(request=request)
+            return {"success": True, "error": None, "data": result.document}
+
+        except Exception as e:
+            error_msg = f"Erro ao processar o documento com o Document AI: {str(e)}"
+            return {"success": False, "error": error_msg, "data": None}
+

csc_cia_stne/google_drive.py

@@ -677,7 +677,7 @@ class GoogleDrive:
             request = self.service.files().export_media(
                 fileId=file.get("id"), mimeType=mimeType
             )
-            file_path = f"{path}{file["name"]}"
+            file_path = f"{path}{file['name']}"
             with open(file_path, "wb") as f:
                 f.write(request.execute())
             return {

{csc_cia_stne-0.1.27.dist-info → csc_cia_stne-0.1.29.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: csc_cia_stne
-Version: 0.1.27
+Version: 0.1.29
 Summary: Biblioteca do time CSC-CIA utilizada no desenvolvimento de RPAs
 License: MIT
 Keywords: karavela,csc,cia,stone,rpa,botcity,stne
@@ -17,9 +17,11 @@ Requires-Dist: zeep
 Requires-Dist: google-cloud-bigquery
 Requires-Dist: google-cloud-storage
 Requires-Dist: google-cloud-bigquery-storage
+Requires-Dist: google-auth
 Requires-Dist: google-auth-oauthlib
 Requires-Dist: google-auth-httplib2
 Requires-Dist: google-api-python-client
+Requires-Dist: google-cloud-documentai
 Requires-Dist: pyjwt
 Requires-Dist: PyYAML
 Requires-Dist: python-dotenv

{csc_cia_stne-0.1.27.dist-info → csc_cia_stne-0.1.29.dist-info}/RECORD

@@ -1,11 +1,12 @@
-csc_cia_stne/__init__.py,sha256=jwLhGpOwFCow_6cqzwLn31WcIrMzutMZtEQpLL4bQtM,2638
+csc_cia_stne/__init__.py,sha256=LITCLPqafF-VUV85wUvJ047ozLnfd82vw5AuuKXYi2s,2713
 csc_cia_stne/bc_correios.py,sha256=s2XjJ0iokMlcUv0mAy9saU3pc_G-6X8wltb_lFHIL6o,24717
 csc_cia_stne/bc_sta.py,sha256=S4EtkSEHP-wTMWRjmmSBH9XY5SnVQ1TwwZFSOE6tI2Q,29551
 csc_cia_stne/email.py,sha256=y4xyPAe6_Mga5Wf6qAsDzYgn0f-zf2KshfItlWe58z8,8481
 csc_cia_stne/ftp.py,sha256=eNkOUEXdw-9NYfuZjNo6Oh7EduD54g8N0cfD0LuOiTU,11516
 csc_cia_stne/gcp_bigquery.py,sha256=foq8azvvv_f7uikMDslX9RcUIrx7RAS-Sn0AGW0QFQc,7231
 csc_cia_stne/gcp_bucket.py,sha256=vMALWiW7IoBCuJAR8bUCpOV6BuBzI9AhRRk3b72OdMk,11515
-csc_cia_stne/google_drive.py,sha256=sfq2arBYrv8OLIfRJTnj067EPbem2gofwIfpbqJ475o,44500
+csc_cia_stne/gcp_document_ai.py,sha256=Dzlk7YR3M_LxE0sHn-Lxz-PA1NsUZN2hgY5PyUfs0IQ,4506
+csc_cia_stne/google_drive.py,sha256=7qwx4_RPEoSJgeVI02aLYNXA7o69_Z3qONvX5bfA4V0,44500
 csc_cia_stne/karavela.py,sha256=jJCYX43D49gGuzmwwK6bN9XVnv2dXdp9iHnnV5H1LMQ,4794
 csc_cia_stne/logger_json.py,sha256=CXxSCOFGMymDi8XE9SKnPKjW4D0wJLqDLnxqePS26i8,3187
 csc_cia_stne/logger_rich.py,sha256=fklgkBb4rblKQd7YZ3q-eWfhGg9eflO2k2-z4pGh_yo,5201
@@ -38,8 +39,8 @@ csc_cia_stne/utilitarios/web_screen/__init__.py,sha256=5QcOPXKd95SvP2DoZiHS0gaU6
 csc_cia_stne/utilitarios/web_screen/web_screen_abstract.py,sha256=PjL8Vgfj_JdKidia7RFyCkro3avYLQu4RZRos41sh3w,3241
 csc_cia_stne/utilitarios/web_screen/web_screen_botcity.py,sha256=Xi5YJjl2pcxlX3OimqcBWRNXZEpAE7asyUjDJ4Oho5U,12297
 csc_cia_stne/utilitarios/web_screen/web_screen_selenium.py,sha256=JLIcPJE9ZX3Pd6zG6oTRMqqUAY063UzLY3ReRlxmiSM,15581
-csc_cia_stne-0.1.27.dist-info/licenses/LICENCE,sha256=LPGMtgKki2C3KEZP7hDhA1HBrlq5JCHkIeStUCLEMx4,1073
-csc_cia_stne-0.1.27.dist-info/METADATA,sha256=MdHxPiFFEt6tkCVCgk3qnDqOyO4g4nybP43vHJ-5cl0,1464
-csc_cia_stne-0.1.27.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-csc_cia_stne-0.1.27.dist-info/top_level.txt,sha256=ldo7GVv3tQx5KJvwBzdZzzQmjPys2NDVVn1rv0BOF2Q,13
-csc_cia_stne-0.1.27.dist-info/RECORD,,
+csc_cia_stne-0.1.29.dist-info/licenses/LICENCE,sha256=LPGMtgKki2C3KEZP7hDhA1HBrlq5JCHkIeStUCLEMx4,1073
+csc_cia_stne-0.1.29.dist-info/METADATA,sha256=sN19T2lC0fc6kk15iwjcobbX_1v_w9jhHWpbpHIvFCA,1530
+csc_cia_stne-0.1.29.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+csc_cia_stne-0.1.29.dist-info/top_level.txt,sha256=ldo7GVv3tQx5KJvwBzdZzzQmjPys2NDVVn1rv0BOF2Q,13
+csc_cia_stne-0.1.29.dist-info/RECORD,,