iatoolkit 0.66.2__py3-none-any.whl → 0.71.2__py3-none-any.whl

iatoolkit/__init__.py

@@ -5,20 +5,18 @@ IAToolkit Package
 # Expose main classes and functions at the top level of the package
 
 # main IAToolkit class
-from .iatoolkit import IAToolkit, current_iatoolkit, create_app
+from .iatoolkit import IAToolkit, create_app, current_iatoolkit
 
 # for registering the client companies
 from .company_registry import register_company
 from .base_company import BaseCompany
-from iatoolkit.repositories.database_manager import DatabaseManager
 
 # --- Services ---
 from iatoolkit.services.query_service import QueryService
-from iatoolkit.services.sql_service import SqlService
 from iatoolkit.services.document_service import DocumentService
 from iatoolkit.services.search_service import SearchService
+from iatoolkit.services.sql_service import SqlService
 from iatoolkit.services.load_documents_service import LoadDocumentsService
-from iatoolkit.services.excel_service import ExcelService
 from iatoolkit.infra.call_service import CallServiceClient
 
 __all__ = [
@@ -27,10 +25,8 @@ __all__ = [
     'current_iatoolkit',
     'register_company',
     'BaseCompany',
-    'DatabaseManager',
     'QueryService',
     'SqlService',
-    'ExcelService',
     'DocumentService',
     'SearchService',
     'LoadDocumentsService',

iatoolkit/base_company.py

@@ -7,10 +7,9 @@
 from abc import ABC, abstractmethod
 from iatoolkit.repositories.profile_repo import ProfileRepo
 from iatoolkit.repositories.llm_query_repo import LLMQueryRepo
-
 from iatoolkit.services.prompt_manager_service import PromptService
 from iatoolkit.repositories.models import Company, Function, PromptCategory
-from iatoolkit import IAToolkit
+from .iatoolkit import IAToolkit
 
 
 class BaseCompany(ABC):
@@ -21,23 +20,16 @@ class BaseCompany(ABC):
         self.llm_query_repo: LLMQueryRepo = injector.get(LLMQueryRepo)
         self.prompt_service: PromptService = injector.get(PromptService)
         self.company: Company | None = None
-
-    def _load_company_by_short_name(self, short_name: str) -> Company:
-        self.company = self.profile_repo.get_company_by_short_name(short_name)
-        return self.company
+        self.company_short_name: str
 
     def _create_company(self,
                         short_name: str,
                         name: str,
                         parameters: dict | None = None,
-                        branding: dict | None = None,
-                        onboarding_cards: dict | None = None,
                         ) -> Company:
         company_obj = Company(short_name=short_name,
                               name=name,
-                              parameters=parameters,
-                              branding=branding,
-                              onboarding_cards=onboarding_cards)
+                              parameters=parameters)
         self.company = self.profile_repo.create_company(company_obj)
         return self.company
 
@@ -77,17 +69,6 @@ class BaseCompany(ABC):
             **kwargs
         )
 
-
-    @abstractmethod
-    # initialize all the database tables  needed
-    def register_company(self):
-        raise NotImplementedError("La subclase debe implementar el método create_company()")
-
-    @abstractmethod
-    # get context specific for this company
-    def get_company_context(self, **kwargs) -> str:
-        raise NotImplementedError("La subclase debe implementar el método get_company_context()")
-
     @abstractmethod
     # get context specific for this company
     def get_user_info(self, user_identifier: str) -> dict:
@@ -98,15 +79,6 @@ class BaseCompany(ABC):
     def handle_request(self, tag: str, params: dict) -> dict:
         raise NotImplementedError("La subclase debe implementar el método handle_request()")
 
-    @abstractmethod
-    # get context specific for the query
-    def start_execution(self):
-        raise NotImplementedError("La subclase debe implementar el método start_execution()")
-
-    @abstractmethod
-    # get context specific for the query
-    def get_metadata_from_filename(self, filename: str) -> dict:
-        raise NotImplementedError("La subclase debe implementar el método get_query_context()")
 
     def register_cli_commands(self, app):
         """

iatoolkit/cli_commands.py

@@ -5,7 +5,7 @@
 
 import click
 import logging
-from iatoolkit import IAToolkit
+from .iatoolkit import IAToolkit
 from iatoolkit.services.profile_service import ProfileService
 
 def register_core_commands(app):

iatoolkit/common/routes.py

@@ -26,7 +26,7 @@ def register_views(injector, app):
     from iatoolkit.views.history_api_view import HistoryApiView
     from iatoolkit.views.help_content_api_view import HelpContentApiView
     from iatoolkit.views.profile_api_view import UserLanguageApiView  # <-- Importa la nueva vista
-
+    from iatoolkit.views.embedding_api_view import EmbeddingApiView
     from iatoolkit.views.login_view import LoginView, FinalizeContextView
     from iatoolkit.views.external_login_view import ExternalLoginView, RedeemTokenApiView
     from iatoolkit.views.logout_api_view import LogoutApiView
@@ -99,6 +99,10 @@ def register_views(injector, app):
     # this endpoint is for upload documents into the vector store (api-key)
     app.add_url_rule('/api/load', view_func=FileStoreApiView.as_view('load_api'))
 
+    # this endpoint is for generating embeddings for a given text
+    app.add_url_rule('/<company_short_name>/api/embedding',
+                     view_func=EmbeddingApiView.as_view('embedding_api'))
+
 
     @app.route('/download/<path:filename>')
     def download_file(filename):

iatoolkit/common/session_manager.py

@@ -3,6 +3,8 @@
 #
 # IAToolkit is open source software.
 
+# This is the Flask connected session manager for IAToolkit
+
 from flask import session
 
 class SessionManager:

iatoolkit/company_registry.py

@@ -31,10 +31,9 @@ class CompanyRegistry:
 
                     # save the created instance in the registry
                     self._company_instances[company_key] = company_instance
-                    logging.info(f"company '{company_key}' instantiated")
 
                 except Exception as e:
-                    logging.error(f"Error instanciando empresa {company_key}: {e}")
+                    logging.error(f"Error while creating company instance for {company_key}: {e}")
                     logging.exception(e)
                     raise
 

iatoolkit/iatoolkit.py

@@ -19,7 +19,7 @@ from werkzeug.middleware.proxy_fix import ProxyFix
 from injector import Binder, Injector, singleton
 from importlib.metadata import version as _pkg_version, PackageNotFoundError
 
-IATOOLKIT_VERSION = "0.66.2"
+IATOOLKIT_VERSION = "0.71.2"
 
 # global variable for the unique instance of IAToolkit
 _iatoolkit_instance: Optional['IAToolkit'] = None
@@ -91,6 +91,9 @@ class IAToolkit:
         # Step 6: initialize dispatcher and registered companies
         self._init_dispatcher_and_company_instances()
 
+        # Re-apply logging configuration in case it was modified by company-specific code
+        self._setup_logging()
+
         # Step 7: Finalize setup within the application context
         self._setup_redis_sessions()
         self._setup_cors()
@@ -155,19 +158,12 @@ class IAToolkit:
                          static_folder=static_folder,
                          template_folder=template_folder)
 
-        is_https = self._get_config_value('USE_HTTPS', 'false').lower() == 'true'
-        is_dev = self._get_config_value('FLASK_ENV') == 'development'
-
-        # get the iatoolkit domain
-        parsed_url = urlparse(os.getenv('IATOOLKIT_BASE_URL'))
-        domain = parsed_url.netloc
-
+        # get the IATOOLKIT_VERSION from the package metadata
         try:
             self.version = _pkg_version("iatoolkit")
         except PackageNotFoundError:
             pass
 
-
         self.app.config.update({
             'VERSION': self.version,
             'SECRET_KEY': self._get_config_value('FLASK_SECRET_KEY', 'iatoolkit-default-secret'),
@@ -180,6 +176,7 @@ class IAToolkit:
             'JWT_EXPIRATION_SECONDS_CHAT': int(self._get_config_value('JWT_EXPIRATION_SECONDS_CHAT', 3600))
         })
 
+        parsed_url = urlparse(os.getenv('IATOOLKIT_BASE_URL'))
         if parsed_url.scheme == 'https':
             self.app.config['PREFERRED_URL_SCHEME'] = 'https'
 
@@ -187,7 +184,7 @@ class IAToolkit:
         self.app.wsgi_app = ProxyFix(self.app.wsgi_app, x_proto=1)
 
         # Configuración para tokenizers en desarrollo
-        if is_dev:
+        if self._get_config_value('FLASK_ENV') == 'dev':
             os.environ["TOKENIZERS_PARALLELISM"] = "false"
 
     def _setup_database(self):
@@ -293,7 +290,6 @@ class IAToolkit:
         from iatoolkit.repositories.document_repo import DocumentRepo
         from iatoolkit.repositories.profile_repo import ProfileRepo
         from iatoolkit.repositories.llm_query_repo import LLMQueryRepo
-
         from iatoolkit.repositories.vs_repo import VSRepo
         from iatoolkit.repositories.tasks_repo import TaskRepo
 
@@ -318,6 +314,8 @@ class IAToolkit:
         from iatoolkit.services.branding_service import BrandingService
         from iatoolkit.services.i18n_service import I18nService
         from iatoolkit.services.language_service import LanguageService
+        from iatoolkit.services.configuration_service import ConfigurationService
+        from iatoolkit.services.embedding_service import EmbeddingService
 
         binder.bind(QueryService, to=QueryService)
         binder.bind(TaskService, to=TaskService)
@@ -333,6 +331,8 @@ class IAToolkit:
         binder.bind(BrandingService, to=BrandingService)
         binder.bind(I18nService, to=I18nService)
         binder.bind(LanguageService, to=LanguageService)
+        binder.bind(ConfigurationService, to=ConfigurationService)
+        binder.bind(EmbeddingService, to=EmbeddingService)
 
     def _bind_infrastructure(self, binder: Binder):
         from iatoolkit.infra.llm_client import llmClient
@@ -359,9 +359,9 @@ class IAToolkit:
         # instantiate all the registered companies
         get_company_registry().instantiate_companies(self._injector)
 
-        # use the dispatcher to start the execution of every company
+        # use the dispatcher to load the company config.yaml file and prepare the execution
         dispatcher = self._injector.get(Dispatcher)
-        dispatcher.start_execution()
+        dispatcher.load_company_configs()
 
     def _setup_cli_commands(self):
         from iatoolkit.cli_commands import register_core_commands

iatoolkit/infra/llm_client.py

@@ -38,12 +38,6 @@ class llmClient:
         self.util = util
         self._dispatcher = None # Cache for the lazy-loaded dispatcher
 
-        # get the model from the environment variable
-        self.model = os.getenv("LLM_MODEL", "")
-        if not self.model:
-            raise IAToolkitException(IAToolkitException.ErrorType.API_KEY,
-                               "La variable de entorno 'LLM_MODEL' no está configurada.")
-
         # library for counting tokens
         self.encoding = tiktoken.encoding_for_model("gpt-4o")
 
@@ -70,6 +64,7 @@ class llmClient:
                context: str,
                tools: list[dict],
                text: dict,
+               model: str,
                context_history: Optional[List[Dict]] = None,
                ) -> dict:
 
@@ -80,13 +75,13 @@ class llmClient:
         force_tool_name = None
         reasoning = {}
 
-        if 'gpt-5' in self.model:
+        if 'gpt-5' in model:
             text['verbosity'] = "low"
             reasoning = {"effort": 'minimal'}
 
         try:
             start_time = time.time()
-            logging.info(f"calling llm model '{self.model}' with {self.count_tokens(context)} tokens...")
+            logging.info(f"calling llm model '{model}' with {self.count_tokens(context)} tokens...")
 
             # get the proxy for the company
             llm_proxy = self.llm_proxy_factory.create_for_company(company)
@@ -99,7 +94,7 @@ class llmClient:
                 }]
 
                 response = llm_proxy.create_response(
-                    model=self.model,
+                    model=model,
                     previous_response_id=previous_response_id,
                     context_history=context_history,
                     input=input_messages,
@@ -136,7 +131,7 @@ class llmClient:
                     logging.info(f"start execution fcall: {function_name}")
                     try:
                         result = self.dispatcher.dispatch(
-                            company_name=company.short_name,
+                            company_short_name=company.short_name,
                             action=function_name,
                             **args
                         )
@@ -186,7 +181,7 @@ class llmClient:
                     tool_choice_value = "required"
 
                 response = llm_proxy.create_response(
-                    model=self.model,
+                    model=model,
                     input=input_messages,
                     previous_response_id=response.id,
                     context_history=context_history,
@@ -200,7 +195,7 @@ class llmClient:
             # save the statistices
             stats['response_time']=int(time.time() - start_time)
             stats['sql_retry_count'] = sql_retry_count
-            stats['model'] = response.model
+            stats['model'] = model
 
             # decode the LLM response
             decoded_response = self.decode_response(response)
@@ -392,6 +387,7 @@ class llmClient:
 
     def add_stats(self, stats1: dict, stats2: dict) -> dict:
         stats_dict = {
+            "model": stats1.get('model', ''),
             "input_tokens": stats1.get('input_tokens', 0) + stats2.get('input_tokens', 0),
             "output_tokens": stats1.get('output_tokens', 0) + stats2.get('output_tokens', 0),
             "total_tokens": stats1.get('total_tokens', 0) + stats2.get('total_tokens', 0),

iatoolkit/infra/llm_proxy.py

@@ -7,6 +7,7 @@ from typing import Dict, List, Any
 from abc import ABC, abstractmethod
 from iatoolkit.common.util import Utility
 from iatoolkit.infra.llm_response import LLMResponse
+from iatoolkit.services.configuration_service import ConfigurationService
 from iatoolkit.infra.openai_adapter import OpenAIAdapter
 from iatoolkit.infra.gemini_adapter import GeminiAdapter
 from iatoolkit.common.exceptions import IAToolkitException
@@ -41,12 +42,16 @@ class LLMProxy:
     _clients_cache_lock = threading.Lock()
 
     @inject
-    def __init__(self, util: Utility, openai_client = None, gemini_client = None):
+    def __init__(self, util: Utility,
+                 configuration_service: ConfigurationService,
+                 openai_client = None,
+                 gemini_client = None):
         """
         Inicializa una instancia del proxy. Puede ser una instancia "base" (fábrica)
         o una instancia de "trabajo" con clientes configurados.
         """
         self.util = util
+        self.configuration_service = configuration_service
         self.openai_adapter = OpenAIAdapter(openai_client) if openai_client else None
         self.gemini_adapter = GeminiAdapter(gemini_client) if gemini_client else None
 
@@ -71,7 +76,11 @@ class LLMProxy:
             )
 
         # Devuelve una NUEVA instancia con los clientes configurados
-        return LLMProxy(util=self.util, openai_client=openai_client, gemini_client=gemini_client)
+        return LLMProxy(
+                    util=self.util,
+                    configuration_service=self.configuration_service,
+                    openai_client=openai_client,
+                    gemini_client=gemini_client)
 
     def create_response(self, model: str, input: List[Dict], **kwargs) -> LLMResponse:
         """Enruta la llamada al adaptador correcto basado en el modelo."""
@@ -103,7 +112,7 @@ class LLMProxy:
                     elif provider == LLMProvider.GEMINI:
                         client = self._create_gemini_client(company)
                     else:
-                        raise IAToolkitException(f"Proveedor no soportado: {provider.value}")
+                        raise IAToolkitException(f"provider not supported: {provider.value}")
 
                     if client:
                         LLMProxy._clients_cache[cache_key] = client
@@ -115,22 +124,41 @@ class LLMProxy:
 
     def _create_openai_client(self, company: Company) -> OpenAI:
         """Crea un cliente de OpenAI con la API key."""
-        if company.openai_api_key:
-            decrypted_api_key = self.util.decrypt_key(company.openai_api_key)
+        decrypted_api_key = ''
+        llm_config = self.configuration_service.get_configuration(company.short_name, 'llm')
+
+        # Try to get API key name from config first
+        if llm_config and llm_config.get('api-key'):
+            api_key_env_var = llm_config['api-key']
+            decrypted_api_key = os.getenv(api_key_env_var, '')
         else:
-            decrypted_api_key = os.getenv("OPENAI_API_KEY", '')
+            # Fallback to old logic
+            if company.openai_api_key:
+                decrypted_api_key = self.util.decrypt_key(company.openai_api_key)
+            else:
+                decrypted_api_key = os.getenv("OPENAI_API_KEY", '')
+
         if not decrypted_api_key:
             raise IAToolkitException(IAToolkitException.ErrorType.API_KEY,
-                               f"La empresa '{company.name}' no tiene API key de OpenAI.")
+                                     f"La empresa '{company.name}' no tiene API key de OpenAI.")
         return OpenAI(api_key=decrypted_api_key)
 
     def _create_gemini_client(self, company: Company) -> Any:
         """Configura y devuelve el cliente de Gemini."""
 
-        if company.gemini_api_key:
-            decrypted_api_key = self.util.decrypt_key(company.gemini_api_key)
+        decrypted_api_key = ''
+        llm_config = self.configuration_service.get_configuration(company.short_name, 'llm')
+
+        # Try to get API key name from config first
+        if llm_config and llm_config.get('api-key'):
+            api_key_env_var = llm_config['api-key']
+            decrypted_api_key = os.getenv(api_key_env_var, '')
         else:
-            decrypted_api_key = os.getenv("GEMINI_API_KEY", '')
+            # Fallback to old logic
+            if company.gemini_api_key:
+                decrypted_api_key = self.util.decrypt_key(company.gemini_api_key)
+            else:
+                decrypted_api_key = os.getenv("GEMINI_API_KEY", '')
 
         if not decrypted_api_key:
             return None

iatoolkit/locales/en.yaml

@@ -77,6 +77,8 @@ ui:
     stop: "Stop"
 
 errors:
+  company_not_found: "The company {company_short_name} does not exist."
+  timeout: "timeout expired."
   auth:
     invalid_password: "The provided password is incorrect."
     user_not_found: "A user with that email address was not found."
@@ -87,7 +89,7 @@ errors:
     no_user_identifier_api: "No user_identifier provided for API call."
   templates:
     company_not_found: "Company not found."
-    home_template_not_found: "The home page template for the company '{company_name}' is not configured."
+    home_template_not_found: "The home page template for the company '{company_short_name}' is not configured."
     template_not_found: "Template not found: '{template_name}'."
 
     processing_error: "An error occurred while processing the custom home page template: {error}"
@@ -95,7 +97,7 @@ errors:
     unexpected_error: "An unexpected error has occurred. Please contact support."
     unsupported_language: "The selected language is not supported."
   signup:
-    company_not_found: "The company {company_name} does not exist."
+    company_not_found: "The company {company_short_name} does not exist."
     incorrect_password_for_existing_user: "The password for the user {email} is incorrect."
     user_already_registered: "The user with email '{email}' already exists in this company."
     password_mismatch: "The passwords do not match. Please try again."
@@ -115,9 +117,29 @@ errors:
     password_no_digit: "Password must contain at least one number."
     password_no_special_char: "Password must contain at least one special character."
 
+  services:
+    no_text_file: "The file is not text or the encoding is not UTF-8"
+    no_output_file: "Missing output file name"
+    no_data_for_excel: "Missing data or it is not a list of dictionaries"
+    no_download_directory: "Temporary directory for saving Excel files is not configured"
+    cannot_create_excel: "Could not create the Excel file"
+    invalid_filename: "Invalid filename"
+    file_not_exist: "File not found"
+    path_is_not_a_file: "The path does not correspond to a file"
+    file_validation_error: "Error validating file"
+    user_not_authorized: "user is not authorized for this company"
+    account_not_verified: "Your account has not been verified. Please check your email."
+    missing_response_id: "Can not found 'previous_response_id' for '{company_short_name}/{user_identifier}'. Reinit context"
+
+
 api_responses:
   context_reloaded_success: "The context has been successfully reloaded."
 
+services:
+  mail_sent: "Email sent successfully."
+  start_query: "Hello, what can I help you with today?"
+
+
 flash_messages:
   password_changed_success: "Your password has been successfully reset. You can now log in."
   login_required: "Please log in to continue."
@@ -142,3 +164,4 @@ js_messages:
   unknown_server_error: "Unknown server error."
   loading: "Loading..."
   reload_init: "init reloading context in background..."
+  no_history_found: "No query history found."

iatoolkit/locales/es.yaml

@@ -73,6 +73,10 @@
         stop: "Detener"
 
     errors:
+      company_not_found: "La empresa {company_short_name} no existe."
+      timeout: "El tiempo de espera ha expirado."
+
+
       auth:
         invalid_password: "La contraseña proporcionada es incorrecta."
         user_not_found: "No se encontró un usuario con ese correo."
@@ -83,15 +87,15 @@
         no_user_identifier_api: "No se proporcionó user_identifier para la llamada a la API."
       templates:
         company_not_found: "Empresa no encontrada."
-        home_template_not_found: "La plantilla de la página de inicio para la empresa '{company_name}' no está configurada."
+        home_template_not_found: "La plantilla de la página de inicio para la empresa '{company_short_name}' no está configurada."
+        processing_error: "Error al procesar el template: {error}"
         template_not_found: "No se encontro el template: '{template_name}'."
-        processing_error: "Ocurrió un error al procesar la plantilla personalizada de la página de inicio: {error}"
 
       general:
-        unexpected_error: "Ha ocurrido un error inesperado. Por favor, contacta a soporte."
+        unexpected_error: "Ha ocurrido un error inesperado: {error}."
         unsupported_language: "El idioma seleccionado no es válido."
       signup:
-        company_not_found: "La empresa {company_name} no existe."
+        company_not_found: "La empresa {company_short_name} no existe."
         incorrect_password_for_existing_user: "La contraseña para el usuario {email} es incorrecta."
         user_already_registered: "El usuario con email '{email}' ya existe en esta empresa."
         password_mismatch: "Las contraseñas no coinciden. Por favor, inténtalo de nuevo."
@@ -111,9 +115,27 @@
         password_no_digit: "La contraseña debe tener al menos un número."
         password_no_special_char: "La contraseña debe tener al menos un carácter especial."
 
+      services:
+        no_text_file: "El archivo no es texto o la codificación no es UTF-8"
+        no_output_file: "falta el nombre del archivo de salida"
+        no_data_for_excel: "faltan los datos  o no es una lista de diccionarios"
+        no_download_directory: "no esta configurado el directorio temporal para guardar excels"
+        cannot_create_excel: "no se pudo crear el archivo excel"
+        invalid_filename: "Nombre de archivo inválido"
+        file_not_exist : "Archivo no encontrado"
+        path_is_not_a_file : "La ruta no corresponde a un archivo"
+        file_validation_error : "Error validando archivo"
+        user_not_authorized: "Usuario no esta autorizado para esta empresa"
+        account_not_verified: "Tu cuenta no ha sido verificada. Por favor, revisa tu correo."
+        missing_response_id: "No se encontró 'previous_response_id' para '{company_short_name}/{user_identifier}'. Reinicia el contexto."
+
     api_responses:
       context_reloaded_success: "El contexto se ha recargado con éxito."
 
+    services:
+      mail_sent: "mail enviado exitosamente."
+      start_query: "Hola, cual es tu pregunta?"
+
     flash_messages:
       password_changed_success: "Tu contraseña ha sido restablecida exitosamente. Ahora puedes iniciar sesión."
       signup_success: "Registro exitoso. Por favor, revisa tu correo para verificar tu cuenta."
@@ -137,4 +159,5 @@
       unknown_server_error: "Error desconocido del servidor."
       loading: "Cargando..."
       reload_init: "Iniciando recarga de contexto en segundo plano..."
+      no_history_found: "No existe historial de consultas."
 

iatoolkit/repositories/database_manager.py

@@ -27,8 +27,8 @@ class DatabaseManager:
             self._engine = create_engine(
                 database_url,
                 echo=False,
-                pool_size=2,  # per worker
-                max_overflow=3,
+                pool_size=10,  # per worker
+                max_overflow=20,
                 pool_timeout=30,
                 pool_recycle=1800,
                 pool_pre_ping=True,
@@ -70,6 +70,11 @@ class DatabaseManager:
     def remove_session(self):
         self.scoped_session.remove()
 
+    def get_all_table_names(self) -> list[str]:
+        # Returns a list of all table names in the database
+        inspector = inspect(self._engine)
+        return inspector.get_table_names()
+
     def get_table_schema(self,
                          table_name: str,
                          schema_name: str | None = None,
@@ -77,7 +82,7 @@ class DatabaseManager:
         inspector = inspect(self._engine)
 
         if table_name not in inspector.get_table_names():
-            raise RuntimeError(f"La tabla '{table_name}' no existe en la BD.")
+            raise RuntimeError(f"Table '{table_name}' does not exist.")
 
         if exclude_columns is None:
             exclude_columns = []

iatoolkit/repositories/document_repo.py

@@ -22,7 +22,7 @@ class DocumentRepo:
     def get(self, company_id, filename: str ) -> Document:
         if not company_id or not filename:
             raise IAToolkitException(IAToolkitException.ErrorType.PARAM_NOT_FILLED,
-                               'Falta empresa o filename')
+                               'missing company_id or filename')
 
         return self.session.query(Document).filter_by(company_id=company_id, filename=filename).first()
 

iatoolkit/repositories/models.py

@@ -53,17 +53,12 @@ class Company(Base):
     id = Column(Integer, primary_key=True)
     short_name = Column(String(20), nullable=False, unique=True, index=True)
     name = Column(String(256), nullable=False)
-    default_language = Column(String(5), nullable=False, default='es')
 
     # encrypted api-key
     openai_api_key = Column(String, nullable=True)
     gemini_api_key = Column(String, nullable=True)
-
-    branding = Column(JSON, nullable=True)
-    onboarding_cards = Column(JSON, nullable=True)
     parameters = Column(JSON, nullable=True)
     created_at = Column(DateTime, default=datetime.now)
-    allow_jwt = Column(Boolean, default=True, nullable=True)
 
     documents = relationship("Document",
                              back_populates="company",
@@ -161,10 +156,10 @@ class Document(Base):
     company_id = Column(Integer, ForeignKey('iat_companies.id',
                     ondelete='CASCADE'), nullable=False)
     filename = Column(String(256), nullable=False, index=True)
-    content = Column(Text, nullable=False)
-    content_b64 = Column(Text, nullable=False)
     meta = Column(JSON, nullable=True)
     created_at = Column(DateTime, default=datetime.now)
+    content = Column(Text, nullable=False)
+    content_b64 = Column(Text, nullable=False)
 
     company = relationship("Company", back_populates="documents")
 
@@ -207,7 +202,10 @@ class VSDoc(Base):
     document_id = Column(Integer, ForeignKey('iat_documents.id',
                         ondelete='CASCADE'), nullable=False)
     text = Column(Text, nullable=False)
-    embedding = Column(Vector(384), nullable=False)  # Ajusta la dimensión si es necesario
+
+    # the size of this vector should be set depending on the embedding model used
+    # for OpenAI is 1536, and for huggingface is 384
+    embedding = Column(Vector(1536), nullable=False)
 
     company = relationship("Company", back_populates="vsdocs")
 

iatoolkit/repositories/profile_repo.py

@@ -74,10 +74,6 @@ class ProfileRepo:
         if company:
             if company.parameters != new_company.parameters:
                 company.parameters = new_company.parameters
-            if company.branding != new_company.branding:
-                company.branding = new_company.branding
-            if company.onboarding_cards != new_company.onboarding_cards:
-                company.onboarding_cards = new_company.onboarding_cards
         else:
             # Si la compañía no existe, la añade a la sesión.
             self.session.add(new_company)