LLM-Driven Extraction of Unstructured Data — Built for API Deployments & ETL Pipeline Workflows
Python ai-agentsdata-engineeringdocument-aigenerative-aiidpjson-extractionllmmcp-serverocrpdf-extractionprompt-engineeringstructured-output
from permissions.permission import IsOwner
from utils.user_context import UserContext
class IsOwnerOrOrganizationMember(IsOwner):
"""Permission that grants access if the user is the owner or belongs to the
same organization.
"""
def has_object_permission(self, request, view, obj):
# Check if the user is the owner via base class logic
if super().has_object_permission(request, view, obj):
return True
# If the object has a 'created_by' field, but the user isn't the owner,
# deny access
if obj.created_by:
return False
# Support legacy API keys where 'created_by' is None by matching the
# organization ID. This allows organization members to access API keys as per
# the existing behavior.
user_organization = UserContext.get_organization()
# Check organization ID for associated `api` or `pipeline`
related_obj = getattr(obj, "api", None) or getattr(obj, "pipeline", None)
if related_obj and related_obj.organization_id == user_organization.id:
return True
return False