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 __future__ import annotations
from enum import Enum
from unstract.core.data_models import ExecutionStatus as SharedExecutionStatus
# Alias shared ExecutionStatus to ensure consistency between backend and workers
ExecutionStatus = SharedExecutionStatus
class WorkflowExecutionMethod(Enum):
INSTANT = "INSTANT"
QUEUED = "QUEUED"
class SchemaType(Enum):
"""Possible types for workflow module's JSON schema.
Values:
src: Refers to the source module's schema
dest: Refers to the destination module's schema
"""
SRC = "src"
DEST = "dest"
class SchemaEntity(Enum):
"""Possible entities for workflow module's JSON schema.
Values:
file: Refers to schema for file based sources
api: Refers to schema for API based sources
db: Refers to schema for DB based destinations
"""
FILE = "file"
API = "api"
DB = "db"
class ColumnModes(Enum):
WRITE_JSON_TO_A_SINGLE_COLUMN = "Write JSON to a single column"
SPLIT_JSON_INTO_COLUMNS = "Split JSON into columns"
class AgentName(Enum):
UNSTRACT_DBWRITER = "Unstract/DBWriter"
class RuleLogic(Enum):
AND = "AND"
OR = "OR"
class TaskType(Enum):
FILE_PROCESSING = "FILE_PROCESSING"
FILE_PROCESSING_CALLBACK = "FILE_PROCESSING_CALLBACK"