97 lines
2.7 KiB
Python
97 lines
2.7 KiB
Python
|
"""Define domain ports for documents."""
|
||
|
|
|
||
|
|
from __future__ import annotations
|
||
|
|
|
||
|
|
from abc import ABC, abstractmethod
|
||
|
|
|
||
|
|
from .models import Chunk, Document, DocumentStatus, ParsedDocument
|
||
|
|
# Keep domain contracts explicit so adapters can swap implementations cleanly.
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
class DocumentRepository(ABC):
|
||
|
|
"""Provide the Document Repository repository implementation."""
|
||
|
|
@abstractmethod
|
||
|
|
def create(self, document: Document) -> Document:
|
||
|
|
"""Handle create for the Document Repository instance."""
|
||
|
|
pass
|
||
|
|
|
||
|
|
@abstractmethod
|
||
|
|
def update(self, document: Document) -> Document:
|
||
|
|
"""Handle update for the Document Repository instance."""
|
||
|
|
pass
|
||
|
|
|
||
|
|
@abstractmethod
|
||
|
|
def get(self, doc_id: str) -> Document | None:
|
||
|
|
"""Handle get for the Document Repository instance."""
|
||
|
|
pass
|
||
|
|
|
||
|
|
@abstractmethod
|
||
|
|
def list(self, limit: int | None = None) -> list[Document]:
|
||
|
|
"""Handle list for the Document Repository instance."""
|
||
|
|
pass
|
||
|
|
|
||
|
|
@abstractmethod
|
||
|
|
def update_status(
|
||
|
|
self,
|
||
|
|
doc_id: str,
|
||
|
|
status: DocumentStatus,
|
||
|
|
*,
|
||
|
|
error_message: str = "",
|
||
|
|
chunk_count: int | None = None,
|
||
|
|
summary: str | None = None,
|
||
|
|
summary_latency_ms: int | None = None,
|
||
|
|
parser_name: str | None = None,
|
||
|
|
index_name: str | None = None,
|
||
|
|
metadata: dict | None = None,
|
||
|
|
) -> Document | None:
|
||
|
|
"""Update status for the Document Repository instance."""
|
||
|
|
pass
|
||
|
|
|
||
|
|
|
||
|
|
class DocumentBinaryStore(ABC):
|
||
|
|
"""Provide the Document Binary Store store implementation."""
|
||
|
|
@abstractmethod
|
||
|
|
def save(
|
||
|
|
self,
|
||
|
|
*,
|
||
|
|
object_name: str,
|
||
|
|
data: bytes,
|
||
|
|
content_type: str,
|
||
|
|
metadata: dict[str, str] | None = None,
|
||
|
|
) -> None:
|
||
|
|
"""Handle save for the Document Binary Store instance."""
|
||
|
|
pass
|
||
|
|
|
||
|
|
@abstractmethod
|
||
|
|
def read(self, object_name: str) -> bytes:
|
||
|
|
"""Handle read for the Document Binary Store instance."""
|
||
|
|
pass
|
||
|
|
|
||
|
|
@abstractmethod
|
||
|
|
def delete(self, object_name: str) -> None:
|
||
|
|
"""Handle delete for the Document Binary Store instance."""
|
||
|
|
pass
|
||
|
|
|
||
|
|
|
||
|
|
class DocumentParser(ABC):
|
||
|
|
"""Provide the Document Parser parser."""
|
||
|
|
@abstractmethod
|
||
|
|
def parse(self, *, file_path: str, doc_id: str, doc_name: str) -> ParsedDocument:
|
||
|
|
"""Handle parse for the Document Parser instance."""
|
||
|
|
pass
|
||
|
|
|
||
|
|
|
||
|
|
class ChunkBuilder(ABC):
|
||
|
|
"""Provide the Chunk Builder builder."""
|
||
|
|
@abstractmethod
|
||
|
|
def build(
|
||
|
|
self,
|
||
|
|
*,
|
||
|
|
parsed_document: ParsedDocument,
|
||
|
|
regulation_type: str,
|
||
|
|
version: str,
|
||
|
|
) -> list[Chunk]:
|
||
|
|
"""Handle build for the Chunk Builder instance."""
|
||
|
|
pass
|