Source code for lalandre_rag.retrieval.context.models

"""
Context Models
Clean separation between act-level metadata, document metadata, and content slices.
"""

from dataclasses import dataclass
from typing import Any, Dict, List, Optional




@dataclass
class DocumentMeta:
    """
    Document-level metadata for a retrieved slice.
    """

    source_kind: str  # subdivision | chunk
    subdivision_id: int
    subdivision_type: str
    sequence_order: int
    chunk_id: Optional[int] = None
    chunk_index: Optional[int] = None
    char_start: Optional[int] = None
    char_end: Optional[int] = None
    payload: Optional[Dict[str, Any]] = None





@dataclass
class ActContext:
    """
    Act-level metadata, shared across multiple slices.
    """

    act_id: int
    celex: str
    title: str
    act_type: str
    regulatory_level: Optional[str] = None  # L1, L2, L3
    url_eurlex: Optional[str] = None
    relations: Optional[List[Dict[str, Any]]] = None
    subjects: Optional[List[str]] = None
    adoption_date: Optional[str] = None
    force_date: Optional[str] = None





@dataclass
class ContextSlice:
    """
    A context slice used by the LLM, with explicit act + doc metadata separation.
    """

    content: str
    score: float
    act: ActContext
    doc: DocumentMeta
    trace: Optional[Dict[str, Any]] = None