knowledge-adapter-service/src/knowledge_adapter/models.py

from typing import Any
from uuid import UUID

from pydantic import BaseModel, ConfigDict, Field, field_validator, model_validator


class RetrievalSetting(BaseModel):
    model_config = ConfigDict(extra="forbid")

    top_k: int = Field(ge=1)
    score_threshold: float = Field(ge=0, le=1)


class MetadataCondition(BaseModel):
    model_config = ConfigDict(extra="forbid")

    logical_operator: str | None = None
    conditions: list[dict[str, Any]] = Field(default_factory=list)

    @model_validator(mode="after")
    def reject_nonempty_conditions(self) -> "MetadataCondition":
        if self.conditions:
            raise ValueError("metadata conditions are not supported")
        return self


class RetrievalRequest(BaseModel):
    model_config = ConfigDict(extra="forbid")

    knowledge_id: UUID
    query: str
    retrieval_setting: RetrievalSetting
    metadata_condition: MetadataCondition | None = None

    @field_validator("query")
    @classmethod
    def strip_nonempty_query(cls, value: str) -> str:
        normalized = value.strip()
        if not normalized:
            raise ValueError("query must not be empty")
        return normalized


class WeKnoraChunk(BaseModel):
    model_config = ConfigDict(extra="ignore", allow_inf_nan=False)

    id: UUID
    content: str
    score: float
    knowledge_base_id: UUID
    knowledge_id: UUID
    knowledge_title: str = ""
    knowledge_filename: str = ""
    chunk_index: int
    chunk_type: str
    image_info: str = ""