"""ГРЛС stand-in: short excerpts of the official drug instruction, per section (SPEC §9, SPEC §7 п.5).

Each `data/meds/grls/<name>.yaml` holds one drug:

    inn: сертралин
    aliases: [sertraline]
    sections:
      side_effects: {text: "...", source: "ГРЛС, инструкция по медицинскому применению, раздел «...»"}

Section codes are the ones `app/llm/med_qa.py` understands (side_effects, onset, alcohol, discontinuation,
administration, interactions, overdose). The bot answers drug questions ONLY from these texts, so they are
the single place to verify against the current ГРЛС instruction (see the header of each file).
"""

import logging
from pathlib import Path

import yaml
from pydantic import BaseModel, Field, ValidationError

from app.config import get_settings
from app.integrations.base import DrugSection

log = logging.getLogger(__name__)


class _SectionFile(BaseModel):
    text: str
    source: str


class _DrugFile(BaseModel):
    inn: str
    aliases: list[str] = Field(default_factory=list)
    sections: dict[str, _SectionFile]


def default_grls_dir() -> Path:
    return get_settings().data_dir / "meds" / "grls"


def normalize_inn(name: str) -> str:
    """Lookup key: «Сертралин 50 мг», «сертралин», «Sertraline» -> first word, lower case, ё -> е."""
    words = name.strip().lower().replace("ё", "е").split()
    return words[0] if words else ""


class GrlsExcerpts:
    """`DrugInfo` over local YAML excerpts. Files are read lazily on the first lookup and cached."""

    def __init__(self, directory: Path | None = None) -> None:
        self._dir = directory
        self._index: dict[str, list[DrugSection]] | None = None

    def _load(self) -> dict[str, list[DrugSection]]:
        if self._index is not None:
            return self._index
        directory = self._dir or default_grls_dir()
        index: dict[str, list[DrugSection]] = {}
        for path in sorted(directory.glob("*.yaml")):
            try:
                drug = _DrugFile.model_validate(yaml.safe_load(path.read_text(encoding="utf-8")))
            except (OSError, yaml.YAMLError, ValidationError) as exc:
                log.warning("grls file skipped file=%s error=%s", path.name, type(exc).__name__)
                continue
            sections = [
                DrugSection(inn=drug.inn, section=code, text=" ".join(s.text.split()), source=s.source)
                for code, s in drug.sections.items()
            ]
            for key in (drug.inn, *drug.aliases):
                index[normalize_inn(key)] = sections
        self._index = index
        return index

    def known(self) -> list[str]:
        """INNs with excerpts (for the dashboard / tests)."""
        return sorted({s[0].inn for s in self._load().values() if s})

    async def sections(self, inn: str) -> list[DrugSection]:
        return [s.model_copy() for s in self._load().get(normalize_inn(inn), [])]