"""OCR backend interface + result type."""
from __future__ import annotations

import time
from dataclasses import dataclass, field
from pathlib import Path


@dataclass
class OCRBackendResult:
    text: str = ""
    engine: str = "none"
    tier: str = "offline"          # vlm | api | local | offline
    pages: int = 0
    confidence: float = 0.0
    latency_ms: float = 0.0
    simulated: bool = False
    cost_usd: float = 0.0
    error: str | None = None
    blocks: list = field(default_factory=list)

    @property
    def available(self) -> bool:
        return bool(self.text and self.text.strip()) and not self.error

    def to_channel_dict(self) -> dict:
        d = {
            "available": self.available,
            "chars": len(self.text.strip()),
            "engine": self.engine,
            "tier": self.tier,
            "confidence": round(self.confidence, 3),
            "latency_ms": round(self.latency_ms, 1),
        }
        if self.simulated:
            d["simulated"] = True
        if self.error:
            d["error"] = self.error
        if self.cost_usd:
            d["cost_usd"] = round(self.cost_usd, 6)
        return d


class OCRBackend:
    name: str = "base"
    tier: str = "offline"
    requires: str = ""  # human-readable description of what must be present

    def available(self) -> bool:  # pragma: no cover - overridden
        return False

    def extract(self, file_path: str | Path) -> OCRBackendResult:  # pragma: no cover
        raise NotImplementedError

    @staticmethod
    def _now() -> float:
        return time.perf_counter()

    def info(self) -> dict:
        return {"name": self.name, "tier": self.tier,
                "available": self.available(), "requires": self.requires}