from __future__ import annotations

import threading
import time
import unittest
from collections.abc import Callable


class TokenBucket:
    """Thread-safe token-bucket rate limiter."""

    def __init__(
        self,
        rate: float,
        capacity: float,
        *,
        clock: Callable[[], float] = time.monotonic,
    ) -> None:
        if rate <= 0:
            raise ValueError("rate must be positive")
        if capacity <= 0:
            raise ValueError("capacity must be positive")

        self.rate = rate
        self.capacity = capacity
        self._tokens = capacity
        self._clock = clock
        self._updated_at = clock()
        self._lock = threading.Lock()

    def _refill(self, now: float) -> None:
        elapsed = max(0.0, now - self._updated_at)
        self._tokens = min(self.capacity, self._tokens + elapsed * self.rate)
        self._updated_at = now

    def consume(self, amount: float = 1.0) -> bool:
        """Consume tokens immediately, returning False when unavailable."""
        if amount <= 0:
            raise ValueError("amount must be positive")
        if amount > self.capacity:
            return False

        with self._lock:
            self._refill(self._clock())
            if self._tokens < amount:
                return False
            self._tokens -= amount
            return True

    def retry_after(self, amount: float = 1.0) -> float:
        """Return seconds until the requested tokens should be available."""
        if amount <= 0:
            raise ValueError("amount must be positive")
        if amount > self.capacity:
            return float("inf")

        with self._lock:
            self._refill(self._clock())
            return max(0.0, (amount - self._tokens) / self.rate)


class TokenBucketTests(unittest.TestCase):
    def test_capacity_is_enforced(self) -> None:
        now = [0.0]
        bucket = TokenBucket(2.0, 2.0, clock=lambda: now[0])

        self.assertTrue(bucket.consume(2.0))
        self.assertFalse(bucket.consume())

    def test_tokens_refill_over_time(self) -> None:
        now = [0.0]
        bucket = TokenBucket(2.0, 2.0, clock=lambda: now[0])

        bucket.consume(2.0)
        now[0] = 0.5

        self.assertEqual(bucket.retry_after(), 0.0)
        self.assertTrue(bucket.consume())
        self.assertAlmostEqual(bucket.retry_after(), 0.5)


if __name__ == "__main__":
    unittest.main()