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._clock = clock
        self._tokens = capacity
        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, tokens: float = 1.0) -> bool:
        """Return True and spend tokens when enough are available."""
        if tokens <= 0:
            raise ValueError("tokens must be positive")

        with self._lock:
            now = self._clock()
            self._refill(now)

            if self._tokens < tokens:
                return False

            self._tokens -= tokens
            return True

    def allow(self) -> bool:
        """Convenience method for one-token requests."""
        return self.consume(1.0)

    def time_until_available(self, tokens: float = 1.0) -> float:
        """Return seconds until the requested number of tokens is available."""
        if tokens <= 0:
            raise ValueError("tokens must be positive")
        if tokens > self.capacity:
            raise ValueError("tokens cannot exceed bucket capacity")

        with self._lock:
            now = self._clock()
            self._refill(now)

            if self._tokens >= tokens:
                return 0.0

            return (tokens - self._tokens) / self.rate


class FakeClock:
    def __init__(self) -> None:
        self.now = 0.0

    def __call__(self) -> float:
        return self.now

    def advance(self, seconds: float) -> None:
        self.now += seconds


class TokenBucketTests(unittest.TestCase):
    def test_burst_then_limit(self) -> None:
        clock = FakeClock()
        bucket = TokenBucket(rate=2.0, capacity=3.0, clock=clock)

        self.assertTrue(bucket.allow())
        self.assertTrue(bucket.allow())
        self.assertTrue(bucket.allow())
        self.assertFalse(bucket.allow())

    def test_refills_over_time(self) -> None:
        clock = FakeClock()
        bucket = TokenBucket(rate=2.0, capacity=3.0, clock=clock)

        self.assertTrue(bucket.consume(3.0))
        self.assertFalse(bucket.allow())

        clock.advance(0.5)
        self.assertTrue(bucket.allow())
        self.assertFalse(bucket.allow())

        clock.advance(1.0)
        self.assertTrue(bucket.consume(2.0))


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