Implement a Leaky Bucket Limiter

Problem Statement

Implement a LeakyBucket class that regulates traffic based on a fixed capacity and a constant leak rate. The bucket holds a maximum of capacity tokens. Tokens are added to the bucket periodically (the 'leak' process), effectively limiting the number of requests allowed within a time window.

API Design

Implement the following methods:

• __init__(self, capacity: int, leak_rate_per_second: float): Initializes the bucket. capacity is the maximum number of requests allowed in the burst. leak_rate_per_second is how many requests can be processed per second.
• allow_request(self) -> bool: Returns True if a request can be processed (if a token is available), otherwise returns False.

Requirements

• The implementation must be thread-safe.
• Use time.time() to track the passage of time.
• Ensure the bucket state is updated lazily (i.e., calculate the number of leaked tokens only when a request arrives).

Examples

# Capacity 3, Leak rate 1.0 request per second
bucket = LeakyBucket(3, 1.0)

bucket.allow_request() # Returns True (Current bucket: 2)
bucket.allow_request() # Returns True (Current bucket: 1)
bucket.allow_request() # Returns True (Current bucket: 0)
bucket.allow_request() # Returns False

# Wait 2 seconds
# Bucket now has 2 tokens (1 per second)
bucket.allow_request() # Returns True

Constraints

• capacity > 0
• leak_rate_per_second > 0
• Handle high-concurrency environments using locks.

Edge Cases

• Requesting at the exact millisecond a token is generated.
• Multiple threads calling allow_request() simultaneously.
• Very high throughput exceeding the capacity significantly.