Last active
October 22, 2019 07:21
-
-
Save sio/5c83f2d409ed448207e3684f32f163b9 to your computer and use it in GitHub Desktop.
Rate limit access to sensitive resources | Development moved to https://github.com/sio/scrapehelper
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
''' | |
Rate limit access to sensitive resources | |
''' | |
# Copyright 2019 Vitaly Potyarkin | |
# | |
# Licensed under the Apache License, Version 2.0 (the "License"); | |
# you may not use this file except in compliance with the License. | |
# You may obtain a copy of the License at | |
# | |
# http://www.apache.org/licenses/LICENSE-2.0 | |
# | |
# Unless required by applicable law or agreed to in writing, software | |
# distributed under the License is distributed on an "AS IS" BASIS, | |
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | |
# See the License for the specific language governing permissions and | |
# limitations under the License. | |
import threading | |
import time | |
from functools import wraps | |
class RateLimitReachedError(RuntimeError): | |
'''Raised when rate limit is reached''' | |
class RateLimiter: | |
''' | |
Honor rate limits when using third party API | |
RateLimiter objects can be used in any of three ways: | |
- Through decorator interface | |
- Through context manager interface | |
- Manually registering new_call() before sending an API request | |
''' | |
# TODO: write tests for RateLimiter | |
REFRESH_INTERVAL = 0.5 # seconds (minimum sleep interval) | |
def __init__(self, calls=15, interval=15*60, wait=True): | |
''' | |
Create a new RateLimiter object | |
calls: | |
Maximum number of API calls allowed within time period | |
interval: | |
The time period (in seconds) for which the number of calls is | |
limited | |
wait: | |
Boolean. If True the RateLimiter will wait until limit terms are | |
satisfied before registering a new call. Will raise | |
RateLimitReachedError otherwise. | |
''' | |
self.call_limit = calls | |
self.call_log = [] | |
self.interval = interval | |
self.wait = wait | |
self.clock = time.monotonic | |
self.lock = threading.RLock() | |
self.next_cleanup = 0 | |
def new_call(self): | |
''' | |
Register new API call | |
If self.wait is True, wait until rate limit terms are satisfied. | |
Raise RateLimitReachedError otherwise. | |
''' | |
while True: | |
try: | |
self._call_attempt() | |
break | |
except RateLimitReachedError as e: | |
if self.wait: | |
time.sleep(max( | |
self.REFRESH_INTERVAL, | |
self.next_cleanup - self.clock() | |
)) | |
else: | |
raise e | |
def _call_attempt(self): | |
''' | |
Try to add a new item to the call log. | |
Raises RateLimitReachedError if the log is full. | |
''' | |
with self.lock: | |
self._cleanup() | |
if len(self.call_log) >= self.call_limit: | |
raise RateLimitReachedError( | |
'can not make more than {num} calls in {interval} seconds'.format( | |
num = self.call_limit, | |
interval = self.interval, | |
) | |
) | |
self.call_log.append(self.clock()) | |
def _cleanup(self): | |
''' | |
Maintain the call log: remove obsolete entries, schedule next cleanup | |
''' | |
if self.call_log and not self.next_cleanup: | |
self.next_cleanup = self.call_log[0] + self.interval | |
while self.next_cleanup\ | |
and self.clock() > self.next_cleanup: | |
try: | |
self.call_log.pop(0) | |
self.next_cleanup = self.call_log[0] + self.interval | |
except IndexError: # pop from empty list | |
self.next_cleanup = 0 | |
def __enter__(self): | |
'''Context manager interface for RateLimiter''' | |
self.new_call() | |
def __exit__(self, exc_type, exc_val, exc_tb): | |
'''Context manager interface for RateLimiter''' | |
return False | |
def __call__(self, function): | |
'''Function decorator interface for RateLimiter''' | |
@wraps(function) | |
def decorated(*a, **kw): | |
self.new_call() | |
return function(*a, **kw) | |
return decorated |
You're welcome! :-)
After I posted this gist I've included its code into my scrapehelper package, installable with pip (though not on PyPI): https://github.com/sio/scrapehelper/blob/master/scrapehelper/limit.py
Currently the code is mostly identical (only one property was added: remaining
), but all further changes will happen there.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Extremely grateful that you went to the effort to implement this alternative, as I was preparing to do something similar after noticing the same issues with the source project. Thank you,