A Server Module to Build RESTful APIs for Python Fnctions.
This framework forks a new process to run the specified function on each api request.
You can specify a maximum concurrency of the function. The api request is denied if the nubmer of running processes for the function is already reaches the concurrency limitation.
Example Code:
import concurrent.futures
import math
from restful_functions import ArgDefinition, ArgType, FunctionServer
def addition(x: int, y: int):
"""Simple Funciton."""
return x+y
def is_prime(n):
if n % 2 == 0:
return False
sqrt_n = int(math.floor(math.sqrt(n)))
for i in range(3, sqrt_n + 1, 2):
if n % i == 0:
return False
return True
def multi():
"""Heavy Function with MultiProcessing."""
PRIMES = [
112272535095293,
112582705942171,
112272535095293,
115280095190773,
115797848077099,
1099726899285419]
with concurrent.futures.ProcessPoolExecutor() as executor:
for number, prime in zip(PRIMES, executor.map(is_prime, PRIMES)):
print('%d is prime: %s' % (number, prime))
return True
if __name__ == '__main__':
server = FunctionServer()
server.add_function(
addition, # Function
[
ArgDefinition('x', ArgType.INTEGER, True, 'A value of X'),
ArgDefinition('y', ArgType.INTEGER, True, 'A value of Y'),
], # Args
1, # Max Concurrency
'Simple Awesome Addition' # Description
)
server.add_function(
multi,
[],
2,
'Heavy Function'
)
server.start()
Example Usage:
$ curl localhost:8888/
/function/list/data
/function/list/text
/function/definition/{function_name}
/function/running-count/{function_name}
/task/info/{task_id}
/task/done/{task_id}
/task/result/{task_id}
/task/list/{function_name}
/terminate/function/{function_name}
/terminate/task/{task_id}
$ curl localhost:8888/function/list/text
addition
URL:
async api: /addition
block api: /addition/keep-connection
Max Concurrency: 1
Description:
Simple Awesome Addition
Args
x INTEGER Requiered
A value of X
y INTEGER Requiered
A value of Y
Timeout: 86400 sec
multi
URL:
async api: /multi
block api: /multi/keep-connection
Max Concurrency: 2
Description:
Heavy Function
No Args
Timeout: 86400 sec
# Call Asynchronous
# Obtain task_id
$ curl -X POST -H "Content-Type: applicaiton/json" -d '{"x":3, "y":6}' http://localhost:8888/addition
{"success": true, "message": "", "task_id": "c3a6a0ef-b19e-4e6f-bce3-8d0e5a9046aa"}
# Obtain the result by task_id
$ curl http://localhost:8888/task/info/c3a6a0ef-b19e-4e6f-bce3-8d0e5a9046aa
{"task_id": "c3a6a0ef-b19e-4e6f-bce3-8d0e5a9046aa", "function_name": "addition", "status": "DONE", "result": 9}
$ curl http://localhost:8888/task/result/c3a6a0ef-b19e-4e6f-bce3-8d0e5a9046aa
9
# Call synchronous
# Keeping the connection until the process ends.
$ curl -X POST -H "Content-Type: applicaiton/json" -d '{"x":3, "y":6}' http://localhost:8888/addition/keep-connection
9
# Over Max Concurrency
$ curl -X POST http://localhost:8888/multi
{"success": true, "message": "", "task_id": "5bbbc1a0-74c2-4828-a843-fa2e2363e341"}
$ curl -X POST http://localhost:8888/multi
{"success": true, "message": "", "task_id": "7729af1f-c766-456e-a516-a75ab5f3a24c"}
$ curl -X POST http://localhost:8888/multi
{"success": false, "message": "Over Max Concurrency 2", "task_id": ""}
You can set timeout to the functions.
FunctionServer checks timeout to all runninng processes every 60 seconds (default). The interval can be changed with an argument polling_timeout_process_interval.
Example Code:
import time
from restful_functions import FunctionServer
def long_process():
time.sleep(100)
if __name__ == '__main__':
# Make polling_timeout_process_interval a small value to confirm a terminated process due to timeout
server = FunctionServer(polling_timeout_process_interval=1.0)
server.add_function(
long_process,
[],
1,
'long_process'
)
server.start()
Example Usage:
$ curl localhost:8888/
/function/list/data
/function/list/text
/function/definition/{function_name}
/function/running-count/{function_name}
/task/info/{task_id}
/task/done/{task_id}
/task/result/{task_id}
/task/list/{function_name}
/terminate/function/{function_name}
/terminate/task/{task_id}
$ curl localhost:8888/function/list/text
long_process
URL:
async api: /long_process
block api: /long_process/keep-connection
Max Concurrency: 1
Description:
timeout test
No Args
Timeout: 10 sec
# Call Asynchronous
# Obtain task_id
$ curl -X POST http://localhost:8888/long_process
{"success": true, "message": "", "task_id": "d6eb93f9-a8e7-4bba-817c-0fd975b3e41a"}
# Obtain the result by task_id
## Before 10 seconds at the function called.
$ curl http://localhost:8888/task/info/d6eb93f9-a8e7-4bba-817c-0fd975b3e41a
{"task_id": "d6eb93f9-a8e7-4bba-817c-0fd975b3e41a", "function_name": "long_process", "status": "RUNNING", "result": "timeout"}
## After 10 seconds at the function called.
$ curl http://localhost:8888/task/info/d6eb93f9-a8e7-4bba-817c-0fd975b3e41a
{"task_id": "d6eb93f9-a8e7-4bba-817c-0fd975b3e41a", "function_name": "long_process", "status": "TIMEOUT", "result": "timeout"}
# Call synchronous
# Keeping the connection until the process ends.
$ curl -X POST http://localhost:8888/long_process/keep-connection
"timeout"
# curl request ends after 10 seconds at the function called
- Using
- Docker
- poetry
- Recommended
- VSCode Remote Container Extension
How to run python scripts:
// 1. use `poetry shell` root@0aa0391397dd:/app# poetry shell Spawning shell within /root/.cache/pypoetry/virtualenvs/restful-functions-9TtSrW0h-py3.6 root@0aa0391397dd:/app# . /root/.cache/pypoetry/virtualenvs/restful-functions-9TtSrW0h-py3.6/bin/activate (restful-functions-9TtSrW0h-py3.6) root@0aa0391397dd:/app# python examples/example.py // 2. type `poetry run` every time root@0aa0391397dd:/app# poetry run python examples/example.py
MIT