Python library to replay the results of (random) function calls.
replay is a simple package that enables you to save the results of time intensive deterministic function calls or random function calls (and by function I mean any kind of callables), that should be replayed, within a file using a simple API.
replay will never execute any code from a replay file. It hashes the calls and looks for the hashes in the replay file, so there's no security risk in changing replay files (other then, that you could get other results...).
Replay files can have different formats. Right now only a human-readable and -editable format is implemented. This format looks mostly like python except that the actual values are encoded with json.
Here's an example script:
#!/usr/bin/env python
# -*- coding: utf-8 -*-
import random
import replay
r = replay.Replay('random.replay')
for i in range(5):
print(r(random.random))
r.save()The output is:
0.471306418755
0.954368067794
0.975113533495
0.155125371588
0.50165523797
Running this for the first time will create the file 'random.replay', which will look like this:
random() = 0.4713064187546834
random() = 0.9543680677943641
random() = 0.9751135334950991
random() = 0.15512537158813866
random() = 0.501655237970431
Running the script for the second time will result in the same output. Here's another example for a replay file:
random() = 0.3
pow(2, 3) = 8
# This is a comment.
random() = 0.2
sorted([7, 2, 3], reverse=true) = [7, 3, 2]
You can freely edit such a file and future calls to these functions will return the results you write into them. Example:
>>> import os
>>> import random
>>> import replay
>>> import tempfile
>>> fd, tmpname = tempfile.mkstemp(suffix='.replay')
>>> with open(tmpname, 'w') as f: n = f.write('random() = 40\n')
>>> r = Replay(tmpname)
>>> r(random.random)
40
>>> os.remove(tmpname)
replay is registred in the Python Package Index (PyPi). You can simply install the package using pip:
pip install replay