Build and cache Stan models smartly (ignoring changes in comments and white spaces)

cachestan.py

import re
import pystan
import hashlib
import pickle
import os

def build_model(code):
    lines = code.split("\n")
    lines = [re.sub('//.*$', '', line).strip() for line in lines]
    lines = [line.replace('    ', ' ').replace('  ', ' ').replace('  ', ' ')
        for line in lines if len(line) > 0]

    slimcode = '\n'.join(lines).strip()
    slimbytes = slimcode.encode(encoding="ascii", errors="ignore")
    slimcode = slimbytes.decode(encoding="ascii")
    hexcode = hashlib.sha224(slimbytes).hexdigest()[:6]
    #print("Stan code[%s]:\n%s" % (hexcode, '\n'.join(['%3d: %s' % (i+1, line) for i, line in enumerate(code.split("\n"))])))
    print("Slimmed Stan code[%s]:\n%s" % (hexcode, '\n'.join(['%3d: %s' % (i+1, line) for i, line in enumerate(slimcode.split("\n"))])))
    
    if not os.path.exists('.modelcache'):
        os.mkdir('.modelcache')
    filename = '.modelcache/%s.pkl' % hexcode
    
    try:
        return pickle.load(open(filename, 'rb'))
    except IOError:
        pass
    
    sm = pystan.StanModel(model_code=slimcode, model_name='m' + hexcode)
    with open(filename, 'wb') as f:
        pickle.dump(sm, f)
    return sm