You allow override of the cache functions, but only on INIT, your internal functions still use your own code.

[gbizeau]

I did a test and passed an override function for reading the cache file, While that worked on Init of the class, It appears that your builtin functions still utilize your existing functions. I need to be able to override your READ/WRITE in order to store this data in a DB for creating a multi user web based application.

[tdorssers]

Okay, I reviewed my code, but I can't find what you are pointing out. Can you provide some example code, so I can reproduce this?

[gbizeau]

Sure,

So in the init of the class, you allow an override of the cache_loader and cache_dumper functions.

def my_loader(): print("Loader") return dict()

def my_dumper(): print("Dumper") return dict()

with teslapy.Tesla('{email Address}', cache_loader=my_loader(), cache_dumper=my_dumper()) as tesla:

This works as expected, and these functions get called, however, you internal code doesn't honor this overloading of the class. Your _token_updater() function calls cache = self.cache_loader() on line 313 and self.cache_dumper(cache) on line 319. These should be referring back to the overloading functions I referred to on the init.

I put an inspect with a print on your _cache_load()

def _cache_load(self):
    """ Default cache loader method """
    import inspect
    print(inspect.stack()[1])
    try:
        with open(self.cache_file) as infile:
            cache = json.load(infile)
    except (IOError, ValueError):
        cache = {}
    return cache

    When I run the program, you can clearly see that it calls my functions, but then your _token_updater() calls your original functions instead.

Loader <- from my loader

Dumper <- from my dumper

Below were printed from your loader functions with the inspect I added.

FrameInfo(frame=<frame at 0x105c3c5b0, file '/PycharmProjects/venv/lib/python3.8/site-packages/teslapy/init.py', line 313, code _token_updater>, filename='/PycharmProjects/venv/lib/python3.8/site-packages/teslapy/init.py', lineno=313, function='_token_updater', code_context=[' cache = self.cache_loader()\n'], index=0)

Which is clearly coming from your "function='_token_updater'"

I know your setting the values to self on init, but they are not sticking.

    self.cache_loader = cache_loader or self._cache_load
    self.cache_dumper = cache_dumper or self._cache_dump    

If I print this at the end of your init

    logger.debug('Using SSO service URL %s', self.sso_base_url)

    print(cache_loader, cache_dumper)
    print(self.cache_loader, self.cache_dumper)

    I get 

{}{} <bound method Tesla._cache_load of <teslapy.Tesla object at 0x1003509d0>> <bound method Tesla._cache_dump of <teslapy.Tesla object at 0x1003509d0>> [{'id': 1492932703981922, 'vehicle_id': 535726691, 'vin': '5YJYGDEE4LF036810', 'display_name': 'Sulley', 'option_codes': 'AD15,AF00,APFB,APH4,AU3P,BC3B,BT37,CDM0,CH16,COCA,DRLH,DV4W,FC01,FG31,FM3B,GLFR,HL31,HM30,ID3W,IL31,LTPB,MDLY,MR30,PPSB,PC30,RENA,RF3G,RS3H,S3PB,SA3P,SC04,STCP,SU3C,TY9A,TM00,TW01,UT3P,WY9S,WR00,ZINV,MI00,PL30,SLR0,ST30,BG31,I36M,USSB,AUF1,RSF1,ILF1,FGF1,CPF0,TSHP,TR00,SR05,RL32,BY00,FCH1', 'color': None, 'access_type': 'OWNER', 'tokens': ['dd1cccf218ec15ea', 'e937c92a888e80ce'], 'state': 'asleep', 'in_service': False, 'id_s': '1492932703981922', 'calendar_enabled': True, 'api_version': 42, 'backseat_token': None, 'backseat_token_updated_at': None}]

Which is weird, if I am giving you an empty dict on the cache_load under _token_loder, it should be raising an error... It's not

Or maybe I am misunderstanding this whole thing?

I hope I explained it better this time :)

[tdorssers]

Crystal! There is an error in your code: with teslapy.Tesla('{email Address}', cache_loader=my_loader(), cache_dumper=my_dumper()) as tesla: Please remove () from my_loader and my_dumper and then it should work.

[gbizeau]

Ah Yes, when I added it with PyCharm it auto completed with the () :)

Thanks