ValueRaider / yfinance-cache

Caching wrapper for yfinance module. Intelligent caching, not dumb caching of web requests.
MIT License
22 stars 9 forks source link

Fix Windows download() & exception propagate #46

Closed ValueRaider closed 1 week ago

ValueRaider commented 8 months ago

Fix multiprocess download() on Windows. Difference to Linux is children don't inherit from parent, so more initialisation needed. Then noticed child exceptions were blocking parent, now they are propagated.

Should resolve #43 - @ana-jiangR test please.

ValueRaider commented 8 months ago

From @ana-jiangR

got a new issue after applying that patch:

Traceback (most recent call last):
  File "C:\finstock\yf.download.test.py", line 134, in <module>
    test_yfc_download('MSFT')
  File "C:\finstock\yf.download.test.py", line 71, in test_yfc_download
    df = yfc.download(tickers,
  File "C:\miniconda3\envs\yfinance\lib\site-packages\yfinance_cache\yfc_multi.py", line 59, in download
    with multiprocessing.Pool(processes=threads, initializer=reinitialize_locks, initargs=(yfcd.exchanges_lock, yfcd.manager)) as pool:
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\context.py", line 119, in Pool
    return Pool(processes, initializer, initargs, maxtasksperchild,
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\pool.py", line 212, in __init__
    self._repopulate_pool()
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\pool.py", line 303, in _repopulate_pool
    return self._repopulate_pool_static(self._ctx, self.Process,
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\pool.py", line 326, in _repopulate_pool_static
    w.start()
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\process.py", line 121, in start
    self._popen = self._Popen(self)
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\context.py", line 327, in _Popen
    return Popen(process_obj)
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\popen_spawn_win32.py", line 93, in __init__
    reduction.dump(process_obj, to_child)
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\reduction.py", line 60, in dump
    ForkingPickler(file, protocol).dump(obj)
TypeError: cannot pickle 'weakref' object
Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\spawn.py", line 116, in spawn_main
    exitcode = _main(fd, parent_sentinel)
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\spawn.py", line 126, in _main
    self = reduction.pickle.load(from_parent)
EOFError: Ran out of input
ValueRaider commented 8 months ago

This could be tricky to fix so can you investigate? I did a little research, indicates that lock and/or manager in reinitialize_locks cannot be pickled. These are used in yfc_time.py, maybe you can figure out a different way to pass global lock and manager to children, or a different solution that doesn't use locks to stop race conditions.

ValueRaider commented 8 months ago

I've pushed a commit that should fix error.

ValueRaider commented 8 months ago

Forgot to git push, now it's available.

ana-jiangR commented 8 months ago

still doesn't work. I'm using the test code in your readme file: https://github.com/ValueRaider/yfinance-cache/blob/main/README.md#interface

Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\spawn.py", line 116, in spawn_main
    exitcode = _main(fd, parent_sentinel)
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\spawn.py", line 125, in _main
    prepare(preparation_data)
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\spawn.py", line 236, in prepare
    _fixup_main_from_path(data['init_main_from_path'])
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\spawn.py", line 287, in _fixup_main_from_path
    main_content = runpy.run_path(main_path,
  File "C:\miniconda3\envs\yfinance\lib\runpy.py", line 288, in run_path
    return _run_module_code(code, init_globals, run_name,
  File "C:\miniconda3\envs\yfinance\lib\runpy.py", line 97, in _run_module_code
    _run_code(code, mod_globals, init_globals,
  File "C:\miniconda3\envs\yfinance\lib\runpy.py", line 87, in _run_code
    exec(code, run_globals)
  File "C:\finstock\yf.download.test.py", line 122, in <module>
    import yfinance_cache as yfc
  File "C:\miniconda3\envs\yfinance\lib\site-packages\yfinance_cache\__init__.py", line 3, in <module>
    from .yfc_dat import Period, Interval
  File "C:\miniconda3\envs\yfinance\lib\site-packages\yfinance_cache\yfc_dat.py", line 362, in <module>
    manager = Manager()
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\context.py", line 57, in Manager
    m.start()
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\managers.py", line 554, in start
    self._process.start()
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\process.py", line 121, in start
    self._popen = self._Popen(self)
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\context.py", line 327, in _Popen
    return Popen(process_obj)
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\popen_spawn_win32.py", line 45, in __init__
    prep_data = spawn.get_preparation_data(process_obj._name)
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\spawn.py", line 154, in get_preparation_data
    _check_not_importing_main()
  File "C:\miniconda3\envs\yfinance\lib\multiprocessing\spawn.py", line 134, in _check_not_importing_main
    raise RuntimeError('''
RuntimeError:
        An attempt has been made to start a new process before the
        current process has finished its bootstrapping phase.

        This probably means that you are not using fork to start your
        child processes and you have forgotten to use the proper idiom
        in the main module:

            if __name__ == '__main__':
                freeze_support()
                ...

        The "freeze_support()" line can be omitted if the program
        is not going to be frozen to produce an executable.
ValueRaider commented 8 months ago

That's a new error though, we're making progress. I've pushed another patch.

ValueRaider commented 1 month ago

@sssstable I saw your multi commit, check this PR to avoid duplication.

sssstable commented 1 month ago

Hey ValueRaider, Glad that I can speak directly to you and have the chance to say "Awesome Work!". I am not using windows but Mac and it is rather the well known Streamlit Multiprocessing issue which is giving me trouble. The only thing I find as a working solution on the internet is this "multiprocess amendment", name == 'main' as in https://discuss.streamlit.io/t/multiprocessing-leads-to-frequent-app-crashes/56965/4

Do you know if & how it could be implemented in your package? I am not a pro (:

ValueRaider commented 1 month ago

I am not a pro (:

Me neither, I use AI a lot.

@estory1 I also see you working on multiprocessing, maybe start from here.

ValueRaider commented 1 month ago

@sssstable Putting if __name__ == '__main__' in yfc_multi.py doesn't make sense, that file just defines functions. Does it work if you move into your main script?