ValueError: invalid literal for int() with base 10: 'caches.json'

[yonikremer]

I get the following error when caching to disk:

test_models_support.py:9: in <module>
    from src.grouped_sampling.config import get_config
../src/grouped_sampling/__init__.py:1: in <module>
    from .base_pipeline import GroupedGenerationPipeLine
../src/grouped_sampling/base_pipeline.py:17: in <module>
    from .tokenizer import get_end_of_text_id, get_padding_id, get_tokenizer
../src/grouped_sampling/tokenizer.py:139: in <module>
    def get_tokenizer_name(
../../../miniconda3/envs/grouped_sampling/lib/python3.10/site-packages/cache_to_disk/cache_to_disk.py:72: in decorator
    delete_old_disk_caches()
../../../miniconda3/envs/grouped_sampling/lib/python3.10/site-packages/cache_to_disk/cache_to_disk.py:50: in delete_old_disk_caches
    max_age_days = int(file.split('_')[-1].replace('.pkl', ''))
E   ValueError: invalid literal for int() with base 10: 'caches.json'
collected 0 items / 1 error
ERROR: not found: /home/yoni/PycharmProjects/grouped_sampling/tests/test_models_support.py::test_unsupported_tokenizers
(no name '/home/yoni/PycharmProjects/grouped_sampling/tests/test_models_support.py::test_unsupported_tokenizers' in any of [<Module test_models_support.py>])

[yonikremer]

I solved it:

I changed get_files_in_directory to:

def get_files_in_directory(directory):
    return [
        f for f in os.listdir(directory) if
        isfile(join(directory, f)) and f.endswith('.pkl')
    ]