captain-kark / python-module-resources Goto Github PK

Add a new directory, module_resources/examples/{json,yaml}/subdirectory to the examples, and ensure that imports are handled orderly.

Account for matching names between a file in the module's top-level directory, and a subdirectory within the module. Use from to distinguish these submodule imports versus named documents in the top-level.

The __iter__ method is currently designed to help dict() work easily, but it makes for-loops deal with dicts, making things clumsy when working with loops.

Find a way to continue using __iter__ as dicts for calls to dict(), while avoiding this type of interaction in loops.

from some_module_resource import module_json_resource

dict(module_json_resource)
# {"first": {"namedtuple": True}, "second": {"namedtuple": True}}

for resource_key, dict_value in module_json_resource:
    namedtuple_value = getattr(module_json_resource, resource_key)
    print(hasattr(namedtuple_value.second))

A resource object should know how to write itself back to the filesystem in the same location and format it was in when it was imported.

Yaml configuration files work as well, with the module_resources.YamlResourceModule class.

Should be module_resources.YamlModuleResource.

You can specify a placeholder Any-typed variable reference in the init file for each module object that is intended to be importable to fix pylint errors.

You should be able to import a module and have the contents of the imported object sourced from someplace like /var/config, /tmp/whatever/, etc.

Imports would still be requesting the same module that the associated __init__.py file is located in. It would just affect where the source json/yaml files are searched for.

This might already be the default behavior. If it isn't, it should support this.

Currently it throws an attribute error, but the message itself should give more details about the source of this fallback dict object, and why the property must be accessed via key.

This came up during manual testing. I fixed it, but a regression test needs to be written.

See references to clean_value in module_resources.py for where I handled this issue in the first place.

Running dict() had a side effect of altering the namedtuple as it iterated through it.