| Name | nshconfig JSON |
| Version |
0.17.2
JSON |
| download |
| home_page | None |
| Summary | Fully typed configuration management, powered by Pydantic |
| upload_time | 2024-10-01 00:49:44 |
| maintainer | None |
| docs_url | None |
| author | Nima Shoghi |
| requires_python | <4.0,>=3.10 |
| license | None |
| keywords |
|
| VCS |
 |
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# nshconfig
Fully typed configuration management, powered by [Pydantic](https://github.com/pydantic/pydantic/)
## Motivation
As a machine learning researcher, I often found myself running numerous training jobs with various hyperparameters for the models I was working on. Keeping track of these parameters in a fully typed manner became increasingly important. While the excellent `pydantic` library provided most of the functionality I needed, I wanted to add a few extra features to streamline my workflow. This led to the creation of `nshconfig`.
## Installation
You can install `nshconfig` via pip:
```bash
pip install nshconfig
```
## Usage
While the primary use case for `nshconfig` is in machine learning projects, it can be used in any Python project where you need to store configurations in a fully typed manner.
Here's a basic example of how to use `nshconfig`:
```python
import nshconfig as C
class MyConfig(C.Config):
field1: int
field2: str
field3: C.AllowMissing[float] = C.MISSING
config = MyConfig.draft()
config.field1 = 42
config.field2 = "hello"
final_config = config.finalize()
print(final_config)
```
For more advanced usage and examples, please refer to the documentation.
## Features
### Draft Configs
Draft configs allow for a nicer API when creating configurations. Instead of relying on JSON or YAML files, you can create your configs using pure Python:
```python
config = MyConfig.draft()
# Set some values
config.a = 10
config.b = "hello"
# Finalize the config
config = config.finalize()
```
This approach enables a more intuitive and expressive way of defining your configurations.
#### Motivation
The primary motivation behind draft configs is to provide a cleaner and more Pythonic way of creating configurations. By leveraging the power of Python, you can define your configs in a more readable and maintainable manner.
#### Usage Guide
1. Create a draft config using the `draft()` class method:
```python
config = MyConfig.draft()
```
2. Set the desired values on the draft config:
```python
config.field1 = value1
config.field2 = value2
```
3. Finalize the draft config to obtain the validated configuration:
```python
final_config = config.finalize()
```
### MISSING Constant
The `MISSING` constant is similar to `None`, but with a key difference. While `None` has the type `NoneType` and can only be assigned to fields of type `T | None`, the `MISSING` constant has the type `Any` and can be assigned to fields of any type.
#### Motivation
The `MISSING` constant addresses a common issue when working with optional fields in configurations. Consider the following example:
```python
import nshconfig as C
# Without MISSING:
class MyConfigWithoutMissing(C.Config):
age: int
age_str: str | None = None
def __post_init__(self):
if self.age_str is None:
self.age_str = str(self.age)
config = MyConfigWithoutMissing(age=10)
age_str_lower = config.age_str.lower()
# ^ The above line is valid code, but the type-checker will complain because `age_str` could be `None`.
```
In the above code, the type-checker will raise a complaint because `age_str` could be `None`. This is where the `MISSING` constant comes in handy:
```python
# With MISSING:
class MyConfigWithMissing(C.Config):
age: int
age_str: C.AllowMissing[str] = C.MISSING
def __post_init__(self):
if self.age_str is C.MISSING:
self.age_str = str(self.age)
config = MyConfigWithMissing(age=10)
age_str_lower = config.age_str.lower()
# ^ No more type-checker complaints!
```
By using the `MISSING` constant, you can indicate that a field is not set during construction, and the type-checker will not raise any complaints.
### Seamless Integration with PyTorch Lightning
`nshconfig` seamlessly integrates with PyTorch Lightning by implementing the `Mapping` interface. This allows you to use your configs directly as the `hparams` argument in your Lightning modules without any additional effort.
## Credit
`nshconfig` is built on top of the incredible [`pydantic`](https://github.com/pydantic/pydantic/) library. Massive credit goes to the [`pydantic`](https://github.com/pydantic/pydantic/) team for creating such a powerful and flexible tool for data validation and settings management.
## Contributing
Contributions are welcome! If you find any issues or have suggestions for improvement, please open an issue or submit a pull request on the GitHub repository.
## License
`nshconfig` is open-source software licensed under the [MIT License](LICENSE).
Raw data
{
"_id": null,
"home_page": null,
"name": "nshconfig",
"maintainer": null,
"docs_url": null,
"requires_python": "<4.0,>=3.10",
"maintainer_email": null,
"keywords": null,
"author": "Nima Shoghi",
"author_email": "nimashoghi@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/28/38/3738890b8c37a18cfdb796f82f45f48e22ad073f2b595cdbd7da2d725650/nshconfig-0.17.2.tar.gz",
"platform": null,
"description": "# nshconfig\n\nFully typed configuration management, powered by [Pydantic](https://github.com/pydantic/pydantic/)\n\n## Motivation\n\nAs a machine learning researcher, I often found myself running numerous training jobs with various hyperparameters for the models I was working on. Keeping track of these parameters in a fully typed manner became increasingly important. While the excellent `pydantic` library provided most of the functionality I needed, I wanted to add a few extra features to streamline my workflow. This led to the creation of `nshconfig`.\n\n\n## Installation\n\nYou can install `nshconfig` via pip:\n\n```bash\npip install nshconfig\n```\n\n## Usage\n\nWhile the primary use case for `nshconfig` is in machine learning projects, it can be used in any Python project where you need to store configurations in a fully typed manner.\n\nHere's a basic example of how to use `nshconfig`:\n\n```python\nimport nshconfig as C\n\nclass MyConfig(C.Config):\n field1: int\n field2: str\n field3: C.AllowMissing[float] = C.MISSING\n\nconfig = MyConfig.draft()\nconfig.field1 = 42\nconfig.field2 = \"hello\"\nfinal_config = config.finalize()\n\nprint(final_config)\n```\n\nFor more advanced usage and examples, please refer to the documentation.\n\n## Features\n\n### Draft Configs\n\nDraft configs allow for a nicer API when creating configurations. Instead of relying on JSON or YAML files, you can create your configs using pure Python:\n\n```python\nconfig = MyConfig.draft()\n\n# Set some values\nconfig.a = 10\nconfig.b = \"hello\"\n\n# Finalize the config\nconfig = config.finalize()\n```\n\nThis approach enables a more intuitive and expressive way of defining your configurations.\n\n#### Motivation\n\nThe primary motivation behind draft configs is to provide a cleaner and more Pythonic way of creating configurations. By leveraging the power of Python, you can define your configs in a more readable and maintainable manner.\n\n#### Usage Guide\n\n1. Create a draft config using the `draft()` class method:\n ```python\n config = MyConfig.draft()\n ```\n\n2. Set the desired values on the draft config:\n ```python\n config.field1 = value1\n config.field2 = value2\n ```\n\n3. Finalize the draft config to obtain the validated configuration:\n ```python\n final_config = config.finalize()\n ```\n\n### MISSING Constant\n\nThe `MISSING` constant is similar to `None`, but with a key difference. While `None` has the type `NoneType` and can only be assigned to fields of type `T | None`, the `MISSING` constant has the type `Any` and can be assigned to fields of any type.\n\n#### Motivation\n\nThe `MISSING` constant addresses a common issue when working with optional fields in configurations. Consider the following example:\n\n```python\nimport nshconfig as C\n\n# Without MISSING:\nclass MyConfigWithoutMissing(C.Config):\n age: int\n age_str: str | None = None\n\n def __post_init__(self):\n if self.age_str is None:\n self.age_str = str(self.age)\n\nconfig = MyConfigWithoutMissing(age=10)\nage_str_lower = config.age_str.lower()\n# ^ The above line is valid code, but the type-checker will complain because `age_str` could be `None`.\n```\n\nIn the above code, the type-checker will raise a complaint because `age_str` could be `None`. This is where the `MISSING` constant comes in handy:\n\n```python\n# With MISSING:\nclass MyConfigWithMissing(C.Config):\n age: int\n age_str: C.AllowMissing[str] = C.MISSING\n\n def __post_init__(self):\n if self.age_str is C.MISSING:\n self.age_str = str(self.age)\n\nconfig = MyConfigWithMissing(age=10)\nage_str_lower = config.age_str.lower()\n# ^ No more type-checker complaints!\n```\n\nBy using the `MISSING` constant, you can indicate that a field is not set during construction, and the type-checker will not raise any complaints.\n\n### Seamless Integration with PyTorch Lightning\n\n`nshconfig` seamlessly integrates with PyTorch Lightning by implementing the `Mapping` interface. This allows you to use your configs directly as the `hparams` argument in your Lightning modules without any additional effort.\n\n## Credit\n\n`nshconfig` is built on top of the incredible [`pydantic`](https://github.com/pydantic/pydantic/) library. Massive credit goes to the [`pydantic`](https://github.com/pydantic/pydantic/) team for creating such a powerful and flexible tool for data validation and settings management.\n\n## Contributing\n\nContributions are welcome! If you find any issues or have suggestions for improvement, please open an issue or submit a pull request on the GitHub repository.\n\n## License\n\n`nshconfig` is open-source software licensed under the [MIT License](LICENSE).\n",
"bugtrack_url": null,
"license": null,
"summary": "Fully typed configuration management, powered by Pydantic",
"version": "0.17.2",
"project_urls": {
"homepage": "https://github.com/nimashoghi/nshconfig"
},
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "a1f228fe8a47146231592c1a0c4749c8c006c48f353bc1f47753e5ce2a0e801f",
"md5": "1e23f389eec5541975a120c561fa3cbb",
"sha256": "24d8bb697a4e05c4c032f6e02e7a90d61521d8a93a534983367e486a7b308526"
},
"downloads": -1,
"filename": "nshconfig-0.17.2-py3-none-any.whl",
"has_sig": false,
"md5_digest": "1e23f389eec5541975a120c561fa3cbb",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": "<4.0,>=3.10",
"size": 15367,
"upload_time": "2024-10-01T00:49:43",
"upload_time_iso_8601": "2024-10-01T00:49:43.124901Z",
"url": "https://files.pythonhosted.org/packages/a1/f2/28fe8a47146231592c1a0c4749c8c006c48f353bc1f47753e5ce2a0e801f/nshconfig-0.17.2-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "28383738890b8c37a18cfdb796f82f45f48e22ad073f2b595cdbd7da2d725650",
"md5": "d1de0c0b88a1878f209588d001ae764f",
"sha256": "8072aa034603a2ab0a6d69590e66363e8e2a0c9dfae2d6f42948af86df746475"
},
"downloads": -1,
"filename": "nshconfig-0.17.2.tar.gz",
"has_sig": false,
"md5_digest": "d1de0c0b88a1878f209588d001ae764f",
"packagetype": "sdist",
"python_version": "source",
"requires_python": "<4.0,>=3.10",
"size": 14291,
"upload_time": "2024-10-01T00:49:44",
"upload_time_iso_8601": "2024-10-01T00:49:44.515972Z",
"url": "https://files.pythonhosted.org/packages/28/38/3738890b8c37a18cfdb796f82f45f48e22ad073f2b595cdbd7da2d725650/nshconfig-0.17.2.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-10-01 00:49:44",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "nimashoghi",
"github_project": "nshconfig",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"lcname": "nshconfig"
}
