# RefConfig
Powered by [SmartDict](https://pypi.org/project/smartdict/).
## Usage
- Configure with references
## Install
`pip install refconfig`
## Description
```yaml
# ./data/zhihu.yaml
dataset: zhihu
store:
data_dir: data/FANS/${config.dataset}
save_dir: saving/${config.dataset}/${exp.model}-E${config.model_config.hidden_size}/
model_config:
hidden_size: 64
num_hidden_layers: 3
num_attention_heads: 8
```
```yaml
# ./exp/train.yaml
exp: step-${exp.tasks.0.params.steps}
model: bert
mode: train
freeze_emb: false
tasks:
-
name: mlm
params:
steps: 5
store:
interval: 10
policy:
epoch: 200
```
```python
import refconfig
config = refconfig.parse_yaml(
exp='./exp/train.yaml',
config='./data/zhihu.yaml',
)
print(config['config']['store']['data_dir']) # => data/FANS/zhihu
print(config['config']['store']['save_dir']) # => saving/zhihu/bert-E64/
print(config['exp']['exp']) # => step-5
# use with the Oba library to achieve a smoother effect
from oba import Obj
config = Obj(config)
config, exp = config.config, config.exp
print(config.store.data_dir) # => data/FANS/zhihu
print(config.store.save_dir) # => saving/zhihu/bert-E64/
print(exp.exp) # => step-5
```
```yaml
# ./data/any.yaml
store:
data_dir: data/FANS/${dataset}
save_dir: saving/${dataset}/${exp.model}-E${config.model_config.hidden_size}/
model_config:
hidden_size: 64
num_hidden_layers: 3
num_attention_heads: 8
```
```python
from refconfig import RefConfig
config = RefConfig().add_yaml(
exp='./exp/train.yaml',
config='./data/any.yaml',
).add_raw(
dataset='youtube',
).parse()
from oba import Obj
config = Obj(config)
config, exp = config.config, config.exp
print(config.store.data_dir) # => data/FANS/youtube
print(config.store.save_dir) # => saving/youtube/bert-E64/
print(exp.exp) # => step-5
```
Raw data
{
"_id": null,
"home_page": "https://github.com/Jyonn/RefConfig",
"name": "refconfig",
"maintainer": "",
"docs_url": null,
"requires_python": "",
"maintainer_email": "",
"keywords": "configuration,refer",
"author": "Jyonn Liu",
"author_email": "i@6-79.cn",
"download_url": "https://files.pythonhosted.org/packages/9a/31/c8a287ac7c12187a2a49386bdbdaaba3aa96f0eb40090808a195ef5b4c11/refconfig-0.0.3.tar.gz",
"platform": "any",
"description": "# RefConfig\n\nPowered by [SmartDict](https://pypi.org/project/smartdict/).\n\n## Usage\n\n- Configure with references\n\n## Install \n\n`pip install refconfig`\n\n## Description\n\n```yaml\n# ./data/zhihu.yaml\n\ndataset: zhihu\nstore:\n data_dir: data/FANS/${config.dataset}\n save_dir: saving/${config.dataset}/${exp.model}-E${config.model_config.hidden_size}/\nmodel_config:\n hidden_size: 64\n num_hidden_layers: 3\n num_attention_heads: 8\n```\n\n```yaml\n# ./exp/train.yaml\n\nexp: step-${exp.tasks.0.params.steps}\nmodel: bert\nmode: train\nfreeze_emb: false\ntasks:\n -\n name: mlm\n params:\n steps: 5\nstore:\n interval: 10\npolicy:\n epoch: 200\n```\n\n```python\nimport refconfig\n\nconfig = refconfig.parse_yaml(\n exp='./exp/train.yaml',\n config='./data/zhihu.yaml',\n)\n\nprint(config['config']['store']['data_dir']) # => data/FANS/zhihu\nprint(config['config']['store']['save_dir']) # => saving/zhihu/bert-E64/\nprint(config['exp']['exp']) # => step-5\n\n# use with the Oba library to achieve a smoother effect\n\nfrom oba import Obj\n\nconfig = Obj(config)\nconfig, exp = config.config, config.exp\nprint(config.store.data_dir) # => data/FANS/zhihu\nprint(config.store.save_dir) # => saving/zhihu/bert-E64/\nprint(exp.exp) # => step-5\n```\n\n```yaml\n# ./data/any.yaml\n\nstore:\n data_dir: data/FANS/${dataset}\n save_dir: saving/${dataset}/${exp.model}-E${config.model_config.hidden_size}/\nmodel_config:\n hidden_size: 64\n num_hidden_layers: 3\n num_attention_heads: 8\n```\n\n```python\nfrom refconfig import RefConfig\n\nconfig = RefConfig().add_yaml(\n exp='./exp/train.yaml',\n config='./data/any.yaml',\n).add_raw(\n dataset='youtube',\n).parse()\n\nfrom oba import Obj\n\nconfig = Obj(config)\nconfig, exp = config.config, config.exp\nprint(config.store.data_dir) # => data/FANS/youtube\nprint(config.store.save_dir) # => saving/youtube/bert-E64/\nprint(exp.exp) # => step-5\n```\n\n\n",
"bugtrack_url": null,
"license": "MIT Licence",
"summary": "a flexible and variable-based tool for multi-type configuration, including json, yaml, and python dict",
"version": "0.0.3",
"split_keywords": [
"configuration",
"refer"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "9a31c8a287ac7c12187a2a49386bdbdaaba3aa96f0eb40090808a195ef5b4c11",
"md5": "9616e5b050536dc5ebe301c275cc0cfa",
"sha256": "343eb56a6065220e7848b28aab524c40ada53a9a1b1fb2161cb3d4e084a7c76f"
},
"downloads": -1,
"filename": "refconfig-0.0.3.tar.gz",
"has_sig": false,
"md5_digest": "9616e5b050536dc5ebe301c275cc0cfa",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 3172,
"upload_time": "2023-01-18T10:23:27",
"upload_time_iso_8601": "2023-01-18T10:23:27.833865Z",
"url": "https://files.pythonhosted.org/packages/9a/31/c8a287ac7c12187a2a49386bdbdaaba3aa96f0eb40090808a195ef5b4c11/refconfig-0.0.3.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-01-18 10:23:27",
"github": true,
"gitlab": false,
"bitbucket": false,
"github_user": "Jyonn",
"github_project": "RefConfig",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"lcname": "refconfig"
}
JSON
