Metadata-Version: 2.1
Name: naapc
Version: 1.5.0
Summary: Nested Automated Argument Parsing Configuration (NAAPC).
Author-email: Bai Huanyu <eiphnix@gmail.com>
License: MIT License
        
        Copyright (c) 2022 Bai Huanyu
        
        Permission is hereby granted, free of charge, to any person obtaining a copy
        of this software and associated documentation files (the "Software"), to deal
        in the Software without restriction, including without limitation the rights
        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
        copies of the Software, and to permit persons to whom the Software is
        furnished to do so, subject to the following conditions:
        
        The above copyright notice and this permission notice shall be included in all
        copies or substantial portions of the Software.
        
        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
        SOFTWARE.
        
Project-URL: repository, https://github.com/eiphy/naapc
Keywords: configuration,config,dictionary,nested,argument parsing
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python
Classifier: Programming Language :: Python :: 3
Requires-Python: >=3.7
Description-Content-Type: text/markdown
License-File: LICENSE

# Nested Automated Argument Parsing Configuration (NAAPC)
[NAAPC](https://pypi.org/project/naapc/) contains two classes: NConfig and NDict.
NDict provides method to easily manipulate nested dictionaries.
NConfig is a subclass of NDict and can **automatically modify configurations according to CLI arguments**.

## Installation
```bash
pip install naapc
```

Or from source code:
```bash
pip install .
```

## Typical Usage.
Assume a configuration file test.yaml:
```yaml
task:
  task: classification
train:
  pretrain: false
  loss_args:
    lr: 0.1
```
The typical usage is as follows:
```python
from naapc import NConfig
from argparse import parser

parser.add_argument("-c", type=str, dest="config")
args, extra_args = parser.parse_known_args(["-c", "test.yaml", "--task;task", "regression", "--train;loss_args;lr", "0.2", "--train;pretrain", "1", "--others", "other"])

with open(args.config, "r") as f:
  raw = yaml.safe_load(f)
config = NConfig(raw)
extra_args = config.parse_update(parser, extra_args)
```
The resulting configurations:
```yaml
task:
  task: regression
train:
  pretrain: true
  loss_args:
    lr: 0.2
```

The data type is determined by the type in the configuration file.
The boolean data is treated as integer number 1 and 0 during parsing.

You may custom the arguments:
```yaml
task:
  task: regression
train:
  pretrain: true
  loss_args:
    lr: 0.2
_ARGUMENT_SPECIFICATION:
  task;task:
    flag: --task
    choices: ["regression", "classification"]
  train;lr:
    flag: lr
```

## NDict Usages
for a sample configuration test.yaml file:
```yaml
task:
  task: classification
train:
  loss_args:
    lr: 0.1
```
and a sample list configuration test_list.yaml file:
```yaml

```

```python
from naapc import NDict

with open("test.yaml", "r") as f:
  raw = yaml.safe_load(f)
nd = NDict(raw, delimiter=";")

nd1 = NDict.from_flatten_dict(nd.flatten_dict) # nd1 == nd
"task;path" in nd                      # "task" in raw and "path" in raw["task"]
del nd["task;path"]                    # del raw["task]["path]
nd["task;path"] = "cwd"                # raw["task"]["path"] = Path(".").absolute()
nd.flatten_dict                        # {"task;task": "classification", "train;loss_args;lr": 0.1}
nd.paths                               # ["task", "task;task", "train", "train;loss_args", "train;loss_args;lr"]
nd.get("task;seed", 1)                 # raw["task"].get("seed", 1)
nd.raw_dict                            # raw
nd.size                                # len(nd.flatten_dict)
nd.update({"task;here": "there"})      # raw["task]["here] = "there
nd.items()                             # raw.items()
nd.keys()                              # raw.keys()
nd.values()                            # raw.values()
len(nd)                                # len(raw)
bool(nd)                               # len(nd) > 0
nd1 == nd                              # nd1.flatten_dict == nd.flatten_dict
nd1["task;path"] = "xcwd"
nd1["task;extra"] = "ecwd"
nd["train;epochs"] = 100
nd.compare_dict(nd1)                   # {"task;path": ("cwd", "xcwd"), "task;extra": (None, ecwd), "train;epochs": (100, None)}
```

## NConfig Usage
NConfig only supports int, str, float, bool, and list of these types.
The NConfig automatically checks data type when modifications are applied.
Note that argument specification ("_ARGUMENT_SPECIFICATION") does not count as part of the configurations but will be saved when use save() method.
The path specified as "_IGNORE_IN_CLI" will not be added to the parser.

```python
config.save("path.yaml")               # Save configurations as a yaml file
config.add_to_argparse(parser)         # Generate cli arguments for every configuration.
config.parse_update(parser, args)      # Parse cli arguments and update corresponding configuration. Extra arguments will be returned.
```

Typical specifications are as follows:
```yaml
_ARGUMENT_SPECIFICATION:
  task;task:
    flag: --task
  task;seed:
    flag: --seed
  task;device:
    flag: -d
  data;dataset:
    choices: ["cifar", "imagenet", "asap"]
  log;comet_ml_key:
    _IGNORE_IN_CLI

```

Other functionalities are the same to NDict.

## Typing
Add a type
```python
NestedOrDict = Union[NDict, dict]
```
