haphazard 0.1.1

A modular framework for registering and running haphazard datasets and models.

Haphazard

A Python package for haphazard dataset and model management.
Provides a standardized interface for loading datasets and models, running experiments, and extending with custom datasets or models.

---

Table of Contents

• Installation
• Quick Start
• Datasets
• Models
• Contributing
• License

---

Installation

Install via pip (after packaging):

pip install haphazard

Or for local development:

git clone <repo_url>
cd haphazard
pip install -e .

---

Project Structure

The Haphazard package has a modular layout:

haphazard/
├── __init__.py                    # Top-level package
├── data/                          # Dataset related modules
│   ├── __init__.py
│   ├── base_dataset.py            # Abstract BaseDataset class
│   └── datasets/                  # All dataset implementations
│       ├── __init__.py
│       └── dummy_dataset/
│           └── __init__.py
├── models/                        # Model related modules
│   ├── __init__.py
│   ├── base_model.py              # Abstract BaseModel class
│   └── model_zoo/                 # All model implementations
│       ├── __init__.py
│       └── dummy_model/
│           └── __init__.py
└── utils/                         # Optional helper functions
    └── ...

Notes:

• data/base_dataset.py defines BaseDataset.
• data/datasets/ contains registered datasets; each dataset is a submodule with __init__.py.
• models/base_model.py defines BaseModel.
• models/model_zoo/ contains registered models; each model is a submodule with __init__.py.
• utils/ is optional, for shared helpers.

This layout allows dynamic registration of datasets and models via decorators.

---

Quick Start

from haphazard import load_dataset, load_model

# Load dataset
dataset = load_dataset("dummy", n_samples=100, n_features=10)

# Load model
model = load_model("dummy")

# Run model
outputs = model(dataset)
print(outputs)

---

Datasets

• All datasets must inherit from BaseDataset.
• Example dataset: DummyDataset.
• Main interface:

from haphazard import load_dataset

dataset = load_dataset("dummy", base_path="./data")
x, y = dataset.load_data()
mask = dataset.load_mask(scheme="probabilistic", availability_prob=0.5)

Dataset Attributes

• name : str — Dataset name.
• task : "classification" | "regression".
• haphazard_type : "controlled" | "intrinsic".
• n_samples, n_features : int.
• num_classes : int (for classification).

---

Models

• All models must inherit from BaseModel.
• Example model: DummyModel.
• Main interface:

from haphazard import load_model

model = load_model("dummy")
outputs = model(dataset)

Output

• Classification: labels, preds, logits, time_taken, is_logit.
• Regression: targets, preds, time_taken.

---

Contributing

Haphazard is designed for easy extensibility. You can add new datasets and models.

Adding a new dataset

1. Create a new folder under haphazard/data/datasets/, e.g., my_dataset/.
2. Add __init__.py:

from ...base_dataset import BaseDataset
from ...datasets import register_dataset
import numpy as np

@register_dataset("my_dataset")
class MyDataset(BaseDataset):
    def __init__(self, base_path="./", **kwargs):
        self.name = "my_dataset"
        self.haphazard_type = "controlled"
        self.task = "classification"
        super().__init__(base_path=base_path, **kwargs)

    def read_data(self, base_path="./"):
        # Load or generate x, y
        x = np.random.random((100, 10))
        y = np.random.randint(0, 2, 100)
        return x, y

3. The dataset is automatically registered and can be loaded with load_dataset("my_dataset").

Adding a new model

1. Create a new folder under haphazard/models/model_zoo/, e.g., my_model/.
2. Add __init__.py:

from ...base_model import BaseModel, BaseDataset
from ...model_zoo import register_model
import numpy as np

@register_model("my_model")
class MyModel(BaseModel):
    def __init__(self, **kwargs):
        self.name = "MyModel"
        self.tasks = {"classification", "regression"}
        self.deterministic = True
        self.hyperparameters = set()
        super().__init__(**kwargs)

    def fit(self, dataset: BaseDataset, mask_params=None, model_params=None, seed=42):
        # Dummy implementation
        x, y = dataset.load_data()
        mask = dataset.load_mask(**mask_params)
        preds = np.random.randint(0, 2, size=y.shape[0])
        if dataset.task == "classification":
            return {
                "labels": y,
                "preds": preds,
                "logits": preds.astype(float),
                "time_taken": 0.0,
                "is_logit": True
            }
        elif dataset.task == "regression":
            return {
                "targets": y,
                "preds": preds,
                "time_taken": 0.0,
            }

3. The model is automatically registered and can be loaded with load_model("my_model").

---

License

MIT License.