Metadata-Version: 2.4
Name: neuralhash
Version: 0.1.0
Summary: Generates a hash from a reverse engineered Apple NeuralHash model
Author-email: Josh Terrill <josh@hacked.codes>
License: MIT
Requires-Python: >=3.8
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: onnxruntime
Requires-Dist: pillow
Requires-Dist: numpy
Dynamic: license-file

# Python NeuralHash

This is a library that exposes the [Apple Neuralhash](https://www.apple.com/child-safety/pdf/CSAM_Detection_Technical_Summary.pdf) model, which can be used to determine similar images, even if the images are not exactly the same.

## Installation

From pip:

```bash
pip install neuralhash
```

From source:

```bash
git clone https://github.com/yourname/neuralhash.git
cd neuralhash
pip install -e .
```

## Usage

Python API:

```python
from neuralhash import get_neuralhash_hex

hash_value = get_neuralhash_hex("tests/sample.png")
print(hash_value)
# ab13fcd08dd3b0a31b07fa2e
```

CLI:

After installation, you also get a command-line tool:

```bash
neuralhash path/to/image.png
# path/to/image.png: a0cca3edec8339d006068f1d
```

Hash multiple files at once:

```bash
neuralhash image1.png image2.jpg
# image1.png: a0cca3edec8339d006068f1d
# image2.png: ab13fcd08dd3b0a31b07fa2e
```

## License

MIT
