Metadata-Version: 2.4
Name: imspy-predictors
Version: 0.5.0
Summary: ML-based predictors for CCS, retention time, and fragment intensity in mass spectrometry.
License-Expression: MIT
Author: theGreatHerrLebert
Author-email: davidteschner@googlemail.com
Requires-Python: >=3.11,<3.14
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Provides-Extra: all
Provides-Extra: cuda
Provides-Extra: koina
Provides-Extra: training
Requires-Dist: datasets (>=2.0.0) ; extra == "all"
Requires-Dist: datasets (>=2.0.0) ; extra == "training"
Requires-Dist: imspy-core (>=0.4.0)
Requires-Dist: koinapy (>=0.0.10) ; extra == "all"
Requires-Dist: koinapy (>=0.0.10) ; extra == "koina"
Requires-Dist: numba (>=0.57.0)
Requires-Dist: numpy (>=1.21.0)
Requires-Dist: pandas (>=1.3.0)
Requires-Dist: scikit-learn (>=1.0)
Requires-Dist: scipy (>=1.7.1)
Requires-Dist: torch (>=2.0.0)
Requires-Dist: torch (>=2.0.0) ; extra == "cuda"
Requires-Dist: tqdm (>=4.66)
Requires-Dist: wandb (>=0.12.1) ; extra == "all"
Requires-Dist: wandb (>=0.12.1) ; extra == "training"
Description-Content-Type: text/markdown

# imspy-predictors

ML-based predictors for CCS, retention time, and fragment intensity in mass spectrometry.

## Installation

```bash
pip install imspy-predictors
```

For remote model access via Koina servers:

```bash
pip install imspy-predictors[koina]
```

## Features

- **CCS Prediction**: Deep learning models for collision cross section / ion mobility prediction
- **Retention Time Prediction**: GRU-based retention time predictors
- **Fragment Intensity Prediction**: Prosit 2023 timsTOF intensity predictor
- **Charge State Prediction**: Binomial and deep learning charge state distribution models
- **Koina Integration**: Access remote prediction models via Koina servers (optional)

## Quick Start

```python
from imspy_predictors import (
    load_deep_ccs_predictor,
    load_deep_retention_time_predictor,
    Prosit2023TimsTofWrapper,
)

# Load CCS predictor
ccs_model = load_deep_ccs_predictor()

# Load RT predictor
rt_model = load_deep_retention_time_predictor()

# Load intensity predictor
intensity_model = Prosit2023TimsTofWrapper()
```

## Submodules

- **ccs/**: CCS / ion mobility prediction
- **rt/**: Retention time prediction
- **intensity/**: Fragment intensity prediction (Prosit)
- **ionization/**: Charge state distribution prediction
- **koina_models/**: Koina remote model access (requires `koinapy`)
- **utilities/**: Tokenizers for ML models

## Dependencies

- **imspy-core**: Core data structures (required)
- **TensorFlow**: Deep learning framework (required)
- **dlomix**: Deep learning for omics (required)
- **koinapy**: Koina API client (optional, for remote models)

## Optional Dependencies

Some functionality requires additional packages:

- **imspy-search**: For PSM-based predictions using sagepy
- **imspy-simulation**: For simulation utilities (e.g., flatten_prosit_array)

## Related Packages

- **imspy-core**: Core data structures and timsTOF readers
- **imspy-search**: Database search functionality
- **imspy-simulation**: Simulation tools for timsTOF data
- **imspy-vis**: Visualization tools

## License

MIT License - see LICENSE file for details.

