Metadata-Version: 2.4
Name: softgraph
Version: 0.3.0
Summary: Graph ML in one call: node/edge/temporal + MLP/GCN/SAGE (optional) + useful plots.
Project-URL: Homepage, https://github.com/soft-tech-talks/softgraph
Author-email: Soft Tech Talks <hello@softtechtalks.org>
License: MIT
Requires-Python: >=3.9
Requires-Dist: matplotlib>=3.7
Requires-Dist: networkx>=3.2
Requires-Dist: numpy>=1.24
Requires-Dist: pandas>=2.1
Requires-Dist: scikit-learn>=1.3
Requires-Dist: scipy>=1.10
Provides-Extra: deep
Requires-Dist: torch>=2.1; extra == 'deep'
Description-Content-Type: text/markdown

# softgraph (one-call API) — v0.3.0

**One call** to generate data → extract features → train a model → evaluate and plot.

## Install (editable dev)
```bash
pip install -e .
```

Optional deep:
```bash
pip install "softgraph[deep]"
# Then install torch-geometric from wheels if you want GCN/SAGE.
```

## Usage
```python
import softgraph

run = softgraph.softgraph(
    task="node_classification",
    dataset="sbm",
    n=800, k=4, p_in=0.10, p_out=0.02,
    features="spectral:16",
    model="mlp",
    plot=True,      # confusion matrix
    curves=True     # learning curve
)
print(run["metrics"])
```
