from pathlib import Path

readme_content = """

🌌 SARAD: Smart Anomaly Detection for SAR Images

SARAD is a modular, intelligent pipeline designed to detect and filter anomalous satellite images — especially noisy, artifact-heavy Synthetic Aperture Radar (SAR) data — using a fusion of classic statistical methods and deep learning.

---

🧠 What It Does

SARAD runs through the following steps:

1. Preprocessing: Cleans noisy raw images (e.g., stripes, edge glitches) with FFT filters and splits them into patches.
2. RX Detection: Uses the Reed-Xiaoli (RX) algorithm to detect statistical outliers based on Mahalanobis distance.
3. Autoencoder Detection: Reconstructs patches using a deep convolutional autoencoder and measures reconstruction error.
4. Fusion & Clustering: Combines RX and AE scores, then clusters them using DBSCAN to determine anomaly groupings.
5. Export: Saves scores, heatmaps, clustering results, and visualizations.

---

🧩 Components

• SARPreprocessor: Cleans and splits raw SAR images
• RXDetector: Computes RX anomaly scores and heatmaps
• AutoencoderAnomalyDetector: Evaluates AE-based reconstruction error
• SARAutoencoderTrainer: Trains a convolutional autoencoder on clean SAR patches
• JointAnomalyClassifier: Combines RX + AE scores and clusters using DBSCAN
• SARADPipelineManager: Orchestrates the full pipeline end-to-end

---

⚡ Quickstart

1. Preprocess Raw SAR Images

from sarad.pipeline import SARPreprocessor

pre = SARPreprocessor("data/raw", "data/cleaned")
pre.process_all()