Prerequisites

# Python 3.8+ required
python --version

# Install dependencies (if needed)
pip install -r requirements.txt

Repository Structure

ECHR_mask_modelname -> Folders containing ECHR(TAB) dataset, script, results and evaluations for each model. data -> ECHR (TAB) dataset (use input.json) edge_cases -> Contains each model script and output for edge cases dataset (test_edge.json) piranha_data_mask/data> Contains Piranha dataset piranha_data_mask_modelname -> Folders containing Piranha dataset, scripts, results and evaluations for each model.

🏃‍♂️ Run a Tool

# Example: Run Presidio (best overall performance)
cd ECHR_mask_presidio
python presidio_script.py --input ../data/input.json --output presidio_output.json

# Evaluate results
python presidio_eval.py --input presidio_output.json --out_detailed presidio_evaluation.json

📊 View Results

Results will be saved as JSON files with detailed metrics:

• True Positives (TP): Correctly identified PII
• False Positives (FP): Incorrectly flagged text
• False Negatives (FN): Missed PII entities
• Precision, Recall, F1: Performance metrics