Add demo script for quick testing

Author: fkalghan

examples/demo.py

@@ -0,0 +1,94 @@
+#!/usr/bin/env python
+"""Demo script to test LocalMod functionality."""
+
+import json
+import sys
+import time
+
+# Check if running with models downloaded
+def main():
+    print("=" * 60)
+    print("LocalMod Demo - Content Moderation API")
+    print("=" * 60)
+    
+    from localmod import SafetyPipeline
+    from localmod.classifiers import PIIDetector
+    
+    # Initialize pipeline (PII only for quick demo - no model downloads needed)
+    print("\n[1] Initializing PII detector (no ML model needed)...")
+    pii = PIIDetector()
+    pii.load()
+    print("    ✓ PII detector ready")
+    
+    # Test PII Detection
+    print("\n[2] Testing PII Detection...")
+    test_cases = [
+        "My email is john.doe@example.com",
+        "Call me at 555-123-4567",
+        "SSN: 123-45-6789",
+        "Credit card: 4111-1111-1111-1111",
+        "Hello, how are you today?",  # No PII
+    ]
+    
+    for text in test_cases:
+        result = pii.predict(text)
+        status = "🚨 FLAGGED" if result.flagged else "✅ Safe"
+        print(f"    {status}: \"{text[:40]}...\"")
+        if result.flagged:
+            print(f"       Types: {result.categories}")
+    
+    # Test PII Redaction
+    print("\n[3] Testing PII Redaction...")
+    text = "Contact John at john@email.com or 555-123-4567"
+    redacted, _ = pii.redact(text)
+    print(f"    Original:  {text}")
+    print(f"    Redacted:  {redacted}")
+    
+    # Full pipeline test (requires model downloads)
+    print("\n[4] Testing Full Pipeline (requires models)...")
+    try:
+        pipeline = SafetyPipeline(classifiers=["pii"])  # Just PII for quick test
+        report = pipeline.analyze(
+            "My email is test@example.com and I hate everyone!",
+            include_explanation=True
+        )
+        print(f"    Flagged: {report.flagged}")
+        print(f"    Severity: {report.severity.value}")
+        print(f"    Summary: {report.summary}")
+        print(f"    Time: {report.processing_time_ms:.2f}ms")
+    except Exception as e:
+        print(f"    ⚠ Full pipeline test skipped: {e}")
+    
+    # Test with ML models if available
+    print("\n[5] Testing ML-based classifiers (if models are downloaded)...")
+    try:
+        from localmod.classifiers import ToxicityClassifier
+        toxicity = ToxicityClassifier(device="cpu")
+        toxicity.load()
+        
+        toxic_texts = [
+            "You're a complete idiot!",
+            "I hope you have a wonderful day!",
+            "Die in a fire you moron",
+        ]
+        
+        print("    Toxicity classifier loaded!")
+        for text in toxic_texts:
+            result = toxicity.predict(text)
+            status = "🚨 TOXIC" if result.flagged else "✅ Safe"
+            print(f"    {status} ({result.confidence:.2%}): \"{text[:40]}\"")
+            
+    except Exception as e:
+        print(f"    ⚠ ML classifiers not available: {type(e).__name__}")
+        print("    Run 'python -m localmod.cli download' to download models")
+    
+    print("\n" + "=" * 60)
+    print("Demo Complete!")
+    print("=" * 60)
+    
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())
+