# tests.py import os import tempfile import numpy as np from pathlib import Path from visualizer import DroneVisualizer from config import config from functions import DroneDetectionSystemFunctions import soundfile as sf import matplotlib.pyplot as plt # -------------------- ENVIRONMENT DETECTION -------------------- def is_notebook(): """Detect if running in Jupyter / Colab""" try: shell = get_ipython().__class__.__name__ return shell == 'ZMQInteractiveShell' except NameError: return False # -------------------- AUDIO FILE SELECTION -------------------- def select_audio_file(): """ Cross-platform audio file selection. - Notebook → uses widgets.FileUpload - Local VS Code / terminal → prompts for file path Returns: (temp_file_path, original_filename) """ if is_notebook(): import ipywidgets as widgets from IPython.display import display, clear_output uploader = widgets.FileUpload( accept='.wav,.mp3', multiple=False, description="Select Audio File" ) display(uploader) print("➡️ Please upload a file using the widget above...") # Wait until user uploads a file while not uploader.value: pass uploaded = list(uploader.value.values())[0] filename = list(uploader.value.keys())[0] content = uploaded['content'] # Save to temporary file ext = Path(filename).suffix.lower() tmp_file = Path(tempfile.gettempdir()) / f"tmp_audio{ext}" with open(tmp_file, 'wb') as f: f.write(content) return str(tmp_file), filename else: # VS Code / terminal prompt file_path = input("Enter path to audio file (.wav/.mp3): ").strip() if not os.path.isfile(file_path): raise FileNotFoundError(f"File not found: {file_path}") return file_path, os.path.basename(file_path) # -------------------- SINGLE FILE TEST -------------------- def run_single_file_test(file_path=None, threshold=0.70, analyze_long=True, show_visualization=True, true_position=None): """ Run drone detection on a single audio file. Automatically handles notebook or terminal environments. If file_path is provided, uses it directly. """ # Notebook environment: ignore file_path → use widget if file_path is None or is_notebook(): file_path, filename = select_audio_file() else: # Local VS Code / terminal file_path = file_path.strip('"') # remove quotes from command line if not os.path.isfile(file_path): raise FileNotFoundError(f"File not found: {file_path}") filename = os.path.basename(file_path) print(f"📄 File: {filename}") print(f"🎚️ Threshold: {threshold:.2f}") print(f"🔍 Long audio analysis: {'ENABLED' if analyze_long else 'DISABLED'}") # Initialize dds = DroneDetectionSystemFunctions() dds.load_best_model() visualizer = DroneVisualizer(config) if show_visualization else None # Run detection if analyze_long: result = dds.detect_and_localize_if_drone_enhanced(file_path, threshold=threshold) else: audio, sr = sf.read(file_path) is_3ch = audio.ndim == 2 and audio.shape[1] >= 3 if is_3ch: result = dds.detect_and_localize_if_drone_enhanced(file_path, threshold=threshold) else: result = dds.detect_and_localize_if_drone_enhanced(file_path, file_path, file_path, threshold=threshold) # Visualization if show_visualization and visualizer: fig, ax = visualizer.create_environment_map(true_position=true_position) if result.get("detected") and "position" in result: pos = result["position"] error = np.linalg.norm(np.array(pos) - np.array(true_position)) if true_position else None visualizer.add_estimated_position(pos, confidence=result["probability"], error=error) visualizer.add_localization_info(probability=result["probability"], filename=filename) visualizer.show() # Print summary print(f"\n✅ Detection result: {'DRONE DETECTED' if result['detected'] else 'NO DRONE'}") if result.get("position"): print(f"📍 Estimated Position: {result['position']}") if true_position: error = np.linalg.norm(np.array(result['position']) - np.array(true_position)) print(f"📏 Localization error: {error:.3f} m") # Cleanup temp file for notebook if is_notebook() and Path(file_path).parent == Path(tempfile.gettempdir()): os.unlink(file_path) # -------------------- AUTOMATED TESTS ON SYNTHETIC FILES -------------------- def run_automated_tests_on_generated_files(): """ Run automated tests on synthetic files generated by DroneDetectionSystemFunctions """ print("🧪 AUTOMATED TESTING ON GENERATED FILES") print("=" * 60) # Initialize dds = DroneDetectionSystemFunctions() dds.load_best_model() visualizer = DroneVisualizer(config) # Generate synthetic test files dds.generate_test_files() test_files = [ ("3-channel recording", "recording_3channel.wav", [0.85, 0.30]), ("Clean drone - 3 files", ["mic1.wav", "mic2.wav", "mic3.wav"], [0.85, 0.30]), ("Indoor close", ["mic1_indoor.wav", "mic2_indoor.wav", "mic3_indoor.wav"], [-0.42, 1.05]), ("Outdoor far", ["mic1_outdoor.wav", "mic2_outdoor.wav", "mic3_outdoor.wav"], [3.2, -1.1]), ] results = [] for test_name, files, true_pos in test_files: print(f"\n🔧 Testing: {test_name}") print(f"📍 Expected position: {true_pos}") try: # Determine if files is list (separate channels) if isinstance(files, list): result = dds.detect_and_localize_if_drone_enhanced(files[0], files[1], files[2], threshold=0.65) else: result = dds.detect_and_localize_if_drone_enhanced(files, threshold=0.65) # Compute error if detected position_error = None if result["detected"] and "position" in result: estimated_pos = result["position"] position_error = np.linalg.norm(np.array(estimated_pos) - np.array(true_pos)) results.append({ "test_name": test_name, "detected": result["detected"], "confidence": result["probability"], "position_error": position_error, "estimated_position": result.get("position"), "true_position": true_pos }) # Visualization fig, ax = visualizer.create_environment_map(true_position=true_pos) if result["detected"] and "position" in result: visualizer.add_estimated_position(result["position"], confidence=result["probability"], error=position_error) visualizer.add_localization_info(probability=result["probability"], filename=test_name) visualizer.show() status = "✅ DETECTED" if result["detected"] else "❌ MISSED" error_info = f" (error: {position_error:.3f}m)" if position_error else "" print(f" Result: {status} - Confidence: {result['probability']:.3f}{error_info}") except Exception as e: print(f" ❌ ERROR: {e}") import traceback traceback.print_exc() results.append({ "test_name": test_name, "detected": False, "confidence": 0.0, "position_error": None, "error": str(e) }) # Summary print("\n🎯 TEST SUMMARY") print("=" * 60) detected_count = sum(1 for r in results if r["detected"]) total_tests = len(results) print(f"📊 Overall: {detected_count}/{total_tests} tests passed") for result in results: status = "✅ DETECTED" if result["detected"] else "❌ MISSED" error_info = f" (error: {result['position_error']:.3f}m)" if result["position_error"] is not None else "" print(f" {result['test_name']:25} → {status} (conf: {result['confidence']:.3f}{error_info})") return results # -------------------- QUICK TESTS -------------------- def quick_test(): run_single_file_test(threshold=0.70, analyze_long=True, show_visualization=True) def test_with_visualization(): run_single_file_test(threshold=0.70, analyze_long=True, show_visualization=True) def test_without_visualization(): run_single_file_test(threshold=0.70, analyze_long=True, show_visualization=False) def test_high_sensitivity(): run_single_file_test(threshold=0.50, analyze_long=True, show_visualization=True) def test_low_sensitivity(): run_single_file_test(threshold=0.85, analyze_long=True, show_visualization=True) # -------------------- MAIN EXECUTION -------------------- if __name__ == "__main__": import argparse parser = argparse.ArgumentParser() parser.add_argument("--auto", action="store_true", help="Run automated tests on synthetic files") parser.add_argument("--file", type=str, help="Path to audio file for single-file test") args = parser.parse_args() if args.auto: run_automated_tests_on_generated_files() elif args.file: # Run single file test in VS Code / terminal run_single_file_test(file_path=args.file,threshold=0.7, analyze_long=True, show_visualization=True) elif is_notebook(): # Run single file test interactively in notebook run_single_file_test() else: print("❌ No file specified. Use --file or run in notebook.")