fix #104

px4/auto-test/evaluation/eval_data_mean_stdev.py

@@ -0,0 +1,66 @@
+import argparse
+import json
+import pandas as pd
+import numpy as np
+import os
+
+def evaluate(actual_mean, actual_stdev, expected_mean, expected_stdev):
+ print(f'actual_mean: {actual_mean}')
+ print(f'actual_stdev: {actual_stdev}')
+ mean_diff_score = max(0, 100 - (abs(actual_mean - expected_mean) / expected_mean * 100))
+ stdev_diff_score = max(0, 100 - (actual_stdev / expected_stdev * 100))
+ print(f'mean_diff_score: {mean_diff_score}')
+ print(f'stdev_diff_score: {stdev_diff_score}')
+ total_score = (mean_diff_score + stdev_diff_score) / 2
+ return total_score
+
+def load_data(file_path, config):
+ data = pd.read_csv(file_path)
+ time_col = config['time_colname']
+ data_col = config['data_colname']
+
+
+ data[time_col] = (data[time_col] - data[time_col].iloc[0])
+ # タイムスタンプを秒単位に変換 (マイクロ秒から秒へ)
+ if config["time_unit"] == "usec":
+ data[time_col] = data[time_col] / 1e6
+ #print(data[1:10])
+
+ offset = config['offset_sec']
+ duration = config['duration_sec']
+ last_time = data.tail(1)[time_col].iloc[0]
+
+
+ # 相対時間を計算
+ if config['data_range_from_end']:
+ relative_start_time = last_time - offset - duration
+ relative_end_time = last_time - offset
+ else:
+ relative_start_time = 0 + offset
+ relative_end_time = 0 + offset + duration
+ print(f'relative_start_time: {relative_start_time}')
+ print(f'relative_end_time: {relative_end_time}')
+
+ # 指定された時間範囲のデータを抽出
+ data = data[(data[time_col] >= relative_start_time) & (data[time_col] <= relative_end_time)]
+ return data[data_col]
+
+def main():
+ parser = argparse.ArgumentParser(description='Evaluate the mean and standard deviation of data.')
+ parser.add_argument('input_json', help='Path to the JSON configuration file')
+ parser.add_argument('data_location_path', help='Path to the data location')
+ args = parser.parse_args()
+
+ with open(args.input_json, 'r') as json_file:
+ config = json.load(json_file)
+
+ filepath = os.path.join(args.data_location_path, config['file'])
+ data = load_data(filepath, config)
+ actual_mean = np.mean(data)
+ actual_stdev = np.std(data)
+
+ score = evaluate(actual_mean, actual_stdev, config['expected_mean_value'], config['expected_stdev_value'])
+ print(f'Score: {score}')
+
+if __name__ == '__main__':
+ main()

px4/auto-test/evaluation_input/eval_height.json

@@ -0,0 +1,11 @@
+{
+ "file": "drone_dynamics.csv",
+ "time_colname": "timestamp",
+ "time_unit": "usec",
+ "data_colname": "Z",
+ "data_range_from_end": true,
+ "offset_sec": 0,
+ "duration_sec": 5, 
+ "expected_mean_value": -5,
+ "expected_stdev_value": 0.1
+}