"""实验 B：随机按行划分与按新样品划分，回答的是不同问题。"""
import csv
import json
import numpy as np
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split, GroupShuffleSplit
from sklearn.neighbors import KNeighborsRegressor
from sklearn.preprocessing import StandardScaler
from sklearn.pipeline import make_pipeline
from sklearn.metrics import mean_absolute_error
from common import OUT

rng = np.random.default_rng(42)
# 每个样品具有一个可重复测得的“指纹”，响应在本例中是随机的，不能外推。
n_samples, repeats = 80, 8
fingerprints = rng.normal(size=(n_samples, 6))
responses = rng.normal(size=n_samples)
groups = np.repeat(np.arange(n_samples), repeats)
X = fingerprints[groups] + rng.normal(0, .005, (len(groups), 6))
y = responses[groups] + rng.normal(0, .02, len(groups))
with (OUT / 'simulated_repeats.csv').open('w', newline='', encoding='utf-8') as f:
    w = csv.writer(f)
    w.writerow(['sample_id'] + [f'fingerprint_{i}' for i in range(6)] + ['teaching_response'])
    w.writerows(np.column_stack([groups, X, y]))
row_train, row_test = train_test_split(np.arange(len(y)), test_size=.25, random_state=42)
group_train, group_test = next(GroupShuffleSplit(n_splits=1, test_size=.25, random_state=42).split(X, y, groups))
results = {}
for name, train, test in [('random_rows', row_train, row_test), ('new_samples', group_train, group_test)]:
    model = make_pipeline(StandardScaler(), KNeighborsRegressor(n_neighbors=1))
    model.fit(X[train], y[train])
    overlap = len(set(groups[train]) & set(groups[test]))
    results[name] = {'mae': float(mean_absolute_error(y[test], model.predict(X[test]))), 'overlapping_samples': overlap}
assert results['new_samples']['overlapping_samples'] == 0
fig, ax = plt.subplots(figsize=(6, 4), constrained_layout=True)
ax.bar(list(results), [v['mae'] for v in results.values()])
ax.set(ylabel='Test MAE', title='Same model, different generalization questions')
fig.savefig(OUT / 'b_split_comparison.png', dpi=150)
plt.close(fig)
(OUT / 'b_results.json').write_text(json.dumps(results, indent=2), encoding='utf-8')
print(json.dumps(results, indent=2))
