| import csv | |
| import numpy as np | |
| from scipy.stats import spearmanr, kendalltau | |
| def add_to_csv(input_files,human_score_csv): | |
| output_rows = [] | |
| for idx in [0,10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, 160, 170, 180, 190]: | |
| for filename in input_files: | |
| with open(filename, newline='', encoding='utf-8') as f: | |
| reader = csv.reader(f) | |
| header = next(reader) # 跳过表头 | |
| rows = list(reader) | |
| if idx < len(rows): | |
| for row in rows: | |
| if row[0] == f"{idx+1:04d}.png": | |
| target_row = row | |
| break | |
| # print(row) | |
| # 倒数第三和倒数第二个元素 | |
| fourth_last = target_row[-4] #reason score | |
| third_last = target_row[-3] #align score | |
| second_last = target_row[-2] #quality score | |
| output_rows.append([fourth_last,third_last, second_last]) | |
| # 2. 读取原CSV内容 | |
| with open(human_score_csv, newline='', encoding='utf-8') as f: | |
| reader = csv.reader(f) | |
| original_rows = list(reader) | |
| # 3. 检查行数是否匹配 | |
| if len(original_rows) - 1 != len(output_rows): | |
| raise ValueError("原CSV数据行数与新列数据行数不一致，请检查！") | |
| # 4. 拼接新列，写入新文件 | |
| with open(human_score_csv, 'w', newline='', encoding='utf-8') as f: | |
| writer = csv.writer(f) | |
| # 写表头 | |
| writer.writerow(original_rows[0][0:10] + ['qwen_reason','qwen_alignment', 'qwen_quality']) | |
| # 写数据行 | |
| for orig_row, new_col in zip(original_rows[1:], output_rows): | |
| writer.writerow(orig_row[0:10] + new_col) | |
| def corr(combined_csv): | |
| avg_align_list = [] | |
| avg_quality_list= [] | |
| qwen_reason_list = [] | |
| qwen_align_list = [] | |
| qwen_quality_list = [] | |
| with open(combined_csv, newline='', encoding='utf-8') as f: | |
| reader = csv.reader(f) | |
| header = next(reader) # 跳过表头 | |
| for row in reader: | |
| # 取倒数第3、4、5列 | |
| alignment_123 = [float(row[-8]), float(row[-7]), float(row[-6])] | |
| quality_123 = [float(row[-5]), float(row[-4]), float(row[-3])] | |
| alignment_avg = np.mean(alignment_123) | |
| quality_avg = np.mean(quality_123) | |
| avg_align_list.append(alignment_avg) | |
| avg_quality_list.append(quality_avg) | |
| # 取倒数第2列 | |
| qwen_reason = float(row[-3]) | |
| qwen_alignment = float(row[-2]) | |
| qwen_quality = float(row[-1]) | |
| qwen_reason_list.append(qwen_reason) | |
| qwen_align_list.append(qwen_alignment) | |
| qwen_quality_list.append(qwen_quality) | |
| # 2. 计算相关性 | |
| weight1 = 0.7 | |
| weight2 = 0.3 | |
| qwen_total_align = [weight1 * v1 + weight2 * v2 for v1, v2 in zip(qwen_reason_list, qwen_align_list)] | |
| align_spearman_corr, align_spearman_p = spearmanr(avg_align_list, qwen_total_align) | |
| align_kendall_corr, align_kendall_p = kendalltau(avg_align_list, qwen_total_align) | |
| quality_spearman_corr, quality_spearman_p = spearmanr(avg_quality_list, qwen_quality_list) | |
| quality_kendall_corr, quality_kendall_p = kendalltau(avg_quality_list, qwen_quality_list) | |
| print(f"align Spearman correlation: {align_spearman_corr:.4f}, p-value: {align_spearman_p:.4g}") | |
| print(f"align Kendall correlation: {align_kendall_corr:.4f}, p-value: {align_kendall_p:.4g}") | |
| print(f"quality Spearman correlation: {quality_spearman_corr:.4f}, p-value: {quality_spearman_p:.4g}") | |
| print(f"quality Kendall correlation: {quality_kendall_corr:.4f}, p-value: {quality_kendall_p:.4g}") | |
| if __name__ == "__main__": | |
| # input_files = [ | |
| # "/group/xihuiliu/sky/reasoning/csv/idiom/bagel.csv", | |
| # "/group/xihuiliu/sky/reasoning/csv/idiom/GPT.csv", | |
| # "/group/xihuiliu/sky/reasoning/csv/idiom/hidream.csv", | |
| # "/group/xihuiliu/sky/reasoning/csv/idiom/janus_pro_7B.csv", | |
| # "/group/xihuiliu/sky/reasoning/csv/idiom/sd30_medium.csv", | |
| # ] | |
| # human_score_csv = "/group/xihuiliu/sky/reasoning/human_eval/human_score/idiom.csv" | |
| # input_files = [ | |
| # "/group/xihuiliu/sky/reasoning/csv/text_image_new/bagel.csv", | |
| # "/group/xihuiliu/sky/reasoning/csv/text_image_new/GPT.csv", | |
| # "/group/xihuiliu/sky/reasoning/csv/text_image_new/hidream.csv", | |
| # "/group/xihuiliu/sky/reasoning/csv/text_image_new/janus_pro_7B.csv", | |
| # "/group/xihuiliu/sky/reasoning/csv/text_image_new/sd30_medium.csv", | |
| # ] | |
| # human_score_csv = "/group/xihuiliu/sky/reasoning/human_eval/human_score/text_image.csv" | |
| # input_files = [ | |
| # "/group/xihuiliu/sky/reasoning/csv/entity/bagel.csv", | |
| # "/group/xihuiliu/sky/reasoning/csv/entity/GPT.csv", | |
| # "/group/xihuiliu/sky/reasoning/csv/entity/hidream.csv", | |
| # "/group/xihuiliu/sky/reasoning/csv/entity/janus_pro_7B.csv", | |
| # "/group/xihuiliu/sky/reasoning/csv/entity/sd30_medium.csv", | |
| # ] | |
| # human_score_csv = "/group/xihuiliu/sky/reasoning/human_eval/human_score/entity.csv" | |
| input_files = [ | |
| "/group/xihuiliu/sky/reasoning/csv/physics/bagel.csv", | |
| "/group/xihuiliu/sky/reasoning/csv/physics/GPT.csv", | |
| "/group/xihuiliu/sky/reasoning/csv/physics/hidream.csv", | |
| "/group/xihuiliu/sky/reasoning/csv/physics/janus_pro_7B.csv", | |
| "/group/xihuiliu/sky/reasoning/csv/physics/sd30_medium.csv", | |
| ] | |
| human_score_csv = "/group/xihuiliu/sky/reasoning/human_eval/human_score/physics.csv" | |
| add_to_csv(input_files,human_score_csv) | |
| corr(human_score_csv) | |
| #idiom | |
| # align Spearman correlation: 0.4674, p-value: 9.474e-07 | |
| # align Kendall correlation: 0.3498, p-value: 1.911e-06 | |
| # quality Spearman correlation: 0.3305, p-value: 0.0007828 | |
| # quality Kendall correlation: 0.2622, p-value: 0.0007976 | |
| #text image | |
| # align Spearman correlation: 0.7675, p-value: 1.246e-20 | |
| # align Kendall correlation: 0.6081, p-value: 9.005e-17 | |
| # quality Spearman correlation: 0.4186, p-value: 1.465e-05 | |
| # quality Kendall correlation: 0.3257, p-value: 3.135e-05 | |
| # entity 0.9 0.1 | |
| # align Spearman correlation: 0.6091, p-value: 1.771e-11 | |
| # align Kendall correlation: 0.4673, p-value: 3.437e-10 | |
| # quality Spearman correlation: 0.3947, p-value: 4.824e-05 | |
| # quality Kendall correlation: 0.3141, p-value: 0.0001425 | |
| # entity 0.8 0.2 | |
| # align Spearman correlation: 0.6106, p-value: 1.529e-11 | |
| # align Kendall correlation: 0.4693, p-value: 2.989e-10 | |
| # quality Spearman correlation: 0.3947, p-value: 4.824e-05 | |
| # quality Kendall correlation: 0.3141, p-value: 0.0001425 | |
| # entity 0.7 0.3 THIS!!!!!!!!!!!! | |
| # align Spearman correlation: 0.6148, p-value: 1.012e-11 | |
| # align Kendall correlation: 0.4732, p-value: 2.094e-10 | |
| # quality Spearman correlation: 0.3947, p-value: 4.824e-05 | |
| # quality Kendall correlation: 0.3141, p-value: 0.0001425 | |
| # entity 0.6 0.4 | |
| # align Spearman correlation: 0.6098, p-value: 1.646e-11 | |
| # align Kendall correlation: 0.4678, p-value: 3.428e-10 | |
| # quality Spearman correlation: 0.3947, p-value: 4.824e-05 | |
| # quality Kendall correlation: 0.3141, p-value: 0.0001425 | |
| # physics 0.9 0.1 | |
| # align Spearman correlation: 0.5756, p-value: 3.746e-10 | |
| # align Kendall correlation: 0.4477, p-value: 6.166e-10 | |
| # quality Spearman correlation: 0.2652, p-value: 0.007651 | |
| # quality Kendall correlation: 0.2080, p-value: 0.007783 | |
| # physics 0.8 0.2 | |
| # align Spearman correlation: 0.5773, p-value: 3.239e-10 | |
| # align Kendall correlation: 0.4496, p-value: 5.381e-10 | |
| # quality Spearman correlation: 0.2652, p-value: 0.007651 | |
| # quality Kendall correlation: 0.2080, p-value: 0.007783 | |
| # physics 0.7 0.3 | |
| # align Spearman correlation: 0.5777, p-value: 3.131e-10 | |
| # align Kendall correlation: 0.4490, p-value: 5.498e-10 | |
| # quality Spearman correlation: 0.2652, p-value: 0.007651 | |
| # quality Kendall correlation: 0.2080, p-value: 0.007783 | |
| # physics 0.6 0.4 | |
| # align Spearman correlation: 0.5839, p-value: 1.811e-10 | |
| # align Kendall correlation: 0.4566, p-value: 3.858e-10 | |
| # quality Spearman correlation: 0.2652, p-value: 0.007651 | |
| # quality Kendall correlation: 0.2080, p-value: 0.007783 | |
| # physics 0.5 0.5 | |
| # align Spearman correlation: 0.5827, p-value: 2.008e-10 | |
| # align Kendall correlation: 0.4594, p-value: 3.317e-10 | |
| # quality Spearman correlation: 0.2652, p-value: 0.007651 | |
| # quality Kendall correlation: 0.2080, p-value: 0.007783 |