first commit

Code/5-ANOVA-SciPy.py

@@ -0,0 +1,150 @@
+import numpy as np
+import openpyxl
+import csv
+import os.path
+from scipy.stats import f_oneway
+
+
+def read_value(patient_dir, atlas, data_type):
+    folder_name = os.path.basename(patient_dir)
+    csv_file_path = os.path.join(patient_dir, f"{data_type}_{atlas}_result", "result_atlas.csv")
+
+    result = [folder_name]
+    with open(csv_file_path, mode='r') as data_csv_file:
+        csv_reader = csv.DictReader(data_csv_file)
+        for row in csv_reader:
+            mean = float(row["meanValue"])
+            result.append(mean)
+
+    return result
+
+
+def read_header(patient_dir, atlas, data_type):
+    header = ['编号']
+
+    csv_file_path = os.path.join(patient_dir, f"{data_type}_{atlas}_result", "result_atlas.csv")
+
+    with open(csv_file_path, mode='r') as data_csv_file:
+        csv_reader = csv.DictReader(data_csv_file)
+        for row in csv_reader:
+            region_name = row["Chinese Name"]
+            header.append(region_name)
+
+    return header
+
+
+def generate_stats_row(name, region_size, function, patient_size, extra_parameter):
+    result = [name]
+
+    for i in range(region_size):
+        letter = openpyxl.utils.cell.get_column_letter(i + 2)
+        result.append(f'={function}({letter}2:{letter}{patient_size + 1}{extra_parameter})')
+
+    return result
+
+
+def write_to_work_sheet(work_sheet, patient_list, atlas, data_type):
+    if len(patient_list) == 0:
+        return
+
+    header = read_header(patient_list[0], atlas, data_type)
+
+    work_sheet.append(header)
+
+    for patient_dir in patient_list:
+        print(data_type, patient_dir)
+        patient_data = read_value(patient_dir, atlas, data_type)
+
+        work_sheet.append(patient_data)
+
+    region_size = len(header) - 1
+    work_sheet.append([])
+    work_sheet.append([])
+    work_sheet.append(generate_stats_row('均值', region_size, 'AVERAGE', len(patient_list), ''))
+    work_sheet.append(generate_stats_row('标准差', region_size, 'STDEV', len(patient_list), ''))
+    work_sheet.append(generate_stats_row('最小值', region_size, 'MIN', len(patient_list), ''))
+    work_sheet.append(generate_stats_row('5%', region_size, 'PERCENTILE', len(patient_list), ',0.05'))
+    work_sheet.append(generate_stats_row('25%', region_size, 'PERCENTILE', len(patient_list), ',0.25'))
+    work_sheet.append(generate_stats_row('50%', region_size, 'PERCENTILE', len(patient_list), ',0.50'))
+    work_sheet.append(generate_stats_row('75%', region_size, 'PERCENTILE', len(patient_list), ',0.75'))
+    work_sheet.append(generate_stats_row('95%', region_size, 'PERCENTILE', len(patient_list), ',0.95'))
+    work_sheet.append(generate_stats_row('最大值', region_size, 'MAX', len(patient_list), ''))
+
+
+def write_anova(workbook, work_sheet, groups, atlas, data_type):
+    header = ['', 'F', 'p']
+    work_sheet.append(header)
+
+    sheet_dict = {}
+    for group_name, group in groups.items():
+        sheet_dict[group_name] = f'{group_name}_{atlas}_{data_type}'
+
+    first_group_sheet = workbook[list(sheet_dict.values())[0]]
+    for column_num in range(2, first_group_sheet.max_column + 1):
+        data_list = []
+        data_row = [first_group_sheet.cell(row=1, column=column_num).value]
+
+        for group_name, sheet_name in sheet_dict.items():
+            group_data = []
+            data_count = len(groups[group_name])
+            for i in range(2, data_count + 2):
+                data_value = workbook[sheet_name].cell(row=i, column=column_num).value
+                if not np.isnan(data_value):
+                    group_data.append(data_value)
+            data_list.append(group_data)
+
+        if len(data_list[0]) == 0:
+            continue
+
+        f_value, p_value = f_oneway(*data_list)
+        data_row += [f_value, p_value]
+        work_sheet.append(data_row)
+
+
+def main():
+    data_types = ['corr-CBF']
+    atlas_list = ['AnImage_AAL3']
+
+    output_file = r'..\Data\csv-ANOVA-SciPy.xlsx'
+    group_file = r'..\Data\group.xlsx'
+    patient_root = r'..\Data\csv-data'
+
+    groups = {}
+
+    group_workbook = openpyxl.load_workbook(group_file)
+    for sheet_name in group_workbook.sheetnames:
+        group = []
+        group_sheet = group_workbook[sheet_name]
+        for i in range(1, group_sheet.max_row + 1):
+            patient_id = group_sheet.cell(row=i, column=1).value
+            patient_path = os.path.join(patient_root, patient_id)
+            group.append(patient_path)
+        groups[sheet_name] = group
+
+    all_group = []
+    for group in groups.values():
+        all_group += group
+
+    work_book = openpyxl.Workbook()
+    work_book.remove(work_book.active)
+
+    for data_type in data_types:
+        for atlas in atlas_list:
+            atlas_new_name = atlas.replace("AnImage_", "")
+            for group_key in groups.keys():
+                work_sheet = work_book.create_sheet(title=f'{group_key}_{atlas_new_name}_{data_type}')
+                write_to_work_sheet(work_sheet, groups[group_key], atlas, data_type)
+
+            work_sheet = work_book.create_sheet(title=f'ALL_{atlas_new_name}_{data_type}')
+            write_to_work_sheet(work_sheet, all_group, atlas, data_type)
+
+    for data_type in data_types:
+        for atlas in atlas_list:
+            atlas_new_name = atlas.replace("AnImage_", "")
+            work_sheet = work_book.create_sheet(title=f'ANOVA_{atlas_new_name}_{data_type}')
+            write_anova(work_book, work_sheet, groups, atlas_new_name, data_type)
+
+    work_book.save(output_file)
+
+
+main()