import pandas as pd
import re

df = pd.read_csv('Data_2024_Tracer_study.csv', sep=';', encoding='utf-8-sig')
cols = list(df.columns)

def normalize(col_name):
    if pd.isna(col_name):
        return ""
    return str(col_name).strip().lower()

def find_column(source_cols, possible_names):
    source_cols_norm = [(normalize(c), c) for c in source_cols]
    for name in possible_names:
        name_norm = normalize(name)
        # Exact match
        for norm_col, orig_col in source_cols_norm:
            if norm_col == name_norm:
                return orig_col
        # Partial match - cari kode di dalam string
        kode_match = re.search(r'[fF]\d+[a-zA-Z0-9]*', name)
        if kode_match:
            kode = kode_match.group(0).lower()
            for norm_col, orig_col in source_cols_norm:
                if kode in norm_col:
                    return orig_col
    return None

# Test
tests = ['F8', 'F502', 'F1761', 'F401', 'F1001', 'f8', 'f502', 'F5a1', 'F1102']
for t in tests:
    found = find_column(cols, [t])
    print(f"find_column(['{t}']) = {repr(found)}")