recommendation.py

from sklearn.preprocessing import MinMaxScaler
from helper import get_scores_optimized


def calculate_final_score(df, job_details):
    # print('Inside final score')
    columns_to_normalize = ['matched_skills', 'experience', 'education_relevance', 'trait_flag']
    df_scored=get_scores_optimized(df,job_details)    
    scaler = MinMaxScaler()
    # print(f'lets see the columns: {df_scored.columns}')
    df_scored[columns_to_normalize] = scaler.fit_transform(df_scored[columns_to_normalize])
    # print(f'scoring done !!!')
    
    # Define weights
    weights = {
        'matched_skills': 0.3,
        'experience': 0.3,
        'education_relevance': 0.3,
        'trait_flag': 0.1
    }
    
    # Compute final score as a weighted sum
    df['final_score'] = (
        df_scored['matched_skills'] * weights['matched_skills'] +
        df_scored['experience'] * weights['experience'] +
        df_scored['education_relevance'] * weights['education_relevance'] +
        df_scored['trait_flag'] * weights['trait_flag']
    )
    df_sorted = df.sort_values(by='final_score', ascending=False)
    df_no_full_duplicates = df_sorted.drop_duplicates(keep="first")
    return df_no_full_duplicates