import pandas as pd

parquet_file = 'data/data_dump_ai_assingment.parquet'
df = pd.read_parquet(parquet_file, engine='pyarrow')

# Setting 3 random campaigns aside as testing examples for final models
campaign_ids = [8, 123, 256]
df_final_testing = df[df['campaign_id'].isin(campaign_ids)==True].copy()

def get_unique_lead_ids(df, campaign_id):
    df_campaign = df_final_testing[df_final_testing['campaign_id'] == campaign_id].copy()
    lead_ids = list(df_campaign['lead_id'].unique())
    return lead_ids

leads_8 = get_unique_lead_ids(df_final_testing, 8)
leads_123 = get_unique_lead_ids(df_final_testing, 123)
leads_256 = get_unique_lead_ids(df_final_testing, 256)

leads_8 = [str(i) for i in leads_8]
leads_123 = [str(i) for i in leads_123]
leads_256 = [str(i) for i in leads_256]