main.ipynb

################ ЯЧЕЙКА 1 ################
import numpy as np
import pandas as pd
import os

import boto3
import os
from pathlib import Path
##########################################

################ ЯЧЕЙКА 2 ################
S3_CREDS = {
    "aws_access_key_id": "........",
    "aws_secret_access_key": "......",
}

bucket = "test-bucket2"

def download_files(s3_client, bucket_name: str, local_path: str, file_name: str) -> None:
    local_path = Path(local_path)

    local_path.mkdir(parents=True, exist_ok=True)

    file_path = Path.joinpath(local_path, file_name)
    file_path.parent.mkdir(parents=True, exist_ok=True)
    s3_client.download_file(
        bucket_name,
        file_name,
        str(file_path)
    )

def get_mean_value(input_data):
    df = pd.read_csv(f'from-s3-folder/{input_data["file_name"]}')
    return df[input_data['column_name']].mean()
##########################################

########## ЯЧЕЙКА 3 для закрепа ##########
input_data = dict(file_name='diabetes.csv', column_name='age')

client = boto3.client(
    service_name='s3',
    endpoint_url='https://storage.yandexcloud.net',
    **S3_CREDS
)

download_files(
    client,
    bucket,
    "from-s3-folder",
    input_data['file_name']
)

output_data = get_mean_value(input_data)
print(output_data)
##########################################