do-data-utils 2.4.0

Functionalities to interact with Google and Azure, and clean data

do-data-utils

This package provides you the functionalities to connect to different cloud sources and data cleaning functions. Package repo on PyPI: do-data-utils - PyPI

Installation

Commands

To install the latest version from main branch, use the following command:

pip install do-data-utils

You can install a specific version, for example,

pip install do-data-utils==2.4.0

Install in requirements.txt

You can also put this source in the requirements.txt.

# requirements.txt

do-data-utils==2.4.0

Available Subpackages

• google – Utilities for Google Cloud Platform.
• azure – Utilities for Azure services.
• pathutils – Utilities related to paths.
• preprocessing – Utilities for data preprocessing.

For a full list of functions, see the overview documentation.

Example Usage

The concept of using this revolves around the idea that:

1. You keep service account JSON secrets (for cloud services) in GCP secret manager
2. You have local JSON secret file for accessing the GCP secret manager
3. Retrive the secret you want to interact with cloud platform from GCP secret manager
4. Do your stuff...

Google

GCS

Download

from do_data_utils.google import get_secret, gcs_to_df


# Load secret key and get the secret to access GCS
secret_path = 'secrets/secret-manager-key.json'
secret = get_secret(secret_id='gcs-secret-id-dev', secret=secret_path, as_json=True)

# Download a csv file to DataFrame
gcspath = 'gs://my-ai-bucket/my-path-to-csv.csv'
df = gcs_to_df(gcspath, secret, polars=False)

from do_data_utils.google import get_secret, gcs_to_dict


# Load secret key and get the secret to access GCS
secret_path = 'secrets/secret-manager-key.json'
secret = get_secret(secret_id='gcs-secret-id-dev', secret=secret_path, as_json=True)

# Download the content from GCS
gcspath = 'gs://my-ai-bucket/my-path-to-json.json'
my_dict = gcs_to_dict(gcspath, secret=secret)

Upload

from do_data_utils.google import get_secret, dict_to_json_gcs


# Load secret key and get the secret to access GCS
secret_path = 'secrets/secret-manager-key.json'

# No need to read in the secret info from version 2.3.0
with open('secrets/secret-manager-key.json', 'r') as f:
    secret_info = json.load(f)

# you can pass in either dict or path to JSON in `secret` argument
secret = get_secret(secret_id='gcs-secret-id-dev', secret=secret_info, as_json=True) 

my_setting_dict = {
    'param1': 'abc',
    'param2': 'xyz',
}

gcspath = 'gs://my-bucket/my-path-to-json.json'
dict_to_json_gcs(dict_data=my_setting_dict, gcspath=gcspath, secret=secret)

GBQ

from do_data_utils.google import get_secret, gbq_to_df


# Load secret key and get the secret to access GCS
with open('secrets/secret-manager-key.json', 'r') as f:
    secret_info = json.load(f)

# you can pass in either dict or path to JSON in `secret` argument
secret = get_secret(secret_id='gbq-secret-id-dev', secret=secret_info, as_json=True)

# Query
query = 'select * from my-project.my-dataset.my-table'
df = gbq_to_df(query, secret, polars=False)

Azure/Databricks

from do_data_utils.azure import databricks_to_df


# Load secret key and get the secret to access GCS
with open('secrets/secret-manager-key.json', 'r') as f:
    secret_info = json.load(f)

secret = get_secret(secret_id='databricks-secret-id-dev', secret=secret_info, as_json=True)

# Download from Databricks sql
query = 'select * from datadev.dsplayground.my_table'
df = databricks_to_df(query, secret, polars=False)

Path utils

from do_data_utils.pathutils import add_project_root

# Adds your root folder to sys.path,
# so you can do imports from the root directory
add_project_root(levels_up=1)

Preprocessing

from do_data_utils.preprocessing import clean_phone, clean_citizenid

phone_numbers = '090-123-4567|0912345678|0901234567-9'
phones_valid = clean_phone(phone_numbers) # Gets the valid phone numbers

citizenid = '0123456789012'
citizenid_cleaned = clean_citizenid(citizenid)