hip-data-tools
© Hipages Group Pty Ltd 2019-2022
Common Python tools and utilities for data engineering, ETL, Exploration, etc. The package is uploaded to PyPi for easy drop and use in various environmnets, such as (but not limited to):
- Running production workloads
- ML Training in Jupyter like notebooks
- Local machine for dev and exploration
Installation
Install from PyPi repo:
pip3 install hip-data-tools
Install from source
pip3 install .
MacOS Dependencies
brew install libev
brew install librdkafka
Connect to aws
You will need to instantiate an AWS Connection:
from hip_data_tools.aws.common import AwsConnectionManager, AwsConnectionSettings, AwsSecretsManager
# to connect using an aws cli profile
conn = AwsConnectionManager(AwsConnectionSettings(region="ap-southeast-2", secrets_manager=None, profile="default"))
# OR if you want to connect using the standard aws environment variables
conn = AwsConnectionManager(settings=AwsConnectionSettings(region="ap-southeast-2", secrets_manager=AwsSecretsManager(), profile=None))
# OR if you want custom set of env vars to connect
conn = AwsConnectionManager(
settings=AwsConnectionSettings(
region="ap-southeast-2",
secrets_manager=AwsSecretsManager(
access_key_id_var="SOME_CUSTOM_AWS_ACCESS_KEY_ID",
secret_access_key_var="SOME_CUSTOM_AWS_SECRET_ACCESS_KEY",
use_session_token=True,
aws_session_token_var="SOME_CUSTOM_AWS_SESSION_TOKEN"
),
profile=None,
)
)
Using this connection to object you can use the aws utilities, for example aws Athena:
from hip_data_tools.aws.athena import AthenaUtil
au = AthenaUtil(database="default", conn=conn, output_bucket="example", output_key="tmp/scratch/")
result = au.run_query("SELECT * FROM temp limit 10", return_result=True)
print(result)
Connect to Cassandra
from cassandra.policies import DCAwareRoundRobinPolicy
from cassandra.cqlengine import columns
from cassandra.cqlengine.management import sync_table
from cassandra.cqlengine.models import Model
from cassandra import ConsistencyLevel
load_balancing_policy = DCAwareRoundRobinPolicy(local_dc='AWS_VPC_AP_SOUTHEAST_2')
conn = CassandraConnectionManager(
settings = CassandraConnectionSettings(
cluster_ips=["1.1.1.1", "2.2.2.2"],
port=9042,
load_balancing_policy=load_balancing_policy,
),
consistency_level=ConsistencyLevel.LOCAL_QUORUM
)
conn = CassandraConnectionManager(
CassandraConnectionSettings(
cluster_ips=["1.1.1.1", "2.2.2.2"],
port=9042,
load_balancing_policy=load_balancing_policy,
secrets_manager=CassandraSecretsManager(
username_var="MY_CUSTOM_USERNAME_ENV_VAR"),
),
consistency_level=ConsistencyLevel.LOCAL_ONE
)
# For running Cassandra model operations
conn.setup_connection("dev_space")
class ExampleModel(Model):
example_type = columns.Integer(primary_key=True)
created_at = columns.DateTime()
description = columns.Text(required=False)
sync_table(ExampleModel)
Connect to Google Sheets
How to connect
You need to go to Google developer console and get credentials. Then the Google sheet need to be shared with client email. GoogleApiConnectionSettings need to be provided with the Google API credentials key json. Then you can access the Google sheet by using the workbook_url and the sheet name.
How to instantiate Sheet Util
You can instantiate Sheet Util by providing GoogleSheetConnectionManager, workbook_url and the sheet name.
sheet_util = SheetUtil(
conn_manager=GoogleSheetConnectionManager(self.settings.source_connection_settings),
workbook_url='https://docs.google.com/spreadsheets/d/cKyrzCBLfsQM/edit?usp=sharing',
sheet='Sheet1')
How to read a dataframe using SheetUtil
You can get the data in the Google sheet as a Pandas DataFrame using the SheetUtil. We have defined a template for the Google sheet to use with this utility.
You need to provide the "field_names_row_number" and "field_types_row_number" to call "get_dataframe()" method in SheetUtil.
sheet_data = sheet_util.get_data_frame(
field_names_row_number=8,
field_types_row_number=7,
row_range="12:20",
data_start_row_number=9)
You can use load_sheet_to_athena() function to load Google sheet data into an Athena table.
GoogleSheetToAthena(GoogleSheetsToAthenaSettings(
source_workbook_url='https://docs.google.com/spreadsheets/d/cKyrzCBLfsQM/edit?usp=sharing',
source_sheet='spec_example',
source_row_range=None,
source_fields=None,
source_field_names_row_number=5,
source_field_types_row_number=4,
source_data_start_row_number=6,
source_connection_settings=get_google_connection_settings(gcp_conn_id=GCP_CONN_ID),
manual_partition_key_value={"column": "start_date", "value": START_DATE},
target_database=athena_util.database,
target_table_name=TABLE_NAME,
target_s3_bucket=s3_util.bucket,
target_s3_dir=s3_dir,
target_connection_settings=get_aws_connection_settings(aws_conn_id=AWS_CONN_ID),
target_table_ddl_progress=False
)).load_sheet_to_athena()
There is an integration test called "integration_test_should__load_sheet_to_athena__when_using_sheetUtil" to test this functionality. You can simply run it by removing the "integration_" prefix.