namara-python

Python SDK for Namara API and namara-pandas

Installation

You need Python 3.7 or above and the package manager pip to install dependencies.

pip3 install -r requirements.txt

Docker is used in this repo only for integration with Jenkins CI, you don't need to run through docker at all.

Authorization

Can read API key from environment variables or directly passed into Client. The Namara API urls are also read from environement variables

export NAMARA_API_KEY=my_namara_api_key
export NAMARA_API_URL=https://api.ee-dev.namara.io
export NAMARA_AUTH_URL=https://account.ee-dev.namara.io

In Python

from namara_python.client import Client

client = Client()

OR

Pass them in directly

from namara_python.client import Client

client = Client(
    api_key='my_namara_api_key',
    auth_url='https://account.ee-dev.namara.io',
    api_url='https://api.ee-dev.namara.io'
)

The client object is now ready to make requests to Namara API

Environment variables

You can create a .env file to manage environment variables more easily, namara-python will read those.

in .env

export NAMARA_API_KEY=my_namara_api_key
export NAMARA_API_URL=https://api.ee-dev.namara.io
export NAMARA_AUTH_URL=https://account.ee-dev.namara.io

Get

You can pass the query statement directly

from namara_python.client import Client, ExportFailedError

c = Client(
    api_key='my_namara_api_key',
    auth_url='https://account.ee-dev.namara.io',
    api_url='https://api.ee-dev.namara.io'
)

try:
    res = client.get(query='SELECT * FROM f78871ca-71f9-4071-859c-7d6ba414972e AS ACM')
    print(res)
except ExportFailedError as e:
    print(e)

datasets = c.list_datasets()
print(datasets)

Put

You can load a dataframe directly onto namara under a given organization id

...
client.put(data_frame=df, dataset_name:'test_dataset', organization_id:<ORG_UUID> warehouse_name:=None, reference_id:=None)

Development

Running tests

make test

You can run a specific test by marking it in the test file, above the function:

@pytest.mark.only # <- add this
def only_test_to_run():
    ...

And then run

make test_only

Static analysis

This project uses Pythons type annotations for some static analysis for types.

Install mypy

pip install mypy

To do type checking

mypy --ignore-missing-imports .

We skip import checks just becuase mypy doesn't quite work as you would expect for imports, especially third party packages

Testing code locally

There's scratch file that you could use to import code and quickly test out the sdk.

There's one already called tester.py, fill in your api_key and a valid dataset id and your setup to start testing things out!

Run:

ipython -i tester.py

to get an interactive shell with tester.py code having already been run.

Pandas extension

Look here: https://github.com/thinkdata-works/namara-python/tree/develop/namara_pandas