This directory contains different examples of using Kaskada. The included examples should work with the latest release of Kaskada and the latest Python Client released to PyPi. However, there are some examples that may be pending/in-development (see section below). The intent is to ensure these examples are kept up to date and are working reference points for users.
There may be some examples that are demonstrate functionality not officially available yet. To view the latest functionality, it may require building the binaries + clients locally to run.
Note: Using local builds requires informing the python client to use the local build (by default uses the latest release). This is achieved by passing the download(False)
build param. See below:
from kaskada.api.session import LocalBuilder
# download(False) skips the fetch of the latest
session = LocalBuilder().download(False).build()
To run everything (build services + build client + run jupyter): make start/jupyter
To build the binaries locally:
make build/services
- Builds all the services.make build/manager
- Builds the manager/API binary.make build/engine
- Builds the compute engine binary.
To build the python client locally:
make venv/build-install-clients
- Build and install the clientsmake venv/create
- Create virtual environment and install dependenciesmake venv/build-client-python
- Builds the python client locally.make venv/install-client-python
- Builds and installs the python client locally.
To clean:
make clean
- Cleans everything (notebooks, build, and virtual environment)make build/clean
- Cleans the built binaries locally and from cache directorymake venv/notebook/clean
- Cleans all output from notebook (use prior to committing)make venv/clean
- Nukes the virtual environment