greenflow - RAPIDS Financial Services Algorithms
Project description
greenflow - Graph Computation Toolkit
What is greenflow?
greenflow is a tool that helps you to organize the workflows.
- It define a TaskGraph file format
.gq.yaml
that describes the workflow. It can be edited easily bygreenflowlab
JupyterLab plugin. - Dynamically compute the input-output ports compatibility, dataframe columns names and types, ports types to prevent connection errors.
- Nodes can have multiple output ports that can be used to generate different output types. E.g. some data loader Node provides both
cudf
anddask_cudf
output ports. The multiple GPUs distributed computation computation is automatically enabled by switching to thedask_cudf
output port. - Provides the standard API to extend your computation Nodes.
- The composite node can encapsulate the TaskGraph into a single node for easy reuse. The composite node can be exported as a regular greenflow node without any coding.
- greenflow can be extended by writing a plugin with a set of nodes for a particular domain. Check
plugins
for examples.
These examples can be used as-is or, as they are open source, can be extended to suit your environments.
Binary pip installation
To install the greenflow graph computation library, run:
pip install greenflow
Or install greenflow
at the root directory:
pip install .
greenflow node plugins can be registered in two ways:
- (Recommended)Write a external plugin using 'entry point' to register it. Check the
external
directory for details - Register the plugin in
greenflowrc
file. Check theSystem environment
for details
System environment
There are a few system environment that the user can overwrite.
The custom module files are specified in the greenflowrc
file. GREENFLOW_CONFIG
enviroment variable points to the location of this file. By default, it points to
$CWD\greenflowrc
.
In the example greenflowrc
, system environment variable MODULEPATH
is used to point to the paths of the module files.
To start the jupyterlab, please make sure MODULEPATH
is set properly.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file greenflow-1.0.5.tar.gz
.
File metadata
- Download URL: greenflow-1.0.5.tar.gz
- Upload date:
- Size: 41.9 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.1.1 pkginfo/1.4.2 requests/2.22.0 setuptools/45.2.0 requests-toolbelt/0.8.0 tqdm/4.30.0 CPython/3.8.10
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | fa8c4c4cfcbe39d5d1e13e2f435f59c1fa506d2099adcfa4d428e60360de6354 |
|
MD5 | c8d6090c7005c2d27ed9b0eb5d464bcc |
|
BLAKE2b-256 | ce6f9e1ef09c7183c7a305655a5d55c654b5f3eeae1b21b1cfa78948b3f8ad85 |
File details
Details for the file greenflow-1.0.5-py3-none-any.whl
.
File metadata
- Download URL: greenflow-1.0.5-py3-none-any.whl
- Upload date:
- Size: 50.2 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.1.1 pkginfo/1.4.2 requests/2.22.0 setuptools/45.2.0 requests-toolbelt/0.8.0 tqdm/4.30.0 CPython/3.8.10
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 9e56b31eeb122dbbd4d9360688aa81f1eb466933c65b4f974ae094e90999ab76 |
|
MD5 | 0956d7ad0990ee5cca403c60bf289b41 |
|
BLAKE2b-256 | 57731bb4ba12fb515833e167919c42c4967310d0573197da81f0fa344004bc19 |