No project description provided
Project description
Flojoy / Python
The Python repo for Flojoy infrastructure. Including the client for Flojoy's desktop app, the @flojoy
decorator, and various utilities for node creation.
The @flojoy
decorator
Usage:
from scipy import signal
import numpy as np
from flojoy import flojoy, DataContainer
@flojoy
def BUTTER(v, params):
''' Apply a butterworth filter to an input vector '''
logger.debug('Butterworth inputs:', v)
x = v[0].x
sig = v[0].y
sos = signal.butter(10, 15, 'hp', fs=1000, output='sos')
filtered = signal.sosfilt(sig, sig)
return DataContainer(x = x, y = filtered)
The @flojoy
decorator automatically injects vector(s) passed from the previous node, as well as any control parameters set in the CTRL panel UI.
Please see https://github.com/flojoy-io/flojoy-desktop/tree/main/PYTHON/FUNCTIONS for more usage examples.
Supported data types
See https://github.com/flojoy-io/flojoy-python/issues/4
Reconcilers
Flojoy nodes should try to accomodate any reasonable combination of inputs that a first-time Flojoy Studio user might try.
For example, the ADD node should make a best effort to do something reasonable when a matrix is added to a DataFrame, or a 2 matrices of a different size are added.
For this reason, we've created the Reconciler
class to handle the process of turning different data types into compatible, easily added objects.
An Example
Here's an example of the Reconciler in action from the ADD node. Just initialize the Reconciler and pass it pairs of primitives from within your DataContainers.
def ADD(dc_inputs: list[DataContainer], params: dict) -> DataContainer:
...
dc_reconciler = Reconciler()
cur_sum = dc_inputs[0]
for operand_n in dc_inputs[1:]:
# reconcile the types of the inputs before calling numpy
reconciled_lhs, reconciled_rhs = dc_reconciler.reconcile(cur_sum.y, operand_n.y)
cur_sum = np.add(reconciled_lhs, reconciled_rhs)
return DataContainer(x=dc_inputs[0].x, y=cur_sum)
Customizing your Reconcilers for different Node Categories
Have a different idea about how you want your DataContainers to be interoperable? Subclass Reconciler
and override the specific type pairs you're interested in.
For example, let's say we're writing image processing nodes, and we want special behavior when dealing with certain image-like DataContainers.
class ImageProcessingReconciler(Reconciler):
def __init__():
super().__init__()
# override only for the pair of types you're interested in!
def image__image(self, lhs, rhs):
# special handling for reconciling two different images
...
return reconciled_lhs, reconciled_rhs
def grayscale__image(self, lhs, rhs):
# special handling for reconciling a grayscale and an image
...
return reconciled_lhs, reconciled_rhs
Publish Package on PYPI
Uploading file via Twine
You have to install twine
package first if not installed already.
To install twine
run following command:
pip install twine
- Update version in setup.py. For example:
version = '0.0.1'
for prod version andversion = '0.0.1-dev'
for dev version release, - Run following command to make a distribution file:
python3 setup.py sdist
- To upload files on
PYPI
, run following command:twine upload dist/*
Note: You'll need username
and password
to make a release. Please get in touch with the team for credentials.
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 flojoy-0.4.1.tar.gz
.
File metadata
- Download URL: flojoy-0.4.1.tar.gz
- Upload date:
- Size: 44.2 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: poetry/1.8.2 CPython/3.11.9 Darwin/23.4.0
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | c29ba6d7e498ccfe6e45a2382fae1fbeec14411364ac7ecb5d37b2388bf331a3 |
|
MD5 | 017a4af998d72a24e47f726a35fc6006 |
|
BLAKE2b-256 | 7530e828eeda1049dfda983476580d3ca57ed16384d69352b8476cd7393db7bc |
File details
Details for the file flojoy-0.4.1-py3-none-any.whl
.
File metadata
- Download URL: flojoy-0.4.1-py3-none-any.whl
- Upload date:
- Size: 51.2 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: poetry/1.8.2 CPython/3.11.9 Darwin/23.4.0
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 4a15f195d4feca055c4ec111a17edd541f73007c8755e6a6e6d65fee94f27d29 |
|
MD5 | a25ce1c11f0c2b086cf9f6122acdaf7c |
|
BLAKE2b-256 | 383878e44fd59cdee85867073affe1932ace7e63b33cb617c09566cb62926381 |