Skip to main content

Making apps from DAGs by just snapping your fingers

Project description

dagapp

Making apps from DAGs by just snapping your fingers

To install: pip install dagapp

What's a DAG?

What's an app?

Enough theory. Here's how it works...

A simple example

First make a DAG

from meshed.dag import DAG

def b(a):
    return 2 ** a


def d(c):
    return 10 - (5 ** c)


def result(b, d):
    return b * d
    
dag = DAG((b, d, result))

Then make an app

from dagapp.base import dag_app
from functools import partial

dags = [dag]

if __name__ == "__main__":
    app = partial(dag_app, dags=dags)
    app()

Then run the app

>>> streamlit run example.py

... and this is what you get

png

A more complicated example

Let's say we want to create two different DAGs and view them on the same streamlit page, keep all the non-root nodes static and represent the number inputs for one of the DAGs as sliders. We will use functions defined in configs_example.py for this example.

Define the DAGs

To start we can create our DAGs just like in the previous example.

profit_dag = DAG((user_clicks, rev, cost, profit))
revenue_dag = DAG((partners, clicks, revenue))

dags = [profit_dag, revnue_dag]

Define the configs

Next we can define some configs for these DAGs. These configs should be a list of dictionaries, with each dictionary representing the configs for each DAG. Each config must contain a dictionary arg_types that matches each of the root nodes in the DAG to an input type (currently the options are: num, slider, text, list, dict). If arg_types is not explicitly defined, then it will try to infer from type annotations in the function definitions, and then default to num. If slider is defined as the arg_type for any of the root nodes, then another dictionary ranges must be defined that matches each of the nodes designated to be slider with a min and max value for that slider. The configs for this example can be seen below.

configs = [
    dict(
        arg_types=dict(
            a="num",
            b="num",
            cost_per_click="num",
            revenue_per_click="num",
        ),
    ),
    dict(
        arg_types=dict(
            max_partners="slider",
            cost_per_click="slider",
            price_elasticity="slider",
            partners="slider",
            clicks_per_partner="slider",
        ),
        ranges=dict(
            max_partners=[0, 2000],
            cost_per_click=[0.0, 1.0],
            price_elasticity=[0, 200],
            partners=[0, 1500],
            clicks_per_partner=[0.0, 10.0],
        ),
    ),
]

Make the app

We can now make our app in a similar manner as the previous example, this time defining our configs, as well as defining StaticPageFunc as our page factory to keep all non-root nodes static.

from dagapp.page_funcs import StaticPageFunc

if __name__ == "__main__":
    app = partial(dag_app, dags=dags, page_factory=StaticPageFunc, configs=configs)
    app()

Run the app

>>> streamlit run configs_example.py

... and this is what you get

png png

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

dagapp-0.1.10.tar.gz (13.9 kB view details)

Uploaded Source

File details

Details for the file dagapp-0.1.10.tar.gz.

File metadata

  • Download URL: dagapp-0.1.10.tar.gz
  • Upload date:
  • Size: 13.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.1 CPython/3.8.14

File hashes

Hashes for dagapp-0.1.10.tar.gz
Algorithm Hash digest
SHA256 c4fc9c8fc0c5d5fc0eb14fa6c968adac14d8077f8750c831c671db9f8c65b964
MD5 044edd994abc85f33ba987d10d2e3a9d
BLAKE2b-256 7afedcfc8b5127a86606c3ea00702c1a459ec48b0f89c00713d4cb1dab1be8c0

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page