Schedula

Common errors

• ModuleNotFoundError: No module named 'schedula'

  cause The 'schedula' package has not been installed in your current Python environment or the environment where it was installed is not active.
  fix Run `pip install schedula` in your terminal to install the package, or activate the correct virtual environment if you've already installed it.

• AttributeError: 'Dispatcher' object has no attribute 'plot' (or 'web', 'to_server', etc.)

  cause You are attempting to use a feature that is part of an optional 'extra' (e.g., `plot`, `web`) without having installed `schedula` with that specific extra.
  fix Install the library with the required extra. For example, to enable plotting, use `pip install 'schedula[plot]'`. For all extras, use `pip install 'schedula[all]'`.

• FileNotFoundError: [Errno 2] No such file or directory: 'dot' (when using plot functionality)

  cause The `plot` functionality in `schedula` relies on the Graphviz command-line tools, specifically the `dot` executable. This error indicates that `dot` is not found in your system's PATH.
  fix Install the Graphviz system package (e.g., via your OS package manager or from graphviz.org/download/) and ensure its installation directory, containing `dot`, is added to your system's PATH environment variable.

Warnings

• gotcha Many advanced functionalities (like plotting, web API generation, or specific parallel execution strategies) are provided via 'extras'. These must be explicitly installed using `pip install 'schedula[extra_name]'` or `pip install 'schedula[all]'`.
  Fix: Review the documentation for required extras and install them, e.g., `pip install 'schedula[plot]'`.
• gotcha The 'plot' extra, which enables visualizing your dataflow models, relies on the external Graphviz system library. Installing `schedula[plot]` only installs the Python bindings; Graphviz itself (and its `dot` executable) must be installed separately on your operating system and be accessible in your system's PATH.
  Fix: Install Graphviz from its official website (graphviz.org/download/) and ensure its `bin` directory is added to your system's PATH.
• gotcha While `schedula` supports asynchronous and parallel dispatching, there is an inherent performance cost associated with creating and managing threads/processes. For very short-duration tasks, this overhead might negate the benefits of concurrency.
  Fix: Benchmark your specific use case to determine if concurrent execution provides a performance gain or if a synchronous execution model is more efficient for your workload.

Install

• pip install schedula Install base package
• pip install 'schedula[all]' Install with all optional dependencies

Imports

• Dispatcher

  import schedula as sh
  dsp = sh.Dispatcher(...)

Quickstart

This quickstart demonstrates how to define a simple dataflow model using `schedula.Dispatcher`, add functions with data dependencies using the `@sh.add_function` decorator, and then execute the workflow by calling `dispatch` with initial inputs and desired outputs. Schedula automatically determines the execution order based on data dependencies.

import schedula as sh

# 1. Define a Dispatcher, which is the main model for your dataflow
dsp = sh.Dispatcher(name='my_simple_workflow')

# 2. Define functions and add them to the Dispatcher using the decorator
@sh.add_function(dsp, outputs=['y'])
def add_one(x):
    print(f"Executing add_one with x={x}")
    return x + 1

@sh.add_function(dsp, outputs=['z'])
def multiply_by_two(y):
    print(f"Executing multiply_by_two with y={y}")
    return y * 2

# 3. Dispatch the workflow with initial inputs and desired outputs
initial_inputs = {'x': 5}
desired_outputs = ['z']

try:
    results = dsp.dispatch(inputs=initial_inputs, outputs=desired_outputs)
    print(f"\nFinal result for 'z': {results['z']}")
except Exception as e:
    print(f"An error occurred during dispatch: {e}")