FlowerPower

Simple Workflow Framework - Hamilton + APScheduler = FlowerPower

---

📚 Table of Contents

1. Overview
2. Installation
3. Getting Started
   • Initialize Project
   • Add Pipeline
   • Setup Pipeline
   • Run Pipeline
   • Schedule Pipeline
   • Start Worker
   • Track Pipeline
4. Development
   • Dev Services

---

🔍 Overview

FlowerPower is a simple workflow framework based on two fantastic Python libraries:

• Hamilton: Creates DAGs from your pipeline functions
• APScheduler: Handles pipeline scheduling

Key Features

• 🔄 Pipeline Workflows: Create and execute complex DAG-based workflows
• ⏰ Scheduling: Run pipelines at specific times or intervals
• ⚙️ Parameterization: Easily configure pipeline parameters
• 📊 Tracking: Monitor executions with Hamilton UI
• 🛠️ Flexible Configuration: Simple YAML-based setup
• 📡 Distributed Execution: Support for distributed environments

More details in Hamilton docs

---

📦 Installation

# Basic installation
pip install flowerpower

# With scheduling support
pip install "flowerpower[scheduler]"

# Additional components
pip install "flowerpower[scheduler,mqtt]"    # MQTT broker
pip install "flowerpower[scheduler,redis]"   # Redis broker
pip install "flowerpower[scheduler,mongodb]" # MongoDB store
pip install "flowerpower[scheduler,ray]"     # Ray computing
pip install "flowerpower[scheduler,dask]"    # Dask computing

---

🚀 Getting Started

Initialize Project

Option 1: Command Line

flowerpower init new-project
cd new-project

Option 2: Python

from flowerpower import init
init("new-project")

This creates basic config files:

• conf/project.yml

📦 Optional: Project Management with UV (Recommended)

It is recommended to use the project manager uv to manage your project dependencies.

Installation

pip install uv

For more installation options, visit: https://docs.astral.sh/uv/getting-started/installation/

Project Initialization

uv init --app --no-readme --vcs git

---

Pipeline Management

Creating a New Pipeline

Option 1: Command Line

flowerpower add my_flow
# or
flowerpower new my_flow

Option 2: Python

# Using PipelineManager
from flowerpower.pipeline import PipelineManager
pm = PipelineManager()
pm.new("my_flow")

# Or using the new function directly
from flowerpower.pipeline import new
new("my_flow")

This creates the new pipeline and configuration file:

• pipelines/my_flow.py
• conf/pipelines/my_flow.yml

Setting Up a Pipeline

1. Add Pipeline Functions Build your pipeline by adding the functions (nodes) to pipelines/my_flow.py that build the DAG, following the Hamilton paradigm.

2. Parameterize Functions

You can parameterize functions in two ways:

Method 1: Default Values

def add_int_col(
    df: pd.DataFrame,
    col_name: str = "foo",
    values: str = "bar"
) -> pd.DataFrame:
    return df.assign(**{col_name: values})

Method 2: Configuration File

In conf/pipelines/my_flow.yml:

...
func:
  add_int_col:
    col_name: foo
    values: bar
...

Add the @parameterize decorator to the function in your pipeline file:

@parameterize(**PARAMS.add_int_col)
def add_int_col(
    df: pd.DataFrame,
    col_name: str,
    values: int
) -> pd.DataFrame:
    return df.assign(**{col_name: values})

---

Running Pipelines

Configuration

You can configure the pipeline parameters inputs, and final_vars, and other parameters in the pipeline configuration file conf/pipelines/my_flow.yml or directly in the pipeline execution function.

Using the Pipeline Configuration

...
run:
  inputs:
    data_path: path/to/data.csv
    fs_protocol: local
  final_vars: [add_int_col, final_df]
  # optional parameters
  with_tracker: false
  executor: threadpool # or processpool, ray, dask
...

Execution Methods

There are three ways to execute a pipeline:

1. Direct Execution

   • Runs in current process
   • No data store required

2. Job Execution

   • Runs as APScheduler job
   • Returns job results
   • Requires data store and event broker

3. Async Job Addition

   • Adds to APScheduler
   • Returns job ID
   • Results retrievable from data store

Command Line Usage

# Note: add --inputs and --final-vars and other optional parameters if not specified in the config file
# Direct execution
flowerpower run my_flow
# Job execution
flowerpower run-job my_flow

# Add as scheduled job
flowerpower add-job my_flow

You can also use the --inputs and --final-vars flags to override the configuration file parameters or if they are not specified in the configuration file.

flowerpower run my_flow \
    --inputs data_path=path/to/data.csv,fs_protocol=local \
    --final-vars final_df \
    --executor threadpool
    --without-tracker

Python Usage

from flowerpower.pipeline import Pipeline, run, run_job, add_job

# Using Pipeline class
p = Pipeline("my_flow")
# Note: add inputs, final_vars, and other optional arguments if not specified in the config file
result = p.run()
result = p.run_job()
job_id = p.add_job()

# Using functions
result = run("my_flow")
result = run_job("my_flow")
job_id = add_job("my_flow")

You can also use the inputs and final-vars arguments to override the configuration file parameters or if they are not specified in the configuration file.

result = run(
    "my_flow",
    inputs={
        "data_path": "path/to/data.csv",
        "fs_protocol": "local"
    },
    final_vars=["final_df"],
    executor="threadpool",
    with_tracker=False
)

---

⏰ Scheduling Pipelines

Setting Up Schedules

Command Line Options

# Run every 30 seconds
flowerpower schedule my_flow \
    --type interval \
    --interval-params seconds=30

# Run at specific date/time
flowerpower schedule my_flow \
    --type date \
    --date-params year=2022,month=1,day=1,hour=0,minute=0,second=0

# Run with cron parameters
flowerpower schedule my_flow \
    --type cron \
    --cron-params second=0,minute=0,hour=0,day=1,month=1,day_of_week=0

# Run with crontab expression
flowerpower schedule my_flow \
    --type cron \
    --crontab "0 0 1 1 0"

Python Usage

from flowerpower.scheduler import schedule, Pipeline

# Using Pipeline class
p = Pipeline("my_flow")
p.schedule("interval", seconds=30)

# Using schedule function
schedule("my_flow", "interval", seconds=30)

---

👷 Worker Management

Starting a Worker

Command Line

flowerpower start-worker

Python

# Using the SchedulerManager class
from flowerpower.scheduler import SchedulerManager
sm = SchedulerManager()
sm.start_worker()

# Using the start_worker function
from flowerpower.scheduler import start_worker
start_worker()

Worker Configuration

Configure your worker in conf/project.yml:

# PostgreSQL Configuration
data_store:
  type: postgres
  uri: postgresql+asyncpq://user:password@localhost:5432/flowerpower

# Redis Event Broker
event_broker:
  type: redis
  uri: redis://localhost:6379
  # Alternative configuration:
  # host: localhost
  # port: 6379

Alternative Data Store Options

SQLite

data_store:
  type: sqlite
  uri: sqlite+aiosqlite:///flowerpower.db

MySQL

data_store:
  type: mysql
  uri: mysql+aiomysql://user:password@localhost:3306/flowerpower

MongoDB

data_store:
  type: mongodb
  uri: mongodb://localhost:27017/flowerpower

In-Memory

data_store:
  type: memory

Alternative Event Broker Options

MQTT

event_broker:
  type: mqtt
  host: localhost
  port: 1883
  username: edge  # optional
  password: edge  # optional

Redis

event_broker:
  type: redis
  uri: redis://localhost:6379
  # Alternative configuration:
  # host: localhost
  # port: 6379

In-Memory

event_broker:
  type: memory

---

📊 Pipeline Tracking

Hamilton UI Setup

Local Installation

# Install UI package
pip install "flowerpower[ui]"

# Start UI server
flowerpower hamilton-ui

Access the UI at: http://localhost:8241

Docker Installation

# Clone Hamilton repository
git clone https://github.com/dagworks-inc/hamilton
cd hamilton/ui

# Start UI server
./run.sh

Access the UI at: http://localhost:8242

Tracker Configuration

Configure tracking in conf/project.yml:

username: my_email@example.com
api_url: http://localhost:8241
ui_url: http://localhost:8242
api_key: optional_key

And specify the tracker parameter in the pipeline configuration `conf/pipelines/my_flow.yml:

...
tracker:
  project_id: 1
  tags:
    environment: dev
    version: 1.0
  dag_name: my_flow_123
...

---

🛠️ Development Services

Local Development Setup

Download the docker-compose configuration:

curl -O https://raw.githubusercontent.com/legout/flowerpower/main/docker/docker-compose.yml

Starting Services

# MQTT Broker (EMQX)
docker-compose up mqtt -d

# Redis
docker-compose up redis -d

# MongoDB
docker-compose up mongodb -d

# PostgreSQL
docker-compose up postgres -d

---

📝 License

MIT License

---

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

---

📫 Support

For support, please open an issue in the GitHub repository.