PyColonies

Python SDK for ColonyOS - build distributed applications with executors that can run anywhere.

Installation

pip3 install pycolonies

Quick Start

Starting a Colonies Server

wget https://raw.githubusercontent.com/colonyos/colonies/main/docker-compose.env
source docker-compose.env
wget https://raw.githubusercontent.com/colonyos/colonies/main/docker-compose.yml
docker-compose up -d

Submitting a Function

from pycolonies import Colonies, func_spec

client = Colonies("localhost", 50080, tls=False)
prvkey = "ddf7f7791208083b6a9ed975a72684f6406a269cfa36f1b1c32045c0a71fff05"

spec = func_spec(
    func="echo",
    args=["hello world"],
    colonyname="dev",
    executortype="container-executor",
    maxexectime=60,
    maxwaittime=60
)

process = client.submit_func_spec(spec, prvkey)
print(f"Process submitted: {process.processid}")

Implementing an Executor

from pycolonies import Colonies
from crypto import Crypto

client = Colonies("localhost", 50080, tls=False)
colony_prvkey = "ba949fa134981372d6da62b6a56f336ab4d843b22c02a4257dcf7d0d73097514"

# Generate executor keys
crypto = Crypto()
executor_prvkey = crypto.prvkey()
executorid = crypto.id(executor_prvkey)

# Register executor
executor = {
    "executorname": "my-executor",
    "executorid": executorid,
    "colonyname": "dev",
    "executortype": "python-executor"
}
client.add_executor(executor, colony_prvkey)
client.approve_executor(executorid, colony_prvkey)

# Process loop
while True:
    process = client.assign("dev", 10, executor_prvkey)
    if process:
        result = process.spec.args[0]  # Echo the first arg
        client.close(process.processid, [result], executor_prvkey)

CLI Examples

List executors:

$ colonies executor ls

╭────────────┬────────────────────┬──────────┬─────────────────────╮
│ NAME       │ TYPE               │ LOCATION │ LAST HEARD FROM     │
├────────────┼────────────────────┼──────────┼─────────────────────┤
│ dev-docker │ container-executor │ Local    │ 2025-12-27 22:13:44 │
╰────────────┴────────────────────┴──────────┴─────────────────────╯

List colonies:

$ colonies colony ls

╭──────┬──────────────────────────────────────────────────────────────────╮
│ NAME │ COLONYID                                                         │
├──────┼──────────────────────────────────────────────────────────────────┤
│ dev  │ 4787a5071856a4acf702b2ffcea422e3237a679c681314113d86139461290cf4 │
╰──────┴──────────────────────────────────────────────────────────────────╯

Architecture

flowchart TB
    subgraph Clients
        C1[Client 1]
        C2[Client 2]
    end

    subgraph "Colonies Server"
        S[(Process Queue)]
    end

    subgraph Executors
        E1[Python Executor]
        E2[Container Executor]
        E3[HPC Executor]
    end

    C1 -->|submit| S
    C2 -->|submit| S
    E1 -->|assign/close| S
    E2 -->|assign/close| S
    E3 -->|assign/close| S

Loading

Features

• Processes: Submit function specifications, assign to executors, track state
• Workflows: Create DAGs with dependencies between functions
• Channels: Real-time bidirectional communication during execution
• Blueprints: Declarative resource management with reconciliation
• File Storage: Distributed file sync with S3-compatible backends

Documentation

• API Reference - Complete API specification
• Channels - Real-time streaming between client and executor
• Blueprints - Declarative resource management
• Workflows - Computational DAGs and MapReduce patterns

Workflows

Create computational DAGs where functions execute across different executors:

from pycolonies import Colonies, func_spec, Workflow

def gen_nums(ctx={}):
    return 1, 2

def sum_nums(n1, n2, ctx={}):
    return n1 + n2

wf = Workflow(colonyname="dev")

f1 = func_spec(func=gen_nums, args=[], colonyname="dev",
               executortype="python-executor", maxexectime=60, maxwaittime=60)
wf.functionspecs.append(f1)

f2 = func_spec(func=sum_nums, args=[], colonyname="dev",
               executortype="python-executor", maxexectime=60, maxwaittime=60)
f2.conditions.dependencies.append("gen_nums")
wf.functionspecs.append(f2)

processgraph = client.submit_workflow(wf, prvkey)

flowchart LR
    G[gen_nums] -->|"(1, 2)"| S[sum_nums]
    S -->|"3"| R[Result]

Loading

Dynamic Workflows (MapReduce)

Functions can dynamically add child processes during execution:

flowchart TB
    MAP[map] -->|spawns| G1[gen_nums_0]
    MAP -->|spawns| G2[gen_nums_1]
    MAP -->|spawns| G3[gen_nums_2]
    MAP -->|spawns| G4[gen_nums_3]
    MAP -->|spawns| G5[gen_nums_4]
    G1 --> RED[reduce]
    G2 --> RED
    G3 --> RED
    G4 --> RED
    G5 --> RED
    RED -->|"15"| R[Result]

Loading

See workflows tutorial for complete examples.

Cleanup

docker-compose down --volumes

Examples

See the examples directory for complete working examples:

• echo_executor.py - Simple echo executor
• python_executor.py - Code-injection executor
• submit_echo.py - Submit echo function
• submit_python.py - Submit Python function
• workflow_example.py - Workflow submission

Testing

make setup  # Install dependencies and setup test environment
make test   # Run all tests

License

MIT