pydhcp

A Python DHCP library and server implementation.

pydhcp is pure Python and targets Python 3.9 and newer. Packet parsing and structured packet tooling are portable; actual DHCP serving still depends on OS socket permissions and platform-specific UDP behavior.

Features

• DHCP Packet Parsing — Full support for parsing and constructing DHCP packets.
• DHCP Server Base — A simple, async-friendly server foundation with overrideable lease and option policy hooks.
• DHCP Client & Capture Tools — Basic packet-client builders and a tshark-like capture command for troubleshooting.

Installation

pip install pydhcp

TOML packet encode/decode support is optional. Install pydhcp[toml] if you want pydhcp packet --format toml; JSON, YAML, and INI support remain available with the base package.

Quick start

Synchronous Server

from pydhcp.server import DhcpServer

server = DhcpServer(listen="*")
server.listen()

The built-in server intentionally keeps allocation policy small. It renews existing leases and responds to client-requested addresses, while applications can subclass DhcpServer or provide a custom lease backend for pools, reservations, and site-specific options.

You can also bind explicit endpoints or multiple ports when you do not want wildcard behavior:

server = DhcpServer(listen=[("127.0.0.1", [6767, 6768])], per_interface=True)
server.listen()

Asynchronous Server

Every listener has an asyncio counterpart -- AsyncDhcpListener, AsyncDhcpServer, AsyncDhcpRelay and AsyncDhcpCapture. They are mixins over the same receive path rather than parallel implementations, so a fix reaches both halves; a test fails if any receive-path line is duplicated between them.

import asyncio
from pydhcp.server import AsyncDhcpServer

async def main():
    server = AsyncDhcpServer()
    await server.start()      # binds and starts receiving; returns immediately
    try:
        await asyncio.Event().wait()   # serve until cancelled (Ctrl-C)
    finally:
        server.stop()

asyncio.run(main())

start() returns as soon as the sockets are bound, so something has to keep the loop alive — await your application's own work there instead of the Event if you have any. Handlers run on a worker thread, so a blocking handle() will not stall the rest of your application.

Basic Packet Client

DhcpClient is a packet-level helper for tests and troubleshooting. It sends DHCP messages and queues matching replies, but it does not configure host network interfaces.

from pydhcp.client import DhcpClient

client = DhcpClient(listen=("127.0.0.1", 6768))
discover = client.build_discover(b"\x00\x11\x22\x33\x44\x55")
client.send(discover, destination="127.0.0.1", port=6767)

Command Line Interface (CLI)

pydhcp includes a command line interface for listing network adapters, decoding packets, and starting servers.

# List all network interfaces
pydhcp interfaces

# Decode a hex-encoded DHCP packet from stdin as JSON
pydhcp packet --decode --input - --format json

# Decode a hex-encoded DHCP packet from stdin as a compact text summary
pydhcp packet --decode --input - --format summary

# Encode structured packet text back to hex
pydhcp packet --encode --input packet.json --format json --output packet.hex

# Capture DHCPDISCOVER packets as newline-delimited JSON on stdout
pydhcp capture --listen 127.0.0.1:6767 --filter msg_type=DHCPDISCOVER --output -

# Write one structured file per capture
pydhcp capture --listen 127.0.0.1:6767 --output "output/{client_id}/{timestamp}_{msg_type}.{format}" --output-mode per-capture --format json

# Invoke a trusted local command hook with each captured packet on stdin
pydhcp capture --listen 127.0.0.1:6767 --hook ./on-dhcp-capture

# Start the DHCP server from JSON or INI config
pydhcp server --config config.json

# Listen on multiple explicit endpoints while debugging
pydhcp server --listen 127.0.0.1:6767,127.0.0.1:6768

# Increase logging while debugging (-v is repeatable; --loglevel targets one logger)
pydhcp server --listen 127.0.0.1:6767 -v
pydhcp server --listen 127.0.0.1:6767 --loglevel pydhcp=DEBUG

Development

See AGENTS.md for environment setup, dependency install, and test commands.

For comprehensive validation in GitHub Actions, the test workflow also supports manual workflow_dispatch runs and safe ci-* tags. Benchmarks stay repo-local and opt-in: use the workflow's run_benchmarks input or a ci-bench-* tag when you want the benchmark harness included.

python benchmarks/run.py                 # print a summary
python benchmarks/run.py --save          # write benchmarks/results/<name>.json
python benchmarks/run.py --suite parse   # one suite instead of all

Results are min/median/max ms per call over repeated samples; see benchmarks/README.md for the schema and for why local timings are not a performance claim.

Releasing

This project follows Semantic Versioning and keeps a CHANGELOG.md. Pushing a tag matching v* triggers the release workflow.

Documentation site

MkDocs builds the API reference from docs/, published on every release. The docs also include a "Common DHCP Options" page with typed examples.

License

MIT — see LICENSE.