Skip to content

tomlegkov/marine-python

Repository files navigation

Marine Python Client

Python client for Marine

Installation

Clone the repo and run:

python setup.py install

Usage

Basic Usage (filter and parse packet)

import marine

passed, result = marine.filter_and_parse_packet(b"your packet", "ip host 1.1.1.1", "tcp.port == 80", ["ip.src", "ip.dst"])
if passed:
    print(f"{result['ip.src']} -> {result['ip.dst']}")

Other Available API

Filter packet

passed = marine.filter_packet(b"your packet", "ip host 1.1.1.1", "tcp.port == 80")
if not passed:
    print("Packet didn't pass filter")

Parse packet

result = marine.parse_packet(b"your packet", ["macro.ip.src", "tcp.port"], {"macro.ip.src": ["ip.src", "ipv6.src"]})
print(f"Parsed IP: {result['macro.ip.src']} and port: {result['tcp.port']}")

BPF validation

validation_result = marine.validate_bpf("ip host 1.1.1.1")
if not validation_result:
    print(f"BPF validation error: {validation_result.error}")

Display filter validation

validation_result = marine.validate_display_filter("tcp.port == 80")
if not validation_result:
    print(f"Display filter validation error: {validation_result.error}")

Fields validation

validation_result = marine.validate_fields(["macro.ip.src", "tcp.port"], {"macro.ip.src": ["ip.src", "ipv6.src"]})
if not validation_result:
    print(f"The following fields don't exist: {validation_result.errors}")

Pool

MarinePool allows to run multiple instances of Marine using multiple cores. The exported API is identical to Marine's: filter, parse, filter_and_parse.

pool = MarinePool(process_count=4)
parsed_packets = pool.filter_and_parse(packets, bpf="udp", fields=["macro.ip.src", "udp.port"], field_templates={"macro.ip.src": ["ip.src", "ipv6.src"]})
for passed, result in parsed_packets:
    if passed:
        print(f"Parsed IP: {result['macro.ip.src']} and UDP port: {result['udp.port']}")

Advanced

For advanced usages (not recommended) see the get_marine() function and the Marine class .

Contributing

Guidelines

Syntax formatting is done using Black

Running Tests

The tests are written using pytest. To run the tests, you need to provide the library file (libmarine.so) and its dependencies. marine expects libmarine.so to reside in marine/.wslibs. Inside a development environment, you can simply create a link from marine/.wslibs to where you compile marine-core.

Then, simply run tox.

Additionally, syntax is checked with flake8 by running flake8 marine tests from the root directory of the project.

Packaging

An x86_64 manylinux2014 wheel is built by our CI, based on a manylinux image supplied by marine-core and patched to meet our needs. There is currently no support for installing on other platforms.

About

Python client for Marine

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published