DetectMate Performance

Library to run high performance operations in DetectMate. It combines C++ functionality with Polars DataFrames.

Installation

CMake allows to build the C++ code:

sudo apt-get install cmake

Intall python dependencies:

pip install uv
uv sync
prek install

Compile the code, make sure that all unit tests are green:

sh compile.sh

To run the C++ and Python unit tests:

pytest

To do a local instalation in a different venv

uv pip install --no-cache-dir <directory_detectmateperformance>

Code example

import detectmateperformance as matep


dataset = {
    "name": "AIT - Audit",
    "path_temp": "data/audit_templates.txt",
    "path_logs": "data/audit.log",
    "regex": r"type=(?P<Type>\w+) msg=audit\((?P<Time>[^:]+):(?P<Serial>\d+)\): (?P<Content>.*)"
}


results = matep.metrics.evaluate(
    logs=dataset["path_logs"],
    ground_templates=dataset["path_temp"],
    templates=dataset["path_temp"],
    regex=dataset["regex"]
)

print(results)  # {'PA': 1.0, 'NED': 1.0, 'TS': 1.0, 'GA': 1.0, 'FR': 1.0}