Pengembangan, Pemaketan, dan Publikasi Perangkat Lunak Forensik Log Penerbangan Drone pada Organisasi DroneNLP

Arianza, Reihan and Aziz, Rafi Faheem (2026) Pengembangan, Pemaketan, dan Publikasi Perangkat Lunak Forensik Log Penerbangan Drone pada Organisasi DroneNLP. Project Report. [s.n.]. (Unpublished)

Text 5025231274_5025231116-Project_Report.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Flight log merupakan salah satu artefak digital terpenting dalam investigasi forensik insiden drone karena merekam pesan-pesan peristiwa dalam bahasa alami yang menggambarkan kondisi drone selama penerbangan. Organisasi DroneNLP di GitHub mengembangkan sejumlah perangkat lunak sumber terbuka untuk menganalisis pesan flight log drone berbasis Natural Language Processing (NLP), di antaranya DFLER, ADFLER, dan LogNexus. Pada saat kerja praktik ini dimulai, ketiga repositori tersebut masih berupa research code yang belum dapat dipasang sebagai paket Python, belum memiliki antarmuka command-line yang dapat diotomasi, belum dilengkapi alur CI/CD, serta belum memiliki dokumentasi penggunaan yang terpusat. Selain itu, LogNexus belum memuat pipeline inferensi identifikasi masalah yang telah dikembangkan pada repositori riset SoPID. Kerja praktik ini melakukan refactoring terhadap ketiga repositori ke dalam format paket Python standar berbasis pyproject.toml, membangun antarmuka command-line menggunakan pustaka argparse, mengotomasi pengunduhan model machine learning dari Hugging Face, mengimplementasikan alur CI/CD berbasis GitHub Actions dengan penerbitan paket otomatis ke PyPI, melakukan porting pipeline inferensi SoPID ke dalam struktur codebase LogNexus, menyusun pengujian otomatis menggunakan pytest, serta menyusun dokumentasi penggunaan pada halaman dokumentasi DroneNLP berbasis MkDocs. Hasil pengujian menunjukkan bahwa ketiga paket berhasil dibangun dan diterbitkan ke PyPI, yaitu DFLER v0.1.3, ADFLER v0.1.1, dan LogNexus v1.0.0, sehingga dapat dipasang oleh pengguna hanya dengan perintah pip install. Pipeline pengujian otomatis berhasil menjalankan seluruh modul uji pytest dengan status lulus, dan dokumentasi penggunaan ketiga perangkat lunak kini tersedia secara publik melalui halaman dokumentasi DroneNLP. Capaian ini menjadikan perangkat forensik drone hasil penelitian lebih mudah diadopsi oleh investigator maupun peneliti tanpa memerlukan konfigurasi lingkungan yang rumit.
==================================================================================================================================
Flight logs are among the most important digital artifacts in drone incident forensic investigations because they record natural-language event messages that describe the drone's condition throughout a flight. The DroneNLP organization on GitHub develops several open-source software tools for analyzing drone flight log messages using Natural Language Processing (NLP), including DFLER, ADFLER, and LogNexus. At the beginning of this internship, all three repositories were still in the form of research code that could not be installed as Python packages, lacked automated command-line interfaces, did not include CI/CD pipelines, and had no centralized user documentation. In addition, LogNexus did not yet incorporate the problem identification inference pipeline previously developed in the SoPID research repository. This internship refactored the three repositories into the standard Python package format based on pyproject.toml, developed a command-line interface using the argparse library, automated the download of machine learning models from Hugging Face, implemented a GitHub Actions-based CI/CD pipeline with automatic package publishing to PyPI, ported the SoPID inference pipeline into the LogNexus codebase, developed automated tests using pytest, and prepared user documentation on the MkDocs-based DroneNLP documentation website. The testing results showed that all three packages were successfully built and published to PyPI—DFLER v0.1.3, ADFLER v0.1.1, and LogNexus v1.0.0—allowing users to install them simply by running the pip install command. The automated testing pipeline successfully executed all pytest test modules with passing results, and the documentation for all three tools is now publicly available through the DroneNLP documentation website. These achievements make the research-based drone forensic tools significantly easier for investigators and researchers to adopt without requiring complex environment configuration.

Item Type:	Monograph (Project Report)
Uncontrolled Keywords:	argparse, CI/CD, Forensik Drone, GitHub Actions, Named Entity Recognition, PyPI, Python
Subjects:	Q Science > QA Mathematics > QA76.758 Software engineering Q Science > QA Mathematics > QA76.9.A25 Computer security. Digital forensic. Data encryption (Computer science)
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Reihan Arianza
Date Deposited:	07 Jul 2026 06:46
Last Modified:	07 Jul 2026 06:46
URI:	http://repository.its.ac.id/id/eprint/134375

Actions (login required)

View Item