Covary-encoder

This repository hosts the obfuscated runtime modules for the Covary-encoder logic, powered by TIPs-VF and with first-level protection using CodeEnigma.

OBFD tag: 03/26

About

The Covary-encoder implements a proprietary sequence encoding logic using the Translator-Interpreter Pre-seeding for Variable-length Fragments (TIPs-VF). Covary-encoder is released as obfuscated runtime ONLY. Each release has a validity period (per quarter). After expiration, users must update to the latest runtime to continue using Covary and other Covary-encoder-dependent implementations, manually or programmatically.

Encoder architecture

The Covary-encoder was built on TIPs-VF, developed by De los Santos (2025) for deep learning applications in synthetic biology. TIPs-VF is a k-mer-derived, non-overlapping, and frequency-independent encoding scheme. It represents genetic sequences based on the relative proximity and directional alignment of k-mer attributes while incorporating sequence, length, and positional awareness. TIPs-VF has demonstrated enhanced performance in truncation and fragmentation analysis, sequence homology detection, motif assessment, and splice junction identification using variable-length sequences.

Covary-encoder updates

Updates and improvements in Covary-encoder are released in a per quarter basis, unless critical security and bug issues are found that warrant unscheduled updates. Covary-encoder updates will also include significant improvements made to TIPs-VF and CodeEnigma, if available.

Obfuscation frequency

As part of security and performance upgrades on Covary-encoder, obfuscation frequency will also occur in a per quarter basis, unless specified in an obfuscation report/release.

Sample implementation (in Covary v2.0 via Colab)

Clone

# Clone Covary-encoder in Google Colab
!git clone https://github.com/mahvin92/Covary-encoder.git
%cd Covary-encoder
!git sparse-checkout init --cone
!git sparse-checkout set Active README.md

Runtime reconstruction

# Reconstruct the runtime directory
import os
%cd Active
os.mkdir("codeenigma_runtime")
!mv /content/Covary-encoder/Active/__init__.py /content/Covary-encoder/Active/codeenigma_runtime

Upload your fasta file

import os
from google.colab import files
uploaded = files.upload()
filename = list(uploaded.keys())[0]
var_name = "input_seq.fasta" # -> this is a required input name
os.rename(filename, var_name)

Run Covary-encoder

# Feed your fasta file, containing the sequences to the Covary-encoder directory
!mv /content/input_seq.fasta /content/Covary-encoder/
%cd /content/Covary-encoder/
!pip install /content/Covary-encoder/Active/codeenigma_runtime-*.whl

# Prevents premature exit
import sys, builtins
builtins.exit = sys.exit

sys.path.append("/content/Covary-encoder/Active")
import Covary_encoder

Reporting

Comments and suggestions to improve Covary-encoder are welcome. If you find any bug or problem, please open an issue.

Acknowledgements

Covary-encoder is powered by TIPs, ChordexBio and CodeEnigma, made with Python, and tested using Google Colab ❤️