Skip to content
This repository has been archived by the owner on Dec 21, 2022. It is now read-only.
/ gcgc Public archive

An ML-feature processing library for biological sequences.

License

Notifications You must be signed in to change notification settings

tshauck/gcgc

Repository files navigation

GCGC

GCGC is a tool for feature processing on Biological Sequences.

code style black

Installation

GCGC is primarily intended to be used as part of a larger workflow inside Python.

To install via pip:

$ pip install gcgc

If you'd like to use code that helps gcgc's tokenizers integrate with common third party libraries, either install those packages separately, or use gcgc's extras.

$ pip install 'gcgc[hf]'

Documentation

The GCGC documentation is at gcgc.trenthauck.com, please see it for examples.

Quick Start

The easiest way to get started is to import the kmer tokenizer, configure it, then start tokenizing.

from gcgc import KmerTokenizer

kmer_tokenizer = KmerTokenizer(alphabet="unambiguous_dna")
encoded = kmer_tokenizer.encode("ATCG")
print(encoded)

sample output:

[1, 6, 7, 8, 5, 2]

This output includes the "bos" token, the "eos" token, and the four nucleotide tokens in between.

You can go the other way and convert the integers to strings.

from gcgc import KmerTokenizer

kmer_tokenizer = KmerTokenizer(alphabet="unambiguous_dna")
decoded = kmer_tokenizer.decode(kmer_tokenizer.encode("ATCG"))
print(decoded)

sample output:

['>', 'A', 'T', 'C', 'G', '<']

There's also the vocab for the kmer tokenizer.

from gcgc import KmerTokenizer

kmer_tokenizer = KmerTokenizer(alphabet="unambiguous_dna")
print(kmer_tokenizer.vocab.stoi)

sample output:

{'|': 0, '>': 1, '<': 2, '#': 3, '?': 4, 'G': 5, 'A': 6, 'T': 7, 'C': 8}