Iupac

[heliziii]

Adding support for IUPAC nucleotide codes to EnsembleTR.

[aryarm]

code looks great! Since the tests seem good and the code makes sense, I tried to see if I could help with speed, but I'm not sure if that's even necessary. Maybe you have a way of benchmarking the runtime?

[aryarm]

You can delete the setup.py file now that you have a pyproject.toml

[aryarm]

could you memoize this function to get a speedup? or would that require too much memory?

Suggested change

	def GetAllPossibleSequences(motif):
	r"""
	Computing all the possible sequences that a motif can be considering IUPAC characters.
	For example, a motif with sequence RGG can be both AGG and GGG. Divide and conquer method is
	used to form all possible sequences.
	Parameters
	----------
	motif : str
	motif
	Returns
	-------
	All possible sequences : list of str
	"""
	possible_seqs = []
	if len(motif) < 1:
	return []
	elif len(motif) == 1:
	if motif[0] in IUPAC_map_dict:
	return IUPAC_map_dict[motif[0]]
	else:
	return motif[0]
	else:
	first_part = GetAllPossibleSequences(motif[0:int(len(motif)/2)])
	second_part = GetAllPossibleSequences(motif[int(len(motif)/2):])
	for seq1 in first_part:
	for seq2 in second_part:
	possible_seqs.append(seq1 + seq2)
	return possible_seqs
	@lru_cache(maxsize=None)
	def GetAllPossibleSequences(motif):
	r"""
	Computing all the possible sequences that a motif can be considering IUPAC characters.
	For example, a motif with sequence RGG can be both AGG and GGG. Divide and conquer method is
	used to form all possible sequences.
	Parameters
	----------
	motif : str
	motif
	Returns
	-------
	All possible sequences : set of str
	"""
	if len(motif) == 1:
	return IUPAC_map_dict.get(motif, [motif])
	first_part = GetAllPossibleSequences(motif[:len(motif)//2])
	second_part = GetAllPossibleSequences(motif[len(motif)//2:])
	return {f1 + f2 for f1 in first_part for f2 in second_part}

you might need to add this at the top:

from functools import lru_cache

[aryarm]

it might be slightly faster to use sets instead of lists?

Suggested change

	potential_sequences_1 = [utils.GetCanonicalMotif(motif) for motif in GetAllPossibleSequences(motif_1)]
	potential_sequences_2 = [utils.GetCanonicalMotif(motif) for motif in GetAllPossibleSequences(motif_2)]
	potential_sequences_1 = {utils.GetCanonicalMotif(motif) for motif in GetAllPossibleSequences(motif_1)}
	potential_sequences_2 = {utils.GetCanonicalMotif(motif) for motif in GetAllPossibleSequences(motif_2)}