Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Update sequence.py #47

Open
wants to merge 4 commits into
base: master
Choose a base branch
from
Open

Update sequence.py #47

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
208fd7f
Update sequence.py
stefanches7 Sep 30, 2019
69f33d6
Fix discrepancies/default no target score specificity
stefanches7 Oct 2, 2019
60a9bfb
fixup! Fix discrepancies/default no target score specificity addition…
stefanches7 Oct 2, 2019
0e7cc23
add logical constraints on labels length
stefanches7 Oct 2, 2019
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
29 changes: 19 additions & 10 deletions kipoiseq/dataloaders/sequence.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,23 +1,19 @@
from collections import OrderedDict
import pandas as pd
import numpy as np
from copy import deepcopy

from kipoi.metadata import GenomicRanges
from kipoi.specs import DataLoaderArgument, ArraySpecialType
from kipoi.plugin import is_installed
from kipoi.data import Dataset, kipoi_dataloader
from kipoi_conda.dependencies import Dependencies
from kipoi.specs import Author
from kipoi_utils.utils import default_kwargs

from kipoiseq.extractors import FastaStringExtractor
from kipoiseq.transforms import SwapAxes, DummyAxis, Compose, OneHot, ReorderedOneHot
from kipoiseq.transforms import ReorderedOneHot
from kipoiseq.transforms.functional import resize_interval
from kipoiseq.utils import to_scalar, parse_dtype

import pybedtools
from pybedtools import BedTool, Interval

# general dependencies
# bioconda::genomelake', TODO - add genomelake again once it gets released with pyfaidx to bioconda
Expand Down Expand Up @@ -165,6 +161,8 @@ class StringSeqIntervalDl(Dataset):
doc: Force uppercase output of sequences
ignore_targets:
doc: if True, don't return any target variables
target_only_score:
doc: if True, only 'score' column of BED file is returned in targets array
output_schema:
inputs:
name: seq
Expand Down Expand Up @@ -194,7 +192,8 @@ def __init__(self,
# max_seq_len=None,
# use_strand=False,
force_upper=True,
ignore_targets=False):
ignore_targets=False,
target_only_score=False):

self.num_chr_fasta = num_chr_fasta
self.intervals_file = intervals_file
Expand All @@ -209,6 +208,7 @@ def __init__(self,
# bed_columns = 6
# else:
# bed_columns = 3
self.target_only_score = target_only_score

self.bed = BedDataset(self.intervals_file,
num_chr=self.num_chr_fasta,
Expand All @@ -226,7 +226,7 @@ def __getitem__(self, idx):
force_upper=self.force_upper)

interval, labels = self.bed[idx]

if self.auto_resize_len:
# automatically resize the sequence to cerat
interval = resize_interval(interval, self.auto_resize_len, anchor='center')
Expand All @@ -237,7 +237,13 @@ def __getitem__(self, idx):

# Run the fasta extractor and transform if necessary
seq = self.fasta_extractors.extract(interval)


if len(labels) >= 2 and labels[2]=="-": #reverse strand
seq = seq[::-1]

if self.target_only_score and len(labels) > 1: # remove "strand" and "name" columns for straightforward ML
labels = labels[1]

return {
"inputs": np.array(seq),
"targets": labels,
Expand Down Expand Up @@ -299,7 +305,8 @@ class SeqIntervalDl(Dataset):
doc: 'defines the numpy dtype of the returned array. Example: int, np.int32, np.float32, float'
ignore_targets:
doc: if True, don't return any target variables

target_only_score:
doc: if True, only 'score' column of BED file is returned in targets array
output_schema:
inputs:
name: seq
Expand Down Expand Up @@ -332,12 +339,14 @@ def __init__(self,
dummy_axis=None,
alphabet="ACGT",
ignore_targets=False,
target_only_score=False,
dtype=None):
# core dataset, not using the one-hot encoding params
self.seq_dl = StringSeqIntervalDl(intervals_file, fasta_file, num_chr_fasta=num_chr_fasta,
label_dtype=label_dtype, auto_resize_len=auto_resize_len,
# use_strand=use_strand,
ignore_targets=ignore_targets)
ignore_targets=ignore_targets,
target_only_score=target_only_score)

self.input_transform = ReorderedOneHot(alphabet=alphabet,
dtype=dtype,
Expand Down