def test_ambiguous_mask2(tmpdir):
# only ambigous regions are present
bed_file = write_tmp(
'chr1\t1\t2\t1\t0\nchr2\t1\t3\t0\t1\nchr3\t1\t3\t-1\t-1', tmpdir)
bt = BedDataset(bed_file, ambiguous_mask=-1)
assert len(bt) == 2
assert np.all(bt.get_targets().max(axis=1) >= 0)
def test_ambiguous_mask(tmpdir):
bed_file = write_tmp(
'chr1\t1\t2\t1\t0\nchr2\t1\t3\t0\t1\nchr3\t1\t3\t0\t-1', tmpdir)
bt = BedDataset(bed_file)
assert len(bt) == 3
assert np.all(bt[2][1] == np.array([0, -1]))
# same as before
bt = BedDataset(bed_file, ambiguous_mask=-1)
assert len(bt) == 3
assert np.all(bt[2][1] == np.array([0, -1]))
assert np.all(bt.get_targets().max(axis=1) >= 0)
def test_bed3_labels(tmpdir):
bed_file = write_tmp('chr1\t1\t2\t1\t0\nchr1\t1\t3\t0\t1', tmpdir)
bt = BedDataset(bed_file)
assert np.all(bt.get_targets() == np.array([[1, 0], [0, 1]]))
assert len(bt) == 2
assert bt.n_tasks == 2
assert np.all(bt.df[0] == 'chr1')
assert bt[0][0] == Interval("chr1", 1, 2)
assert np.all(bt[0][1] == np.array([1, 0]))
assert bt[1][0] == Interval("chr1", 1, 3)
assert np.all(bt[1][1] == np.array([0, 1]))
assert len(bt) == 2
def test_label_dtype(tmpdir):
bed_file = write_tmp('chr1\t1\t2\t1\t0\nchr2\t1\t3\t0\t1', tmpdir)
bt = BedDataset(bed_file, label_dtype=bool)
assert len(bt) == 2
assert bt[0][1].dtype == bool
assert bt.get_targets().dtype == bool
class ActivityDataset(Dataset):
"""
Args:
intervals_file: bed4 file containing chrom start end name
fasta_file: file path; Genome sequence
label_dtype: label data type
num_chr_fasta: if True, the tsv-loader will make sure that the chromosomes
don't start with chr
"""
def __init__(self,
intervals_file,
fasta_file,
bigwigs,
track_width=2000,
incl_chromosomes=None,
excl_chromosomes=None,
num_chr_fasta=False):
self.num_chr_fasta = num_chr_fasta
self.intervals_file = intervals_file
self.fasta_file = fasta_file
self.bigwigs = bigwigs
self.incl_chromosomes = incl_chromosomes
self.excl_chromosomes = excl_chromosomes
self.track_width = track_width
self.tsv = BedDataset(
self.intervals_file,
num_chr=self.num_chr_fasta,
bed_columns=4,
ignore_targets=True,
incl_chromosomes=incl_chromosomes,
excl_chromosomes=excl_chromosomes,
)
self.fasta_extractor = None
self.bigwig_extractors = None
def __len__(self):
return len(self.tsv)
def __getitem__(self, idx):
if self.fasta_extractor is None:
self.fasta_extractor = FastaExtractor(self.fasta_file)
self.bigwig_extractors = {
a: [BigwigExtractor(f) for f in self.bigwigs[a]]
for a in self.bigwigs
}
interval, labels = self.tsv[idx]
interval = resize_interval(interval, 1000)
# Intervals need to be 1000bp wide
assert interval.stop - interval.start == 1000
# Run the fasta extractor
seq = np.squeeze(self.fasta_extractor([interval]))
interval_wide = resize_interval(deepcopy(interval), self.track_width)
return {
"inputs": {
"seq": seq
},
"targets": {
a:
sum([e([interval_wide])[0]
for e in self.bigwig_extractors[a]]).sum()
for a in self.bigwig_extractors
},
"metadata": {
"ranges":
GenomicRanges(interval.chrom, interval.start, interval.stop,
str(idx)),
"ranges_wide":
GenomicRanges.from_interval(interval_wide),
"name":
interval.name
}
}
def get_targets(self):
return self.tsv.get_targets()