Skip to content

Commit

Permalink
Fix #384
Browse files Browse the repository at this point in the history
  • Loading branch information
@lucventurini
lucventurini committed Mar 3, 2021
1 parent f3043d3 commit 4a05df5
Show file tree
Hide file tree
Showing 2 changed files with 190 additions and 66 deletions.
5 changes: 5 additions & 0 deletions Mikado/_transcripts/transcript_base.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -477,6 +477,11 @@ def __initialize_with_gf(self, transcript_row: (GffLine, GtfLine)):
raise ValueError(transcript_row)
self.feature = intern(transcript_row.feature)

def _set_expandable(self, val):
if not val in (True, False):
raise ValueError("Invalid value, it must be boolean")
self.__expandable = val

def __repr__(self):
return pprint.saferepr(self.__dict__)

Expand Down
251 changes: 185 additions & 66 deletions Mikado/subprograms/util/convert.py
View file
Original file line number Diff line number Diff line change
@@ -1,14 +1,184 @@
#!/usr/bin/env python3
import argparse
import sys
from ...exceptions import InvalidTranscript
from ...exceptions import InvalidTranscript, InvalidParsingFormat
from ...parsers import parser_factory
from ...parsers.GFF import GFF3
from ...parsers.GTF import GTF
from ...parsers.bed12 import BED12, Bed12Parser
from ...loci import Transcript, Gene


def launch(args):
def _convert_bam(parser, args, out_format):
current = None
mock_gene_counter = 0
_line_counter = 0
for line in parser:
if line.is_unmapped is True:
continue

mock_gene_counter += 1
gene = "gene_{mock_gene_counter}".format(**locals())
try:
transcript = Transcript(line)
except KeyboardInterrupt:
raise
except Exception as exc:
raise InvalidTranscript("Invalid BAM line at position {} for file {}. Error: {}".format(
_line_counter, parser.name, exc
))
transcript.parent = gene
print(Gene(transcript).format(out_format, transcriptomic=args.transcriptomic), file=args.out)
continue


def _convert_gtf(parser: GTF, args, out_format):
genes = dict()
current = None
for line in parser:
if line.header is True:
continue
elif hasattr(line, "feature") and "superlocus" in line.feature or line.feature in ("chromosome", "region"):
continue
if line.gene in genes:
if line.is_transcript is True and line.transcript in genes[line.gene]:
gene = genes[line.gene]
assert isinstance(gene, Gene)
assert line.strand == gene.strand
assert line.chrom == gene.chrom
gene.transcripts[line.transcript].start, gene.transcripts[line.transcript].end = line.start, line.end
gene.transcripts[line.transcript]._set_expandable(False)
gene.transcripts[line.transcript].attributes.update(line.attributes)
elif line.is_transcript is True:
genes[line.gene].add(Transcript(line))
elif line.is_gene:
genes[line.gene].start, genes[line.gene].end = line.start, line.end
genes[line.gene].attributes.update(line.attributes)
elif line.is_exon:
if line.transcript in genes[line.gene].transcripts:
genes[line.gene].add_exon(line)
else:
genes[line.gene].add(Transcript(line))
else:
if line.is_gene:
genes[line.gene] = Gene(line)
else:
genes[line.gene] = Gene(Transcript(line))

if current and current != line.gene and args.assume_sorted is True:
print(genes[current].format(out_format, transcriptomic=args.transcriptomic), file=args.out)
del genes[current]

current = line.gene

if args.assume_sorted is False:
for gid, gene in genes.items():
print(gene.format(out_format, transcriptomic=args.transcriptomic), file=args.out)
elif current is not None:
print(genes[current].format(out_format, transcriptomic=args.transcriptomic), file=args.out)


def _convert_bed12(parser: Bed12Parser, args, out_format):
mock_gene_counter = 0
for line in parser:
if line.header is True:
continue
gid = "gene_{mock_gene_counter}".format(**locals())
line.parent = gid
gene = Gene(Transcript(line))
print(gene.format(out_format, transcriptomic=args.transcriptomic), file=args.out)


def _convert_gff(parser: GFF3, args, out_format):
orphaned = dict()
tid2gene = dict()
genes = dict()

current = None
for line in parser:
if line.header is True:
continue
elif line.feature in ("region", "chrom", "superlocus", "protein"):
continue
# We need to have better checks here regarding sorted/unsorted inputs here
if line.is_gene is False and args.assume_sorted is True:
if line.is_transcript is False:
if current is None or line.parent[0] not in genes[current].transcripts:
raise InvalidParsingFormat("The provided GFF is not sorted.")
genes[current].add_exon(line)
else:
assert genes[current].id in line.parent[0]
genes[current].add(Transcript(line))
elif line.is_gene is True:
if args.assume_sorted is True and current is not None:
print(genes[current].format(out_format, transcriptomic=args.transcriptomic), file=args.out)
del genes[current]
genes[line.gene] = Gene(line)
current = line.gene
if args.assume_sorted is False:
orph = list(orphaned.keys())
for tid in orph:
if orphaned[tid].parent and orphaned[tid].parent[0] == current:
genes[current].add(orphaned[tid])
del orphaned[tid]
elif line.is_transcript is True and args.assume_sorted is False:
transcript = Transcript(line)
tid2gene[transcript.id] = transcript.parent[0]
if transcript.id in orphaned:
assert transcript.chrom == orphaned[transcript.id].chrom
assert transcript.strand == orphaned[transcript.id].strand
orphaned[transcript.id].start, orphaned[transcript.id].end = transcript.start, transcript.end
orphaned[transcript.id].attributes.update(transcript.attributes)
orphaned[transcript.id]._set_expandable(False)
orphaned[transcript.id].parent = transcript.gene
if orphaned[transcript.id].parent and orphaned[transcript.id].parent[0] in genes:
genes[orphaned[transcript.id].parent[0]].add(orphaned[transcript.id])
del orphaned[transcript.id]
elif transcript.parent[0] in genes:
genes[transcript.parent[0]].add(transcript)
else:
orphaned[transcript.id] = transcript
else:
assert args.assume_sorted is False
# Exons might have multiple parents ...
if any(_ in tid2gene for _ in line.parent):
gids = set([tid2gene[_] for _ in line.parent])
assert len(gids) == 1, line
gid = gids.pop()
if gid in genes:
genes[gid].add_exon(line)
for parent in line.parent:
if parent not in genes[gid]:
genes[gid].add(Transcript(line))
else:
for parent in line.parent:
if parent in orphaned:
orphaned[parent].add_exon(line)
else:
orphaned[parent] = Transcript(line)
else:
for parent in line.parent:
if parent in orphaned:
orphaned[parent].add_exon(line)
else:
orphaned[parent] = Transcript(line)

if current and args.assume_sorted is True:
print(genes[current].format(out_format, transcriptomic=args.transcriptomic), file=args.out)

if args.assume_sorted is False:
for tid in orphaned:
for parent in orphaned[tid].parent:
if parent in genes:
genes[parent].add(orphaned[tid])
else:
genes[parent] = Gene(orphaned[tid])

for gid, gene in genes.items():
print(gene.format(out_format, transcriptomic=args.transcriptomic), file=args.out)

from ...parsers import parser_factory
from ...parsers.bed12 import BED12
from ...loci import Transcript, Gene

def launch(args):

if args.gf == "-":
if args.in_format is None:
Expand Down Expand Up @@ -36,67 +206,15 @@ def launch(args):
if out_format == "gff3":
print("##gff-version\t3", file=args.out)

mock_gene_counter = 0
_line_counter = 0

done = 0
for line in parser:
_line_counter += 1
if line.header is True:
continue
if parser.__annot_type__ == "bam":
# BAM file. We need to do things a little bit differently
if line.is_unmapped is True:
continue

mock_gene_counter += 1
gene = "gene_{mock_gene_counter}".format(**locals())
try:
transcript = Transcript(line)
except KeyboardInterrupt:
raise
except Exception as exc:
raise InvalidTranscript("Invalid BAM line at position {} for file {}. Error: {}".format(
_line_counter, parser.name, exc
))
transcript.parent = gene
print(Gene(transcript).format(out_format, transcriptomic=args.transcriptomic), file=args.out)
done += 1
continue

elif isinstance(line, BED12) or (line.is_exon is False and line.gene is None):
mock_gene_counter += 1
gene = "gene_{mock_gene_counter}".format(**locals())
line.parent = gene

if current is None or ((line.is_gene or line.is_transcript) and line.gene != current.id):
if current:
print(current.format(out_format, transcriptomic=args.transcriptomic), file=args.out)
if hasattr(line, "feature") and "superlocus" in line.feature: # Hack for Mikado files
continue
elif line.is_gene is True:
current = Gene(line)
elif hasattr(line, "feature") and line.feature in ("chromosome", "region"):
continue
else:
if line.parent is None:
line.parent = "{}.gene".format(line.id) # Hack for BED12 files
current = Gene(Transcript(line))
elif parser.__annot_type__ == "gtf" and line.is_exon is True and (
current is None or current.id != line.gene):
if current:
print(current.format(out_format, transcriptomic=args.transcriptomic), file=args.out)
current = Gene(Transcript(line))
elif (parser.__annot_type__ == "gtf" and
line.is_exon is True and line.transcript not in current.transcripts):
current.add(Transcript(line))
elif line.is_transcript is True:
current.add(Transcript(line))
elif line.is_exon is True:
current.add_exon(line)

if current is not None:
print(current.format(out_format, transcriptomic=args.transcriptomic), file=args.out)
if parser.__annot_type__ == "bam":
_convert_bam(parser, args, out_format)
return
elif parser.__annot_type__ == "gtf":
_convert_gtf(parser, args, out_format)
elif parser.__annot_type__ == "bed12":
_convert_bed12(parser, args, out_format)
else:
_convert_gff(parser, args, out_format)

if not args.out == sys.stdout:
args.out.flush()
Expand All @@ -111,6 +229,7 @@ def convert_parser():

parser = argparse.ArgumentParser(
"Utility to covert across GTF, GFF3 and BED12.")
parser.add_argument("-as", "--assume-sorted", default=False, action="store_true")
parser.add_argument("-of", "--out-format", dest="out_format",
choices=["bed12", "gtf", "gff3"], default=None)
parser.add_argument("-if", "--in-format", dest="in_format",
Expand Down

0 comments on commit 4a05df5

Please sign in to comment.