def test_summarize_labels():
figure = Figure()
figure.add(Track(1, 200, label="Foo"))
figure.add(Track(1, 200, label="Bar", color='none'))
expected_summary = """\
Foo[1-200]
Bar(1-200)
"""
summary = summarize_figure(figure)
assert summary == expected_summary
def test_summarize_label_objects():
figure = Figure()
figure.add(Track(0, 0, label=Label(25, "Foo:")))
figure.add(Track(0, 0, label="Bar:"))
expected_summary = """\
Foo:
Bar:
"""
summary = summarize_figure(figure)
assert summary == expected_summary
def test_add_track(self):
figure = Figure()
figure.add_track(
Track(50,
300,
direction='f',
label="Another\
sequence",
regions=[(50, 100, 'lightblue')]))
figure.add_track(Track(110, 410, direction='r', label="Sequence 1",\
regions=[(150, 200, 'salmon')]))
figure.show()
def test_multitrack(self):
figure = Figure()
for i in range(0, 10):
figure.add_track(
Multitrack([
Track(
i, i + 10, direction='f', label='Track {}F'.format(i)),
Track(i + 20,
i + 30,
direction='r',
label='Track {}R'.format(i))
],
join=True))
figure.show()
def test_summarize_multitracks():
figure = Figure()
figure.add(Track(0, 0, label="Foo:"))
figure.add(
Multitrack([Track(10, 20, label="Bar"),
Track(30, 40, label="Baz")]))
expected_summary = """\
Foo:
Bar[10-20], Baz[30-40]
"""
summary = summarize_figure(figure)
assert summary == expected_summary
def add_partial_banner(f, position_offset, max_position):
""" Build a dashed line with dashes 500 wide. """
dash_width = 500
banner_width = max_position - position_offset
subtracks = [
Track(i * dash_width + position_offset + 1,
min((i + 1) * dash_width + position_offset, max_position))
for i in range((banner_width + dash_width) // dash_width) if not i % 2
]
subtracks.append(
Track(position_offset + 1,
max_position,
label='Partial Blast Results',
color='none'))
f.add(Multitrack(subtracks))
def test_summarize_zero_coverage():
figure = Figure()
figure.add(Coverage(10, 20, [0, 0, 0]), gap=-4)
figure.add(Track(10, 20, label="Bar"))
with pytest.raises(ZeroDivisionError):
summarize_figure(figure)
def test_arrow_group(svg_differ):
expected_figure = Figure()
expected_figure.add(Track(1, 500, label='Header'))
h = 30
expected_figure.add(Arrow(1, 200, label='X', h=h), gap=-h)
expected_figure.add(Arrow(300, 500, label='Y', h=h))
expected_svg = expected_figure.show()
f = Figure()
f.add(Track(1, 500, label='Header'))
f.add(
ArrowGroup(
[Arrow(1, 200, label='X', h=h),
Arrow(300, 500, label='Y', h=h)]))
svg = f.show()
svg_differ.assert_equal(svg, expected_svg, 'test_arrow_group')
def test_summarize_multitracks_with_separate_label():
figure = Figure()
figure.add(Track(0, 0, label="Foo:"))
figure.add(
Multitrack([
Track(10, 20),
Track(30, 40),
Track(10, 40, label="Bar", color='none')
]))
expected_summary = """\
Foo:
[10-20], [30-40], Bar(10-40)
"""
summary = summarize_figure(figure)
assert summary == expected_summary
def test_draw_coverage(svg_differ):
expected_figure = Figure()
expected_figure.add(Track(0, 1, color='', h=-4)) # Just a spacer.
expected_figure.add(Track(100, 200, label='Bar'))
expected_svg = expected_figure.show()
expected_svg.insert(0, draw.Rectangle(100, 20, 25, 5, fill='blue'))
expected_svg.insert(1, draw.Rectangle(125, 20, 25, 10, fill='blue'))
expected_svg.insert(2, draw.Rectangle(175, 20, 25, 1, fill='blue'))
figure = Figure()
coverage_depths = 25 * [5] + 25 * [10] + 25 * [0] + 25 * [1]
figure.add(SmoothCoverage(100, 200, coverage_depths), gap=-4)
figure.add(Track(100, 200, label="Bar"))
svg = figure.show()
svg_differ.assert_equal(svg, expected_svg, 'test_draw_coverage')
def test_arrow_group_reverse_overlap(svg_differ):
expected_figure = Figure()
expected_figure.add(Track(1, 500, label='Header'))
h = 20
expected_figure.add(Arrow(1, 300, label='X', h=h), gap=3)
expected_figure.add(Arrow(400, 250, label='Y', h=h))
expected_svg = expected_figure.show()
f = Figure()
f.add(Track(1, 500, label='Header'))
f.add(
ArrowGroup(
[Arrow(1, 300, label='X', h=h),
Arrow(400, 250, label='Y', h=h)]))
svg = f.show()
svg_differ.assert_equal(svg, expected_svg, 'test_arrow_group')
def test_arrow_group_small_neighbour(svg_differ):
expected_figure = Figure()
expected_figure.add(Track(1, 500, label='Header'))
h = 20
expected_figure.add(Arrow(301, 315, elevation=-1, label='1.2', h=h),
gap=-h)
expected_figure.add(Arrow(1, 300, elevation=-1, label='1.1', h=h))
expected_svg = expected_figure.show()
f = Figure()
f.add(Track(1, 500, label='Header'))
f.add(
ArrowGroup([
Arrow(1, 300, elevation=-1, label='1.1', h=h),
Arrow(301, 315, elevation=-1, label='1.2', h=h)
]))
svg = f.show()
svg_differ.assert_equal(svg, expected_svg, 'test_arrow_group')
def test_summarize_smooth_coverage_ten_percent():
figure = Figure()
figure.add(SmoothCoverage(10, 20, [100, 110, 111, 50]), gap=-4)
figure.add(Track(12, 22, label="Bar"))
expected_summary = """\
Coverage 100x2, 111, 50
Bar[12-22]
"""
summary = summarize_figure(figure)
assert summary == expected_summary
def test_summarize_smooth_coverage():
figure = Figure()
figure.add(SmoothCoverage(10, 20, [11, 11, 21, 1, 1, 1]), gap=-4)
figure.add(Track(12, 22, label="Bar"))
expected_summary = """\
Coverage 11x2, 21, 1x3
Bar[12-22]
"""
summary = summarize_figure(figure)
assert summary == expected_summary
def test_summarize_regions():
figure = Figure()
figure.add(
Track(1,
200,
label="Foo",
regions=[(50, 100, 'lightgreen'), (110, 120, 'red')]))
expected_summary = """\
Foo[1-200], lightgreen{50-100}, red{110-120}
"""
summary = summarize_figure(figure)
assert summary == expected_summary
def start_drawing(width, height):
expected_svg = Drawing(width, height, origin=(0, 0))
expected_svg.append(
Rectangle(0,
height - 15,
200,
10,
stroke='lightgrey',
fill='lightgrey'))
expected_svg.append(
Text('Header',
10,
width / 2,
height - 15,
font_family='monospace',
text_anchor='middle'))
f = Figure()
f.add(Track(0, width, label='Header'))
return f, expected_svg
def build_contig(reader, f, contig_name, max_position, position_offset,
blast_rows):
contig_matcher = ContigMatcher(contig_name)
blast_ranges = [] # [[start, end, blast_num]]
blast_starts = defaultdict(set) # {start: {blast_num}}
blast_ends = defaultdict(set) # {end: {blast_num}}
if not contig_name.startswith('contig-'):
for blast_row in blast_rows:
if not contig_matcher.is_match(blast_row):
continue
blast_num = len(blast_ranges) + 1
blast_ranges.append([None, None, blast_num])
blast_starts[blast_row['start']].add(blast_num)
blast_ends[blast_row['end']].add(blast_num)
event_positions = set(blast_starts)
event_positions.update(blast_ends)
event_positions = sorted(event_positions, reverse=True)
insertion_size = 0
insertion_ranges = [] # [(start, end)]
unmatched_ranges = [] # [[start, end]]
for contig_name2, contig_rows in groupby(reader, itemgetter('contig')):
if contig_name2 != contig_name:
continue
contig_rows = list(contig_rows)
coordinates_name = contig_rows[0]['coordinates']
if coordinates_name:
pos_field = 'refseq_nuc_pos'
else:
pos_field = 'query_nuc_pos'
for contig_row in contig_rows:
for field_name in (pos_field, 'coverage', 'dels'):
field_text = contig_row[field_name]
field_value = None if field_text == '' else int(field_text)
contig_row[field_name] = field_value
start = contig_rows[0][pos_field]
end = contig_rows[-1][pos_field]
coverage = [0] * (end - start + 1)
pos = 0
for contig_row in contig_rows:
pos = contig_row[pos_field]
if pos is None:
insertion_size += 1
else:
if insertion_size:
insertion_ranges.append((pos, pos + insertion_size - 1))
insertion_size = 0
if contig_row['coverage'] is not None:
coverage[pos - start] = (contig_row['coverage'] -
contig_row['dels'])
contig_pos = int(contig_row['query_nuc_pos'])
while event_positions and event_positions[-1] <= contig_pos:
event_pos = event_positions.pop()
for blast_num in blast_starts[event_pos]:
blast_ranges[blast_num - 1][0] = pos
for blast_num in blast_ends[event_pos]:
blast_ranges[blast_num - 1][1] = pos
link = contig_row.get('link')
if link == 'U':
# Position is unmatched, add to list.
if not unmatched_ranges or unmatched_ranges[-1][-1] != pos - 1:
unmatched_ranges.append([pos, pos])
else:
unmatched_ranges[-1][-1] = pos
while event_positions:
# Use up any events that went past the end of the contig.
event_pos = event_positions.pop()
for blast_num in blast_starts[event_pos]:
blast_ranges[blast_num - 1][0] = pos
for blast_num in blast_ends[event_pos]:
blast_ranges[blast_num - 1][1] = pos
arrows = []
for arrow_start, arrow_end, blast_num in blast_ranges:
arrows.append(
Arrow(arrow_start + position_offset,
arrow_end + position_offset,
elevation=-1,
label=f'{contig_matcher.num}.{blast_num}'))
if arrows:
f.add(ArrowGroup(arrows))
subtracks = []
for has_coverage, group_positions in groupby(
enumerate(coverage), lambda item: item[1] != 0):
if has_coverage:
group_positions = list(group_positions)
group_start, _ = group_positions[0]
group_end, _ = group_positions[-1]
subtracks.append(
Track(start + group_start + position_offset,
start + group_end + position_offset))
if not subtracks:
group_start = prev_pos = None
included_positions = [row[pos_field] for row in contig_rows]
included_positions.append(None) # Trigger final section.
for pos in included_positions:
if group_start is None:
group_start = pos
else:
if pos != prev_pos + 1:
subtracks.append(
Track(group_start + position_offset,
prev_pos + position_offset))
group_start = pos
prev_pos = pos
if max(coverage) <= 0:
track_label = contig_name
else:
f.add(ShadedCoverage(start + position_offset,
end + position_offset, coverage),
gap=-4)
track_label = f"{contig_name} - depth {max(coverage)}"
subtracks.append(
Track(1,
max_position,
label=track_label,
color='none',
regions=[
(a + position_offset, b + position_offset, 'lightgreen')
for a, b in insertion_ranges
] + [(a + position_offset, b + position_offset, 'yellow')
for a, b in unmatched_ranges]))
f.add(Multitrack(subtracks))
break
def build_coverage_figure(genome_coverage_csv, blast_csv=None):
min_position, max_position = 1, 500
coordinate_depths = Counter()
contig_depths = Counter()
contig_groups = defaultdict(set) # {coordinates_name: {contig_name}}
reader = DictReader(genome_coverage_csv)
for row in reader:
query_nuc_pos = int(row['query_nuc_pos'])
if row['refseq_nuc_pos']:
refseq_nuc_pos = int(row['refseq_nuc_pos'])
else:
refseq_nuc_pos = min_position
min_position = min(min_position, refseq_nuc_pos, query_nuc_pos)
max_position = max(max_position, refseq_nuc_pos, query_nuc_pos)
coordinates_name = row['coordinates']
contig_name = row['contig']
if row['coverage'] != '':
row_coverage = int(row['coverage']) - int(row['dels'])
coordinate_depths[coordinates_name] = max(
coordinate_depths[coordinates_name], row_coverage)
contig_depths[contig_name] = max(contig_depths[contig_name],
row_coverage)
contig_groups[coordinates_name].add(contig_name)
if '' in coordinate_depths:
# Force partial contigs to come last.
coordinate_depths[''] = -1
position_offset = -min_position + 1
max_position += position_offset
blast_rows = []
if blast_csv is not None:
for blast_row in DictReader(blast_csv):
for field_name in ('start', 'end', 'ref_start', 'ref_end'):
# noinspection PyTypeChecker
blast_row[field_name] = int(blast_row[field_name])
blast_rows.append(blast_row)
blast_rows.sort(key=itemgetter('start', 'ref_start'))
landmarks_path = (Path(__file__).parent.parent / "data" /
"landmark_references.yaml")
landmark_groups = yaml.safe_load(landmarks_path.read_text())
projects = ProjectConfig.loadDefault()
f = Figure()
for _, coordinates_name in sorted(
(-depth, name) for name, depth in coordinate_depths.items()):
for reference_set in landmark_groups:
if coordinates_name != reference_set['coordinates']:
continue
prev_landmark = None
for i, landmark in enumerate(
sorted(reference_set['landmarks'],
key=itemgetter('start'))):
landmark.setdefault('frame', 0)
if prev_landmark and 'end' not in prev_landmark:
prev_landmark['end'] = landmark['start'] - 1
prev_landmark = landmark
for frame, frame_landmarks in groupby(reference_set['landmarks'],
itemgetter('frame')):
subtracks = []
for landmark in frame_landmarks:
landmark_colour = landmark.get('colour')
if landmark_colour is None:
continue
subtracks.append(
Track(landmark['start'] + position_offset,
landmark['end'] + position_offset,
label=landmark['name'],
color=landmark_colour))
max_position = max(max_position,
landmark['end'] + position_offset)
f.add(Multitrack(subtracks))
break
else:
add_partial_banner(f, position_offset, max_position)
contig_names = contig_groups[coordinates_name]
sorted_contig_names = sort_contig_names(contig_names, contig_depths)
ref_arrows = []
for contig_name in sorted_contig_names:
if contig_name.startswith('contig-'):
# No arrows on original contig tracks.
continue
contig_matcher = ContigMatcher(contig_name)
ref_positions = None
arrow_count = 0
for blast_row in blast_rows:
if not contig_matcher.is_match(blast_row):
continue
if (ref_positions is None and coordinates_name != ''
and blast_row['ref_name'] != coordinates_name):
ref_positions = map_references(blast_row['ref_name'],
coordinates_name, projects)
arrow_count += 1
ref_start = int(blast_row['ref_start'])
ref_end = int(blast_row['ref_end'])
if ref_positions is None:
coordinate_start = ref_start
coordinate_end = ref_end
else:
coordinate_start = ref_positions[ref_start]
coordinate_end = ref_positions[ref_end]
ref_arrows.append(
Arrow(coordinate_start + position_offset,
coordinate_end + position_offset,
elevation=1,
label=f'{contig_matcher.num}.{arrow_count}'))
if ref_arrows:
f.add(ArrowGroup(ref_arrows))
for contig_name in sorted_contig_names:
genome_coverage_csv.seek(0)
reader = DictReader(genome_coverage_csv)
build_contig(reader, f, contig_name, max_position, position_offset,
blast_rows)
if not f.elements:
f.add(Track(1, max_position, label='No contigs found.', color='none'))
return f