def parse_position(string: str):
""" Converts a positiong from a string into a Position subclass """
if string[0] == '<':
return BeforePosition(int(string[1:]))
if string[0] == '>':
return AfterPosition(int(string[1:]))
if string == "UnknownPosition()":
return UnknownPosition()
return ExactPosition(int(string))
def test_unknown_position(self):
location = FeatureLocation(ExactPosition(1),
UnknownPosition(),
strand=1)
new_location = self.convert(location)
assert isinstance(new_location.start, ExactPosition)
assert new_location.start == 1
assert isinstance(new_location.end, UnknownPosition)
def test_start_before_end(self):
expected = "must be greater than or equal to start location"
with self.assertRaises(ValueError) as err:
FeatureLocation(42, 23, 1)
self.assertIn(expected, str(err.exception))
with self.assertRaises(ValueError) as err:
FeatureLocation(42, 0, 1)
self.assertIn(expected, str(err.exception))
with self.assertRaises(ValueError) as err:
FeatureLocation(BeforePosition(42), AfterPosition(23), -1)
self.assertIn(expected, str(err.exception))
with self.assertRaises(ValueError) as err:
FeatureLocation(42, AfterPosition(0), 1)
self.assertIn(expected, str(err.exception))
# Features with UnknownPositions should pass check
FeatureLocation(42, UnknownPosition())
FeatureLocation(UnknownPosition(), 42)
# Same start and end should pass check
FeatureLocation(42, 42)
def _read_ft(record, line):
name = line[5:13].rstrip()
if name:
if line[13:21] == " ": # new-style FT line
location = line[21:80].rstrip()
try:
isoform_id, location = location.split(":")
except ValueError:
isoform_id = None
try:
from_res, to_res = location.split("..")
except ValueError:
from_res = location
to_res = ""
qualifiers = {}
else: # old-style FT line
from_res = line[14:20].lstrip()
to_res = line[21:27].lstrip()
isoform_id = None
description = line[34:75].rstrip()
qualifiers = {"description": description}
if from_res == "?":
from_res = UnknownPosition()
elif from_res.startswith("?"):
position = int(from_res[1:]) - 1 # Python zero-based counting
from_res = UncertainPosition(position)
elif from_res.startswith("<"):
position = int(from_res[1:]) - 1 # Python zero-based counting
from_res = BeforePosition(position)
else:
position = int(from_res) - 1 # Python zero-based counting
from_res = ExactPosition(position)
if to_res == "":
position = from_res + 1
to_res = ExactPosition(position)
elif to_res == "?":
to_res = UnknownPosition()
elif to_res.startswith("?"):
position = int(to_res[1:])
to_res = UncertainPosition(position)
elif to_res.startswith(">"):
position = int(to_res[1:])
to_res = AfterPosition(position)
else:
position = int(to_res)
to_res = ExactPosition(position)
location = FeatureLocation(from_res, to_res, ref=isoform_id)
feature = FeatureTable(
location=location, type=name, id=None, qualifiers=qualifiers
)
record.features.append(feature)
return
# this line is a continuation of the previous feature
feature = record.features[-1]
if line[5:34] == " ": # old-style FT line
description = line[34:75].rstrip()
if description.startswith("/FTId="):
# store the FTId as the feature ID
feature.id = description[6:].rstrip(".")
return
# this line is a continuation of the description of the previous feature
old_description = feature.qualifiers["description"]
if old_description.endswith("-"):
description = "%s%s" % (old_description, description)
else:
description = "%s %s" % (old_description, description)
if feature.type in ("VARSPLIC", "VAR_SEQ"): # special case
# Remove unwanted spaces in sequences.
# During line carryover, the sequences in VARSPLIC/VAR_SEQ can get
# mangled with unwanted spaces like:
# 'DISSTKLQALPSHGLESIQT -> PCRATGWSPFRRSSPC LPTH'
# We want to check for this case and correct it as it happens.
try:
first_seq, second_seq = description.split(" -> ")
except ValueError:
pass
else:
extra_info = ""
# we might have more information at the end of the
# second sequence, which should be in parenthesis
extra_info_pos = second_seq.find(" (")
if extra_info_pos != -1:
extra_info = second_seq[extra_info_pos:]
second_seq = second_seq[:extra_info_pos]
# now clean spaces out of the first and second string
first_seq = first_seq.replace(" ", "")
second_seq = second_seq.replace(" ", "")
# reassemble the description
description = first_seq + " -> " + second_seq + extra_info
feature.qualifiers["description"] = description
else: # new-style FT line
value = line[21:].rstrip()
if value.startswith("/id="):
qualifier_type = "id"
value = value[4:]
assert value.startswith('"')
assert value.endswith('"')
feature.id = value[1:-1]
return
elif value.startswith("/evidence="):
value = value[10:]
assert value.startswith('"')
if value.endswith('"'):
value = value[1:-1]
else: # continues on the next line
value = value[1:]
assert "evidence" not in feature.qualifiers
feature.qualifiers["evidence"] = value
return
elif value.startswith("/note="):
value = value[6:]
assert value.startswith('"')
if value.endswith('"'):
value = value[1:-1]
else: # continues on the next line
value = value[1:]
assert "note" not in feature.qualifiers
feature.qualifiers["note"] = value
return
# this line is a continuation of the description of the previous feature
keys = list(feature.qualifiers.keys())
key = keys[-1]
description = value.rstrip('"')
old_description = feature.qualifiers[key]
if key == "evidence" or old_description.endswith("-"):
description = "%s%s" % (old_description, description)
else:
description = "%s %s" % (old_description, description)
if feature.type == "VAR_SEQ": # see VARSPLIC above
try:
first_seq, second_seq = description.split(" -> ")
except ValueError:
pass
else:
extra_info = ""
# we might have more information at the end of the
# second sequence, which should be in parenthesis
extra_info_pos = second_seq.find(" (")
if extra_info_pos != -1:
extra_info = second_seq[extra_info_pos:]
second_seq = second_seq[:extra_info_pos]
# now clean spaces out of the first and second string
first_seq = first_seq.replace(" ", "")
second_seq = second_seq.replace(" ", "")
# reassemble the description
description = first_seq + " -> " + second_seq + extra_info
feature.qualifiers[key] = description