def test_sequences_without_commonality_have_no_alignment(s1, s2):
u1 = s1.difference(s2)
u2 = s1.intersection(s2)
assert u1.intersection(u2) == set()
_, als = align(list(u1), list(u2), simple_score, simple_gap)
assert (len(list(als))) == 0
def test_imperfect_alignment_is_found(chunk):
separator = max(chunk) + 1
a = list(chunk) * 2
b = list(chunk) + [separator] + list(chunk)
_, alignments = align(a, b, simple_score, simple_gap)
als = list(alignments)
# We should only find one alignment
assert len(als) == 1
# The alignment for both strings should start at 0
assert als[0][0] == (0, 0)
def update(anchor, handle=None):
"""Update an anchor based on the current contents of its source file.
Args:
anchor: The `Anchor` to be updated.
handle: File-like object containing contents of the anchor's file. If
`None`, then this function will open the file and read it.
Returns: A new `Anchor`, possibly identical to the input.
Raises:
ValueError: No alignments could be found between old anchor and new
text.
AlignmentError: If no anchor could be created. The message of the
exception will say what the problem is.
"""
if handle is None:
with anchor.file_path.open(mode='rt') as fp:
source_text = fp.read()
else:
source_text = handle.read()
handle.seek(0)
ctxt = anchor.context
a_score, alignments = align(ctxt.full_text, source_text, score,
gap_penalty)
# max_score = len(ctxt.full_text) * 3
try:
alignment = next(alignments)
except StopIteration:
raise AlignmentError('No alignments for anchor: {}'.format(anchor))
anchor_offset = ctxt.offset - len(ctxt.before)
source_indices = tuple(s_idx for (a_idx, s_idx) in alignment
if a_idx is not None if s_idx is not None
if _index_in_topic(a_idx + anchor_offset, anchor))
if not source_indices:
raise AlignmentError(
"Best alignment does not map topic to updated source.")
return make_anchor(file_path=anchor.file_path,
offset=source_indices[0],
width=len(source_indices),
context_width=anchor.context.width,
metadata=anchor.metadata,
handle=handle)
def test_alignment_finds_perfect_subset(prefix, match, suffix):
larger = prefix + match + suffix
_, alignments = align(match, larger, simple_score, simple_gap)
als = [
al for al in alignments
# only perfect alignments
if len(al) == len(match)
# only alignments that start after `prefix`
if al[0][1] == len(prefix)
]
assert len(als) == 1
actual = als[0]
expected = tuple(
zip(range(len(match)), range(len(prefix),
len(prefix) + len(match))))
assert actual == expected
def test_multiple_alignments(match, size):
match = list(match)
larger = match * size
_, als = align(match, larger, simple_score, simple_gap)
assert len(list(als)) == size
def test_alignments_are_no_longer_than_longest_input(s1, s2):
max_input_len = max(len(s1), len(s2))
_, als = align(s1, s2, simple_score, simple_gap)
for al in als:
assert len(al) <= max_input_len
def test_sequences_with_commonality_have_at_least_one_alignment(seqs):
_, als = align(seqs[0], seqs[1], simple_score, simple_gap)
assert len(list(als)) >= 1
def test_empty_sequences_have_no_alignment(seq):
_, als = align(seq, [], simple_score, simple_gap)
assert len(list(als)) == 0
_, als = align([], seq, simple_score, simple_gap)
assert len(list(als)) == 0