def equals(alpha, beta):
if isinstance(alpha, str) and isinstance(beta, str):
if alpha == beta:
return True
elif isinstance(beta, list):
sym = translate(alpha)
if sym in beta:
return True
elif isinstance(alpha, list):
sym = translate(beta)
if sym in alpha:
return True
return False
def _translate(s):
r=[]
for c in s:
if isinstance(c,str):
r.append([translate(c)])
elif isinstance(c, list):
r.append(c)
return r
def match_score2(alpha, beta):
if alpha == '-' and beta != '-':
return gap_penalty
if alpha != '-' and beta == '-':
return gap_penalty
if isinstance(alpha, str) and isinstance(beta, str):
if alpha == beta:
return match_award
elif isinstance(beta, list):
sym = translate(alpha)
if sym in beta:
return match_award
elif isinstance(alpha, list):
sym = translate(beta)
if sym in alpha:
return match_award
return mismatch_penalty
def translate_all(inputs,
filter_empty=True,
sort=True,
trans_table=all_table_ascii_unicode):
p = [
translate(i, trans_table=trans_table) for i in inputs
if not filter_empty or len(i.strip()) > 0
]
if sort:
p = sorted(p, key=functools.cmp_to_key(pattern_comparator))
return p
def symbol_to_patter(orig, symbol):
m=[]
for i, s in enumerate(symbol):
if isinstance(s, str):
m.append(s)
elif isinstance(s, list):
if len(s) == 1:
# normally only happens if we compare two patterns with the same symbols
m.append(s)
elif len(s) != 2:
print("WARNIGN {}".format(s))
else:
a1, a2 = s[0], s[1]
if "-" in s:
opt=a2 if a1 =="-" else a1
if isinstance(opt,str):
opt= utils.translate(opt)
m.append((opt,0,1))
else:
if isinstance(a1, str) and isinstance(a2, str):
# update vcounts
t = [utils.translate(a1), utils.translate(a2)]
m.append(list(set(t)))
if isinstance(a1, list) and isinstance(a2, str):
# we keep the a2 value in the freq dict,
# but need to merge the symbol
t = [utils.translate(a2)]
m.append(list(set(a1 + t)))
if isinstance(a1, str) and isinstance(a2, list):
# we keep the a2 value in the freq dict,
# but need to merge the symbol
t = [utils.translate(a1)]
m.append(list(set(a2 + t)))
print("O:{} S:{}".format(orig, symbol))
return m
def update_symbol(self, symbol):
#TODO length check
m = []
if not self._vcounts:
self._vcounts = [{} for i in range(0, len(symbol))]
for i, s in enumerate(symbol):
if isinstance(s, str):
m.append(s)
elif isinstance(s, list):
if len(s) == 1:
#normally only happens if we compare two patterns with the same symbols
m.append(s)
print("SINGLE VALUE {}".format(s))
elif len(s) != 2:
print("WARNIGN {}".format(s))
else:
a1 = s[0]
#
a2 = s[1]
if isinstance(a1, str) and isinstance(a2, str):
#
self._vcounts[i].setdefault(a1, 0)
self._vcounts[i].setdefault(a2, 0)
self._vcounts[i][a1] += 1
self._vcounts[i][a2] += 1
t = [translate(a1), translate(a2)]
m.append(list(set(t)))
if isinstance(a1, list) and isinstance(a2, str):
#we keep the a2 value in the freq dict,
# but need to merge the symbol
t = [translate(a2)]
m.append(list(set(a1 + t)))
self.symbol = m
self._count += 1
def match_score2(alpha, beta, score_matrix=score_matrix):
if alpha == '-' and beta != '-':
return score_matrix[gap_penalty]
if alpha != '-' and beta == '-':
return score_matrix[gap_penalty]
if isinstance(alpha, str) and isinstance(beta, str):
if alpha == beta:
return score_matrix[match_award]
elif isinstance(alpha, list) and isinstance(beta, list):
if set(beta) < set(alpha):
return score_matrix[match_award]
elif isinstance(beta, list):
sym = translate(alpha)
if sym in beta:
return score_matrix[cset_match_award]
elif isinstance(alpha, list):
sym = translate(beta)
if sym in alpha:
return score_matrix[cset_match_award]
return score_matrix[mismatch_penalty]
def merge_alignment(self, symbol):
m = []
for s in symbol:
if isinstance(s, str):
m.append(s)
elif isinstance(s, list):
a1 = s[0]
a2 = s[1]
# if isinstance(a1,list) and isinstance(a2, str):
##a1 is already a merge or optional
t = set(translate("".join(s)))
m.append([c for c in t])
return m
def from_string(cls,pstring):
_p= Pattern()
ins=''
parseValue=True
valuestring=''
for c in pstring:
if c == "'":
if not parseValue:
#start
parseValue=True
else:
#end
_p._values.append(valuestring)
_p._l1patterns.append(translate(valuestring))
parseValue=False
else:
if parseValue:
valuestring+=c
else:
pass
def parse_value(self, value):
logger.debug("CREATE PATTERN")
if not value:
return
t = translate(value)
logger.debug("{:>2} translated:{}".format('',t))
#o=unique_order(t)
#logger.debug("O:{}".format(o))
l1_grouped = [(k, sum(1 for i in g)) for k, g in itertools.groupby(t)]
logger.debug("{:>2} L1-groups:{}".format('',l1_grouped))
c=0
for g in l1_grouped:
self._l2patterns.append((g[0],c,g[1]))
self._l1patterns.append(t[c:c+g[1]])
self._values.append(value[c:c + g[1]])
c+=g[1]
self._l2string= "".join([p[0] for p in self._l2patterns])
self._l1string = "".join(self._l1patterns)
self._valuestring = "".join(self._values)
def find_best_alignment(self, alpha_list, beta_list):
identity, score, align1, symbol2, align2 = needle(
alpha_list, beta_list)
self.data['raw'] = {
'score': score,
'identity': identity,
'align1': align1,
'align2': align2,
'symbol': symbol2
}
if 0 < identity < 100:
ctrans = False
#translate the non matching symbols in alpha
alpha_ct = []
for i in range(0, len(align1)):
if len(symbol2[i]) == 1:
alpha_ct.append(align1[i])
else:
if symbol2[i][0] != '-':
if isinstance(symbol2[i][0], str):
ctrans = True
alpha_ct.append([translate(symbol2[i][0])])
else:
alpha_ct.append(symbol2[i][0])
if ctrans:
identity, score, align1, symbol2, align2 = needle(
alpha_ct, beta_list)
self.data['partl1'] = {
'score': score,
'identity': identity,
'align1': align1,
'align2': align2,
'symbol': symbol2
}
elif identity == 0:
#no matching characters:
identity, score, align1, symbol2, align2 = needle(
self._translate(alpha_list), beta_list)
self.data['l1'] = {
'score': score,
'identity': identity,
'align1': align1,
'align2': align2,
'symbol': symbol2
}
if len(self.data) > 1:
def compare(item1, item2):
res = item1[1]['identity'] - item2[1]['identity']
if res == 0:
res = item1[1]['score'] - item2[1]['score']
return res
_s_al = sorted(enumerate(list(self.data.values())),
key=functools.cmp_to_key(compare))
self.data['best'] = _s_al[-1][1]
else:
self.data['best'] = self.data['raw']