Minor bug fixes and some regex preprocessing of input notations

glyles/glycans/mono/reactor.py

@@ -490,9 +490,9 @@ def set_fg(self, c_or_o, pos, bond_elem, name):
             return True
 
         # in case of a carbon chain
-        elif (len(name) >= 5 and name[1] == "C" and name[2:4].isnumeric()) or \
-                (len(name) >= 6 and name[1:3] in "aCiC" and name[3:5].isnumeric()) or \
-                (len(name) >= 7 and name[1:4] == "aiC" and name[4:6].isnumeric()):
+        elif len(name) > 1 and (name[1] == "C" and name[2:4].isnumeric()) or \
+                (name[1:3] in "aCiC" and name[3:5].isnumeric()) or \
+                (name[1:4] == "aiC" and name[4:6].isnumeric()):
             if bond_elem == "P":
                 bond_elem = "OP(=O)(O)"
             self.side_chains[pos][c_or_o] += bond_elem + self.parse_poly_carbon(name)

glyles/glycans/poly/glycan.py

@@ -1,5 +1,6 @@
 import logging
 from collections import Counter
+import re
 
 import networkx as nx
 from glyles.iupac.IUPACParser import IUPACParser
@@ -98,6 +99,10 @@ def recipe_equality(
     return False
 
 
+def prepare(iupac):
+    return re.sub(r'(?<!\d)dTal', '6dTal', iupac)
+
+
 class Glycan:
     """
     This class is like an interaction with the Parser for the IUPAC representation of the glycan. The grammar for
@@ -370,8 +375,9 @@ def parse(self):
             # parse the remaining structure description following the grammar, also add the dummy characters
             if not isinstance(self.iupac, str):
                 raise ParseError("Only string input can be parsed: " + str(self.iupac))
+            self.iupac = prepare('#' + self.iupac + '#')
 
-            stream = InputStream(data='#' + self.iupac + '#')
+            stream = InputStream(data=self.iupac)
             lexer = IUPACLexer(stream)
             token = CommonTokenStream(lexer)
             parser = IUPACParser(token)
@@ -411,4 +417,4 @@ def parse(self):
 
 
 if __name__ == "__main__":
-    print(Glycan("Fuc(a1-2)[GalNAc(a1-3)]Gal(b1-4)GlcNAc(b1-3)[Fuc(a1-2)[GalNAc(a1-3)]Gal(b1-4)GlcNAc(b1-6)]Gal(b1-3)[GlcNAc(a1-4)Gal(b1-4)GlcNAc6S(b1-6)]GalNAc", tree_only=True).summary())
+    print(Glycan("dTal").get_smiles())

glyles/glycans/poly/gwb_glycan.py

@@ -8,7 +8,7 @@
 from glyles.glycans.utils import ParseError
 from glyles.gwb.GWBLexer import GWBLexer
 from glyles.gwb.GWBParser import GWBParser
-from glyles.glycans.poly.glycan import Glycan
+from glyles.glycans.poly.glycan import Glycan, prepare
 
 
 def graph_to_string_int(graph: nx.DiGraph, node2label: callable) -> str:
@@ -78,6 +78,7 @@ def parse(self):
             if "$" in self.iupac:
                 self.iupac = self.iupac[:self.iupac.index("$")]
             self.iupac += "$"
+            prepare(self.iupac)
             stream = InputStream(data=self.iupac)
             lexer = GWBLexer(stream)
             token = CommonTokenStream(lexer)

tests/test_gwb.py

@@ -33,7 +33,7 @@ def test_gwb_iupac(data):
 @pytest.mark.fuzzy
 @pytest.mark.parametrize("gwb", fuzzy_data)
 def test_gwb_fuzzy(gwb):
-    if any(x in gwb for x in {"dTal", "End--??1P$", "End--??1S$", "[", "]"}):  # skip specific stuff and repeats
+    if any(x in gwb for x in {"[", "]"}):  # skip repeats
         pytest.skip()
     iupac = GWBGlycan(gwb).to_iupac()
     assert iupac is not None

tests/test_parser.py

@@ -78,12 +78,12 @@ def test_extreme_2(mode):
 def test_grammar_main(iupac, mode):
     assert Glycan(reduce_notation(iupac, mode), tree_only=True).get_tree() is not None
 
-def test_parse_1(self):
+def test_parse_1():
     g = Glycan("Man").get_tree()
 
     check_initial(g, "Man", 0, Config.UNDEF, lactole=Lactole.PYRANOSE)
 
-def test_parse_1_2(self):
+def test_parse_1_2():
     g = Glycan("Manpa").get_tree()
 
     check_initial(g, "Man", 0, Config.ALPHA, lactole=Lactole.PYRANOSE)
@@ -460,7 +460,7 @@ def test_parse_ternary_branching_2(mode):
     check_child(g, id_child_1, id_child_22, "Glc", "(a1-4)", 0, Lactole.PYRANOSE)
     check_child(g, id_child_1, id_child_23, "Gal", "(a1-6)", 0, Lactole.PYRANOSE)
 
-def test_parsing_error(self, caplog):
+def test_parsing_error(caplog):
     iupac = "Alt(a1-2)[Glc(a1-4)][Gal(a1-6)]Gul(a1-4)M*#$s'\\d ;«]as;an"  # Invalid IUPAC string!
     g = Glycan(iupac).get_tree()
     assert g is None