MAINT: improve error with SMILES

* The error message when processing conventional SMILES strings
can be confusing, as observed at:
marrink-lab/polyply_1.0#390 (comment)

* For cases like that where there are no polymeric fragments (i.e., no
use of braces), this patch improves the error message (and adds a test
for it) to nudge the user to BigSMILES or CGSmiles formats.

cgsmiles/resolve.py

@@ -430,6 +430,9 @@ def from_string(cls, cgsmiles_str, last_all_atom=True, legacy=False):
         """
         # here we figure out how many resolutions we are dealing with
         elements = re.findall(r"\{[^\}]+\}", cgsmiles_str)
+        if len(elements) == 0:
+            raise SyntaxError("No polymeric fragments detected; likely "
+                              "not a valid CGSmiles/BigSMILES string.")
         # the first one describes our lowest resolution
         molecule = read_cgsmiles(elements[0])
         # the rest are fragment lists

cgsmiles/tests/test_molecule_resolve.py

@@ -310,6 +310,8 @@ def _atomname_match(n1, n2):
 (("{[#A][#B]}.{#A=CC[$]}", "Found node #B but no corresponding fragment."),
  ("{[#A][#B]1}.{#A=CC[$],#B=OC[$]}", "You have a dangling ring index."),
  ("{[#A]1[#B]1}{#A=CC[$],#B=OC[$]}", "You define two edges between the same node. Use bond order symbols instead."),
+ # see: https://github.com/marrink-lab/polyply_1.0/issues/390#issuecomment-2455472121
+ ("CC(c1ccccc1)CC(c1ccccc1)CCCCCCCCCC(CC)CC(CC)CC(c1ccccc1)CC(c1ccccc1)", "No polymeric fragments detected; likely not a valid CGSmiles/BigSMILES string."),
 )))
 def test_syntax_errors(cgsmiles_str, error_message):
     with pytest.raises(SyntaxError) as e_message: