Merge pull request #344 from bigict/data

fix: filter not supported chain type when creating dataset

Author: chungongyu

tools/dataset_from_pdb.py

@@ -24,6 +24,13 @@
 logger = logging.getLogger(__file__)
 
 
+CHAIN_TYPE_DICT = {
+    'mol:protein': residue_constants.PROT,
+    'mol:dna': residue_constants.DNA,
+    'mol:rna': residue_constants.RNA
+}
+
+
 def output_get_basename(filename):
   filename = os.path.basename(filename)
   if filename.endswith('.gz'):
@@ -175,6 +182,8 @@ def _make_domain(start, end, delta=0):
     chain_type = chain_type_dict.get(chain_id)
     if not exists(chain_type):  # FIX: 146d
       continue
+    if CHAIN_TYPE_DICT.get(chain_type) not in residue_constants.restype_list:
+      continue
 
     int_resseq = label_seq_id_list[i]
     residue_id = residue_id_list[i]