sanskrit_parser vakya namonamaH --input HK Interpreting input loosely (strict_io set to false)
INFO Input String: namonamaH
INFO Input String in SLP1: namonamaH
Splits:
INFO Lexical Split: ['namaH', 'namaH']
Traceback (most recent call last):
File "c:\python37\lib\runpy.py", line 193, in _run_module_as_main
"__main__", mod_spec)
File "c:\python37\lib\runpy.py", line 85, in _run_code
exec(code, run_globals)
File "C:\Python37\Scripts\sanskrit_parser.exe\__main__.py", line 9, in <module>
File "c:\python37\lib\site-packages\sanskrit_parser\cmd_line.py", line 225, in cmd_line
eval(getattr(args, 'command')+"(rest)")
File "<string>", line 1, in <module>
File "c:\python37\lib\site-packages\sanskrit_parser\cmd_line.py", line 63, in vakya
for pi, parse in enumerate(split.parses()):
File "c:\python37\lib\site-packages\sanskrit_parser\api.py", line 161, in parses
max_parse_dc=self.parser.split_above)
File "c:\python37\lib\site-packages\sanskrit_parser\parser\datastructures.py", line 289, in __init__
self.add_edges()
File "c:\python37\lib\site-packages\sanskrit_parser\parser\datastructures.py", line 342, in add_edges
self.add_avyayas(bases)
File "c:\python37\lib\site-packages\sanskrit_parser\parser\datastructures.py", line 531, in add_avyayas
nextset = self.partitions[i+1]
IndexError: list index out of range