The CRFsuiteDumpParser parses a dump line per line with the assumption that each line is complete. Whenever an attribute contains a newline however, some lines will be broken up over multiple lines. Take for example the following broken up state feature line
(0) -1:word.lower:
--> O: 0.000003
This will break parse_STATE_FEATURES because the first line will not match the regex.
The
CRFsuiteDumpParserparses a dump line per line with the assumption that each line is complete. Whenever an attribute contains a newline however, some lines will be broken up over multiple lines. Take for example the following broken up state feature lineThis will break
parse_STATE_FEATURESbecause the first line will not match the regex.