Closed brendano closed 9 years ago
processing afp_eng_200206 (gigaword 5) with 17539 input documents, using ner + shift-reduce parsing
first try failed after 5099 lines output next try, after adding timeouts for the socket connect, failed with 14808 lines output, try again got 14810 lines output
Traceback (most recent call last):
File "proc.py", line 26, in <module>
head_jdoc = ss.parse_doc(headline_text)
File "/home/brenocon/gigaword/new_proc_of_rawgw/stanford_corenlp_pywrapper/sockwrap.py", line 143, in parse_doc
return self.send_command_and_parse_result(cmd, timeout, raw=raw)
File "/home/brenocon/gigaword/new_proc_of_rawgw/stanford_corenlp_pywrapper/sockwrap.py", line 162, in send_command_and_parse_result
data = self.send_command_and_get_string_result(cmd, timeout)
File "/home/brenocon/gigaword/new_proc_of_rawgw/stanford_corenlp_pywrapper/sockwrap.py", line 183, in send_command_and_get_string_result
sock = self.get_socket()
File "/home/brenocon/gigaword/new_proc_of_rawgw/stanford_corenlp_pywrapper/sockwrap.py", line 157, in get_socket
assert False, "couldnt connect socket"
all lines went through with the prev commit
socket timeout doesn't seem to get caught? i think this is happening in a place it's not supposed to -- when it's connecting the new socket. this isn't related to the 5-minute timeout on parsing a document.