ValueError: invalid literal for int() with base 10: '!!'

[otariidae]

Minimal reproduce code

knp = KNP()
knp.parse("!!")

Traceback (most recent call last):
  File "hoge.py", line 99, in <module>
    knp.parse("!!")
  File "/root/.local/share/virtualenvs/app-4PlAip0Q/lib/python3.8/site-packages/pyknp/knp/knp.py", line 70, in parse
    return self.parse_juman_result(juman_str, juman_format)
  File "/root/.local/share/virtualenvs/app-4PlAip0Q/lib/python3.8/site-packages/pyknp/knp/knp.py", line 97, in parse_juman_result
    return BList(knp_lines, self.pattern, juman_format)
  File "/root/.local/share/virtualenvs/app-4PlAip0Q/lib/python3.8/site-packages/pyknp/knp/blist.py", line 39, in __init__
    self.parse(spec)
  File "/root/.local/share/virtualenvs/app-4PlAip0Q/lib/python3.8/site-packages/pyknp/knp/blist.py", line 116, in parse
    synnodes = SynNodes(string)
  File "/root/.local/share/virtualenvs/app-4PlAip0Q/lib/python3.8/site-packages/pyknp/knp/syngraph.py", line 21, in __init__
    self.tagids = [int(n) for n in tagid.split(',')]
  File "/root/.local/share/virtualenvs/app-4PlAip0Q/lib/python3.8/site-packages/pyknp/knp/syngraph.py", line 21, in <listcomp>
    self.tagids = [int(n) for n in tagid.split(',')]
ValueError: invalid literal for int() with base 10: '!!'

KNP output

$ echo '!!' | jumanpp | knp -tab
# S-ID:1 KNP:5.0-165d699 DATE:2020/09/23 SCORE:-23.04048
* -1D <文頭><文末><体言><用言:判><体言止><レベル:C><区切:5-5><ID:（文末）><裸名詞><提題受:30><主節><状態述語><正規化代表表記:!!/!!><主辞代表表記:!!/!!>
+ -1D <文頭><文末><体言><用言:判><体言止><レベル:C><区切:5-5><ID:（文末）><裸名詞><提題受:30><主節><状態述語><判定詞句><名詞項候補><先行詞候補><正規化代表表記:!!/!!><主辞代表表記:!!/!!><用言代表表記:!!/!!><CF_NOT_FOUND><節-区切><節-主辞><時制:非過去><格解析結果:!!/!!:判0><標準用言代表表記:!!/!!>
!! !! !! 名詞 6 普通名詞 1 * 0 * 0 "品詞推定:名詞 疑似代表表記 代表表記:!!/!! 品詞変更:!!-!!-!!-15-1-0-0" <品詞推定:名詞><疑似代表表記><代表表記:!!/!!><正規化代表表記:!!/!!><品詞変更:!!-!!-!!-15-1-0-0-"品詞推定:名詞 疑似代表表記 代表表記:!!/!!"><品曖-その他><未知語><記英数カ><英記号><記号><名詞相当語><文頭><文末><表現文末><自立><内容語><タグ単位始><文節始><文節主辞><用言表記先頭><用言表記末尾><用言意味表記末尾>
EOS

JUMAN++ 1.02
KNP current HEAD of master ku-nlp/knp@165d699
pyknp current HEAD of master 6ba00ea
Python 3.8.5
OS Ubuntu 20.04

[polm]

I ran into a similar issue using Juman++ (latest release) and pyknp from pip.

Traceback (most recent call last):
  File "./benchmark-jumanpp.py", line 10, in <module>
    for word in tok.analysis(line.strip()).mrph_list():
  File "/mnt/pool/code/tokenizer-benchmark/env/lib/python3.8/site-packages/pyknp/juman/juman.py", line 89, in analysis
    return self.juman(input_str, juman_format)
  File "/mnt/pool/code/tokenizer-benchmark/env/lib/python3.8/site-packages/pyknp/juman/juman.py", line 76, in juman
    result = MList(self.juman_lines(input_str), juman_format)
  File "/mnt/pool/code/tokenizer-benchmark/env/lib/python3.8/site-packages/pyknp/juman/mlist.py", line 29, in __init__
    mrph = Morpheme(line, mid, juman_format)
  File "/mnt/pool/code/tokenizer-benchmark/env/lib/python3.8/site-packages/pyknp/juman/morpheme.py", line 79, in __init__
    self._parse_spec(spec.strip("\n"))
  File "/mnt/pool/code/tokenizer-benchmark/env/lib/python3.8/site-packages/pyknp/juman/morpheme.py", line 142, in _parse_spec
    self.hinsi_id = int(parts[4])
ValueError: invalid literal for int() with base 10: 'input'

[nobu-g]

This problem seems to be fixed now.
I tested in the following environments and confirmed that pyknp works well.

JUMAN++ 1.02 / 2.0.0-rc3
KNP current HEAD of master ku-nlp/knp@2ad4f6d / 4.2
pyknp current HEAD of master 38469c8 / latest version from pip (0.4.5)
Python 3.7.9
OS macOS Bug Sur (11.0.1) / Ubuntu 20.04.1