What causes "only support settings for CTB and PK now." in stanfordnlp/CoreNLP?

Turn off useASBCChar2/useHKChar2/useMSRChar2 and keep only useCTBChar2 or usePKChar2

stanfordnlp/CoreNLP · error · RuntimeException

only support settings for CTB and PK now.

Error message

only support settings for CTB and PK now.

What it means

TagAffixDetector picks character-list and affix dictionary paths for char-based POS features. When useChPos/useCTBChar2/usePKChar2 is on but one of the unsupported ASBC/HK/MSR char2 flags is also set, it throws RuntimeException because path lists exist only for CTB and PK.

Solutions

Turn off useASBCChar2/useHKChar2/useMSRChar2 and keep only useCTBChar2 or usePKChar2
If you need ASBC/HK/MSR support, supply your own character-list and in.* dictionary paths (requires code change)
Review the properties file so exactly one corpus char2 flag is enabled

Example fix

// before (Properties)
useCTBChar2=true
useMSRChar2=true
// after (Properties)
useCTBChar2=true
useMSRChar2=false

Defensive patterns

Strategy: validation

Validate before calling

Properties p = /* props */;
boolean char2on = Boolean.parseBoolean(p.getProperty("useChPos","false"))
  || Boolean.parseBoolean(p.getProperty("useCTBChar2","false"))
  || Boolean.parseBoolean(p.getProperty("usePKChar2","false"));
boolean unsupported = Boolean.parseBoolean(p.getProperty("useASBCChar2","false"))
  || Boolean.parseBoolean(p.getProperty("useHKChar2","false"))
  || Boolean.parseBoolean(p.getProperty("useMSRChar2","false"));
if (char2on && unsupported) {
  throw new IllegalArgumentException("TagAffixDetector supports only CTB/PK char2 settings");
}

Try / catch

try {
  new TagAffixDetector(flags);
} catch (RuntimeException e) {
  if (e.getMessage().contains("CTB and PK")) {
    flags.useMSRChar2 = false;
    new TagAffixDetector(flags);
  } else throw e;
}

Prevention

Enable exactly one corpus char2 flag (CTB or PK) per configuration
Never combine ASBC/HK/MSR char2 flags with useChPos
Validate props files with a small preflight script before launching training

When it happens

Trigger: Constructing TagAffixDetector with (useChPos || useCTBChar2 || usePKChar2) AND (useASBCChar2 || useHKChar2 || useMSRChar2) true.

Common situations: Enabling multiple corpus flags together by mistake (e.g. copying a properties file for MSR data while leaving CTB/PK flags on); trying to use tag-affix detection with ASBC/HK/MSR corpora.

Understand the failure class

Background: Conflicting config options: "cannot be used together" — configuration validation errors across open-source libraries — this error's family across 162 libraries.

Related errors

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/ead25221e0d878f0. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/wordseg/TagAffixDetector.java:35

  public TagAffixDetector(SeqClassifierFlags flags) {
    String corporaDict;
    if (flags.sighanCorporaDict != null) {
      corporaDict = flags.sighanCorporaDict;
    } else {
      corporaDict = DEFAULT_CORPORA_DICT;
    }

    if ( ! corporaDict.isEmpty() && ! corporaDict.endsWith("/")) {
      corporaDict = corporaDict + '/';
    }

    String ccPath;
    String adPath;
    if (flags.useChPos || flags.useCTBChar2 || flags.usePKChar2) {
      // if we're using POS information, override the ccPath
      // For now we only have list for CTB and PK
      if (flags.useASBCChar2 || flags.useHKChar2 || flags.useMSRChar2) {
        throw new RuntimeException("only support settings for CTB and PK now.");
      } else if (flags.useCTBChar2) {
        ccPath = corporaDict+"dict/character_list";
        adPath = corporaDict+"dict/in.ctb";
      } else if (flags.usePKChar2) {
        ccPath = corporaDict+"dict/pos_open/character_list.pku.utf8";
        adPath = corporaDict+"dict/in.pk";
      } else {
        throw new RuntimeException("none of flags.useXXXChar2 are on");
      }
    } else {
      ccPath = corporaDict+"dict/pos_close/char.ctb.list";
      adPath = corporaDict+"dict/in.ctb";
    }
    if (VERBOSE) {
      logger.info("TagAffixDetector: useChPos=" + flags.useChPos +
              " | useCTBChar2=" + flags.useCTBChar2 + " | usePKChar2=" + flags.usePKChar2);
      logger.info("TagAffixDetector: building TagAffixDetector from " + ccPath + " and " + adPath);
    }

View on GitHub (pinned to 1b7edd19c4)