hankcs/HanLP · error · RuntimeError

eval.pl exited with error code {exitcode} and error message

Error message

eval.pl exited with error code {exitcode} and error message {err} and output {out}.

What it means

The dependency-parsing evaluator shells out to the external Perl script eval.pl (CONLLX_EVAL) to compute UAS/LAS. If the Perl process exits with a nonzero status, HanLP re-raises its stderr/stdout as a RuntimeError.

Source

Thrown at hanlp/metrics/parsing/conllx_eval.py:33

    Args:
      gold_file(str): The gold conllx file
      pred_file(str): The pred conllx file

    Returns:

    
    """
    gold_file = get_resource(gold_file)
    fixed_pred_file = tempfile.NamedTemporaryFile().name
    copy_cols(gold_file, pred_file, fixed_pred_file, keep_comments=False)
    if gold_file.endswith('.conllu'):
        fixed_gold_file = tempfile.NamedTemporaryFile().name
        copy_cols(gold_file, gold_file, fixed_gold_file, keep_comments=False)
        gold_file = fixed_gold_file

    exitcode, out, err = get_exitcode_stdout_stderr(f'perl {CONLLX_EVAL} -q -b -g {gold_file} -s {fixed_pred_file}')
    if exitcode:
        raise RuntimeError(f'eval.pl exited with error code {exitcode} and error message {err} and output {out}.')
    lines = out.split('\n')[-4:]
    las = int(lines[0].split()[3]) / int(lines[0].split()[5])
    uas = int(lines[1].split()[3]) / int(lines[1].split()[5])
    return uas, las


def copy_cols(gold_file, pred_file, copied_pred_file, keep_comments=True):
    """Copy the first 6 columns from gold file to pred file

    Args:
      gold_file: 
      pred_file: 
      copied_pred_file: 
      keep_comments:  (Default value = True)

    Returns:

View on GitHub (pinned to ddb1299bdd)

Solutions

Ensure perl is installed and on PATH (perl -v).
Inspect the reported stderr message — it usually states the real cause (e.g. 'No such file or directory' for missing eval.pl or an unreadable input file).
Validate that both gold and prediction files exist and are well-formed CoNLL-X (tab-separated, sentences separated by blank lines) before evaluating.
If you cannot install perl, use a pure-Python UAS/LAS implementation instead of this wrapper.

Defensive patterns

Strategy: try-catch

Validate before calling

import shutil
assert shutil.which('perl'), 'perl is required for conllx eval'

Try / catch

try:
    uas, las = evaluate(gold, pred)
except RuntimeError as e:
    log.error('eval.pl failed: %s', e)
    raise

Prevention

Pre-check shutil.which('perl') before evaluation.
Validate CoNLL files (column counts, blank-line separators) with a small parser before eval.

When it happens

Trigger: Calling conllx_eval.evaluate(gold_file, pred_file) when perl is missing/broken, eval.pl is not found at the expected path, or the gold/pred CoNLL-X files are malformed enough that eval.pl itself fails.

Common situations: Environments without perl installed (Windows, slim Docker images); running from a directory where the bundled eval.pl cannot be resolved; corrupted or empty prediction files produced by a failed parse step.

Related errors

AI-assisted analysis of hankcs/HanLP@ddb1299bdd (2026-08-27). Data as JSON: /api/errors/7781c56a119fda7c. Report an issue: GitHub.