HKUDS/Vibe-Trading · error · RegistryError

{path.name}: {size}B exceeds {_MAX_YAML_BYTES}B YAML cap

Error message

{path.name}: {size}B exceeds {_MAX_YAML_BYTES}B YAML cap

What it means

_safe_yaml_load parses YAML config files with a hard 5 MB cap as defence in depth against oversized (potentially malicious) YAML bombs. Files exceeding _MAX_YAML_BYTES raise RegistryError before yaml.safe_load runs.

Source

Thrown at agent/src/factors/registry.py:192

    except (ValueError, SyntaxError) as exc:
        raise RegistryError(f"{path.name}: __alpha_meta__ not a literal: {exc}") from exc

    if not isinstance(raw, dict):
        raise RegistryError(f"{path.name}: __alpha_meta__ must be dict, got {type(raw).__name__}")

    try:
        return AlphaMeta(**raw)
    except ValidationError as exc:
        raise RegistryError(f"{path.name}: AlphaMeta validation failed: {exc}") from exc


def _safe_yaml_load(path: Path) -> Any:
    """yaml.safe_load with hard 5 MB size cap (defence in depth)."""
    import yaml  # local import keeps registry import-light

    size = path.stat().st_size
    if size > _MAX_YAML_BYTES:
        raise RegistryError(f"{path.name}: {size}B exceeds {_MAX_YAML_BYTES}B YAML cap")
    text = path.read_text(encoding="utf-8")
    return yaml.safe_load(text)


def _zoo_dir_default() -> Path:
    return Path(__file__).parent / "zoo"


class Registry:
    """In-memory registry of all discoverable alphas across zoo subdirectories."""

    def __init__(self, zoo_root: Path | None = None) -> None:
        default_root = _zoo_dir_default()
        self._zoo_root = (zoo_root or default_root).resolve()
        # When the registry points at the bundled zoo dir, modules are loaded
        # by package name (warm import cache, supports relative imports). For
        # any other zoo_root (tests, plugins) we load by file path so the
        # caller doesn't have to fiddle with sys.path.

View on GitHub (pinned to 80ffdda44c)

Solutions

1. Split the YAML into multiple smaller files
2. Trim stale entries or move bulk data out of config
3. Regenerate the config programmatically to remove duplication

from factors.registry import _MAX_YAML_BYTES
assert path.stat().st_size <= _MAX_YAML_BYTES, f'{path} too large'

try:
    _safe_yaml_load(path)
except RegistryError as e:
    if 'YAML cap' in str(e): split_yaml(path)

When it happens

Trigger: A registry/theme YAML config larger than 5 MB, e.g. thousands of alphas in one file or YAML alias bombs.

Common situations: Machine-generated config files accumulating entries over time, merging many configs into one, or accidentally committing a data dump as YAML.

Related errors

• {path.name}: {size}B exceeds {_MAX_PY_BYTES}B cap
• YAML config is not available because PyYAML is missing
• Agent config must decode to a JSON/YAML object
• unknown panel column: {column}
• invalid {kind} {token!r}: must match {_ID_RE.pattern}

AI-assisted analysis of HKUDS/Vibe-Trading@80ffdda44c (2026-08-28). Data as JSON: /api/errors/5cc147907d69e3a7. Report an issue: GitHub.