Input Validation API¶

OpenQP validates parsed inputs before a calculation starts. The same diagnostic engine can be called from Python to build input editors, web tools, notebooks, or agent workflows that need fast feedback.

Validating a Config¶

Use check_input_values when you already have a typed OpenQP configuration.

from oqp.utils.input_checker import check_input_values

report = check_input_values(
    config,
    raise_error=False,
    emit=False,
)

if not report.ok:
    print(report.to_text())

check_input_values(
    config,
    *,
    raise_error=True,
    emit=True,
    input_dir=None,
) -> CheckReport

Argument	Meaning
`config`	Parsed OpenQP configuration dictionary.
`raise_error`	If `True`, validation errors raise `SystemExit(1)`.
`emit`	If `True`, diagnostics are printed as text.
`input_dir`	Directory used to resolve paths stored relative to the input file.

Parsing Then Validating¶

For raw sectioned dictionaries, use OQPConfigParser with OQP_CONFIG_SCHEMA, then pass the typed config into the checker.

from oqp.molecule.oqpdata import OQP_CONFIG_SCHEMA
from oqp.utils.input_parser import OQPConfigParser
from oqp.utils.input_checker import check_input_values

raw = {
    "input": {
        "system": "\nH 0 0 0\nH 0 0 0.74",
        "basis": "6-31g*",
        "method": "hf",
        "runtype": "energy",
    },
    "scf": {
        "type": "rhf",
    },
}

parser = OQPConfigParser(schema=OQP_CONFIG_SCHEMA, allow_no_value=True)
parser.load_dict(raw)
config = parser.validate()

report = check_input_values(config, raise_error=False, emit=False)
print(report.to_text())

The parser applies schema defaults before validation, so the report sees the same effective input that a real OpenQP run would see.

Diagnostic Objects¶

check_input_values returns a CheckReport.

Attribute or method	Meaning
`report.ok`	`True` when there are no `ERROR` diagnostics.
`report.errors`	List of error diagnostics.
`report.warnings`	List of warning diagnostics.
`report.diagnostics`	Full diagnostic list, including informational messages.
`report.to_text()`	Human-readable report.

Each diagnostic contains:

Field	Meaning
`severity`	Usually `ERROR`, `WARNING`, or an informational severity.
`path`	Keyword path such as `input.runtype` or `tdhf.type`.
`message`	Short description of the problem.
`value`	Current value, when useful.
`expected`	Expected value or allowed values, when useful.
`action`	Suggested fix.
`wiki`	Extra help text for common user-facing errors.

Front-End Pattern¶

For an input builder or web service, keep validation non-fatal and render each diagnostic next to the relevant field.

report = check_input_values(config, raise_error=False, emit=False)

errors_by_path = {}
for item in report.errors:
    errors_by_path.setdefault(item.path, []).append(item.action or item.message)

This pattern avoids launching a calculation when the input is inconsistent, and it keeps the suggested fix close to the keyword that needs attention.