AutoMovie API
    Preparing search index...

    Roll every run of one design revision into one bounded verdict.

    The report cannot be cleared by silence. A required domain nobody submitted a run for becomes a gap of its own; a run that read a superseded revision is counted as stale rather than as an answer; a metric that produced no value carries its reason up. Only when every required domain answered against this revision, every metric produced a value, and every judged metric met its target does the status read meets.

    Declaring at least one required domain is mandatory, which is what removes the last way to pass: a report over nothing would otherwise clear everything.

    Gaps come from every submitted run, whether or not its domain is required. Any misses metric makes the status misses; otherwise any gap makes it incomplete. Gaps are ordered by run and by metric within a run, then by missing required domain in domain-table order; the first maxGaps of that list are listed and the rest only counted. The digest is SHA-256 over one JSON array per line (header, each rollup, each listed gap, then the omitted count), so it covers the listed gaps and that count but not the text of the omitted ones.

    Throws on a blank revision, a gap bound that is not a positive safe integer, an empty or unknown required domain, a run that fails validation, and a run id reported twice.

    type IProps = {
        maxGaps?: number;
        required: readonly AutoMovieAnalysisDomain[];
        revision: string;
        runs: readonly IAutoMovieAnalysisRun[];
    }
    Index
    maxGaps?: number

    Positive safe-integer bound on listed gaps; defaults to AUTOMOVIE_ANALYSIS_REPORT_MAX_GAPS.

    required: readonly AutoMovieAnalysisDomain[]

    Domains the production requires an answer for; at least one, all known.

    revision: string

    Non-blank design revision the report is a verdict about.

    runs: readonly IAutoMovieAnalysisRun[]

    Runs to roll up; each is validated before it is counted.