Circle V2 API Docs
    Preparing search index...
    type SplitNoaPacketOptions = {
        duplicatedPieceSteps?: string[];
        parseRunIdInUsageOnly?: boolean;
        parseRunIdOnSplitOnly?: boolean;
        retriedValue?: Record<string, unknown>;
        withFailedPieceSteps?: string[];
        withOrphanPieceStep?: boolean;
        withoutClassifyStep?: boolean;
        withoutDocumentStep?: boolean;
        withoutParentFileIds?: boolean;
        withoutPieceSteps?: string[];
        withUnreadableSplits?: boolean;
    }
    Index

    Properties

    duplicatedPieceSteps?: string[]

    Piece types whose extraction step appears twice, as after an Extend retry; the later copy carries retriedValue.

    parseRunIdInUsageOnly?: boolean

    No step names a parse run at all; usage.breakdown is the only place the id appears.

    parseRunIdOnSplitOnly?: boolean

    As production sends it: the parse run id rides on the SPLIT step alone, every other step's is null.

    retriedValue?: Record<string, unknown>
    withFailedPieceSteps?: string[]

    Piece types whose extraction step is present but produced no result, as when the sub-step failed.

    withOrphanPieceStep?: boolean

    A piece step whose file belongs to no split — the only way a piece can go unrecorded.

    withoutClassifyStep?: boolean
    withoutDocumentStep?: boolean
    withoutParentFileIds?: boolean

    No step file names its parent, so pieces can only be told apart from the document by the split list.

    withoutPieceSteps?: string[]

    Piece types whose extraction step is left out, as when Extend never ran that extractor.

    withUnreadableSplits?: boolean

    The SPLIT output with its page numbers as strings, which the parser cannot read.