Member Junction
    Preparing search index...

    One block of extracted content. Several blocks (e.g. archive members) become several records.

    interface ContentBlock {
        Columns?: Record<string, unknown>;
        Content?: Uint8Array;
        ExtractorKeyOverride?: string;
        FileType?: string;
        Key?: string;
        Request?: PipelineStageRequest;
        Text: string;
        Title?: string;
        TitleConfidence?: number;
    }
    Index

    Properties

    Columns?: Record<string, unknown>

    Extra or custom columns on MJ: Content Items for the child record this block produces.

    Content?: Uint8Array

    Raw bytes of a non-text block (e.g. a PDF inside an archive), which gets its own extraction.

    ExtractorKeyOverride?: string

    Forces an extractor for this block, set by the extractor that produced it.

    FileType?: string

    The block's file type, when a splitting extractor knows it differs from its parent's.

    Key?: string

    The block's identity within its parent (e.g. an archive member path), appended to the parent's URL.

    Explicit request for downstream pipeline work on the child record this block produces.

    Text: string

    The block's text. Empty for a non-text block (see Content); never decode arbitrary bytes into it.

    Title?: string

    A title the extractor could determine from the block's own structure.

    TitleConfidence?: number

    How much the extractor trusts that title. Omitted means it has no opinion.