Constructor DocumentSourceItem

Namespace

Qavren.Edge.Ingestion

Assembly

Qavren.Edge.Ingestion.dll

DocumentSourceItem(string, string, Func<CancellationToken, ValueTask<Stream>>, string?, long?, DateTimeOffset?, IReadOnlyDictionary<string, string>?)

One item a source yields. SP3 never interprets Path.

public DocumentSourceItem(string DocumentId, string MediaType, Func<CancellationToken, ValueTask<Stream>> OpenAsync, string? Path = null, long? SizeBytes = null, DateTimeOffset? LastModifiedUtc = null, IReadOnlyDictionary<string, string>? Metadata = null)

Parameters

DocumentId string

Stable within the source. The state row's key, with the source id.

MediaType string

Drives registry resolution before the extension does.

OpenAsync Func<CancellationToken, ValueTask<Stream>>

Called TWICE per document (hash pass, then extraction pass) and so MUST be re-openable. SHOULD return a seekable stream at position 0; a non-seekable one is buffered once below NonSeekableBufferLimitBytes and raises 6053 above it (spec 7.1).

Path string

The originating path or handle, for diagnostics only.

SizeBytes long?

Declared size, when the source knows it for free.

LastModifiedUtc DateTimeOffset?

Declared mtime, when the source knows it for free.

Metadata IReadOnlyDictionary<string, string>

Extra per-document facts a consumer wants carried through.