OfficeIMO.Reader.Image 3.6.1

OfficeIMO.Reader.Image

OfficeIMO.Reader.Image adds standalone image files to an isolated OfficeDocumentReader. It uses the existing header-only OfficeIMO.Drawing identification API to expose format, dimensions, DPI, a materializable source asset, and optional OCR readiness without decoding pixels or running OCR.

using OfficeIMO.Reader;
using OfficeIMO.Reader.Image;

OfficeDocumentReader reader = new OfficeDocumentReaderBuilder()
    .AddImageHandler()
    .Build();

OfficeDocumentReadResult result = reader.ReadDocument("diagram.png");
Console.WriteLine(result.Markdown);
Console.WriteLine(result.Assets[0].MediaType);

PNG, JPEG, GIF, BMP, TIFF, SVG, EMF, WMF, ICO, PCX, and WebP extensions are registered. Identification is local and header-only. OCR execution remains an explicit host choice through OfficeIMO.Reader.Ocr and the shared OfficeIMO.Ocr.IOcrEngine contract.

The handler applies a 128 MiB default input ceiling when ReaderOptions.MaxInputBytes is not set. Set IncludePayload = false together with CreateOcrCandidate = false when only metadata is required.

Enable detailed inspection to project the shared Core metadata evidence into located Reader metadata:

OfficeDocumentReadResult inspected = reader.ReadDocument("photo.jpg", new ReaderOptions {
    Inspection = new ReaderInspectionOptions()
});
foreach (OfficeDocumentMetadataEntry property in inspected.Metadata) {
    Console.WriteLine($"{property.Category}: {property.Name} = {property.Value}");
}

The image.exif, image.exif.timestamp, image.exif.gps, image.xmp, image.icc and image.resolution categories preserve source counts/offsets, raw timestamp components, timezone uncertainty, RDF collection/language qualifiers and separate ICC validity/transform-support facts. Coverage and diagnostics use the existing Reader document contract. ReaderInspectionOptions applies operation-wide record and character limits; ReaderImageOptions.MetadataInspection controls Core's finite image/profile limits. Its default detailed-metadata input limit is 16 MiB, independently of the handler's general input ceiling. Detailed metadata does not execute OCR or decode image pixels, and explicitly reports unsupported BigTIFF, private MakerNote and additional-IFD evidence.

ReaderOptions.Projection = ReaderProjection.TextAndMetadata returns image identification as plain text chunks and metadata without constructing an asset, visual, Markdown or OCR candidate. Enable Inspection for the detailed metadata above. The default Full projection retains the asset and optional OCR-readiness behavior.