DocSharpParsers 0.0.1

DocSharpParsers

A fork of DocSharp that adds a small, unified façade over its binary Office parsers.

Open legacy binary Microsoft Office files — .doc, .ppt, .xls — and:

• browse their internal OLE / compound-file structure,
• extract text and images,
• convert them to modern OOXML (.docx / .pptx / .xlsx).

Everything is lazy: opening only reads the compound-file directory, the format model is parsed on first text/convert access, and image bytes are decoded only when requested.

Quick start

using DocSharpParsers;

using var doc = OfficeDocument.Open("presentation.ppt");

Console.WriteLine(doc.Format);              // Word / PowerPoint / Excel

// 1. Browse the OLE compound-file structure (lazy).
foreach (var e in doc.Root.DescendantsAndSelf())
    Console.WriteLine($"{e.Kind,-7} {e.Path} ({e.Size} bytes)");

// Read a raw stream's bytes:
byte[] raw = doc.EnumerateStreams().First().ReadBytes();

// 2. Extract text.
string text = doc.GetText();

// 3. Extract images (bytes decoded on demand).
foreach (var img in doc.GetImages())
    File.WriteAllBytes($"img.{img.Extension}", img.GetBytes());

// 4. Convert to OOXML.
doc.ConvertToFile("presentation.pptx");
// or: doc.ConvertTo(someStream);

Opening from a stream

using var fs = File.OpenRead("report.xls");
using var doc = OfficeDocument.Open(fs, fs.Name);          // doc closes fs on dispose

// Keep ownership of the stream (it is NOT closed with the document):
using var doc2 = OfficeDocument.Open(fs, fs.Name, leaveOpen: true);

The stream must stay open and seekable for the lifetime of the OfficeDocument (its contents are read on demand). Pass the file name — or just its extension — to aid format detection when the content check is inconclusive.

Notes

• Format is auto-detected from the streams inside the OLE container (falling back to the file extension).
• Images are extracted for all three formats (Word, PowerPoint and Excel).
• OfficeImage.Format / Extension report the encoding (PNG, JPEG, GIF, TIFF, BMP, EMF, WMF, PICT); metafiles are decompressed and headerless DIB blips are wrapped into standalone BMPs.

License

MIT — see LICENSE. This project is a fork of DocSharp (© Manfredi Marceca), also MIT-licensed.