DocX

A small framework that converts NSAttributedString to .docx Word files on iOS and macOS.

Motivation

On iOS, NSAttributedString.DocumentType supports only HTML and Rich Text, while on macOS .doc and .docx are available options. Even then the .docx exporter on macOS supports only a subset of the attributes of NSAttributedString.

This library is used in SimpleFurigana for macOS and SimpleFurigana for iOS, hence the focus on furigana annotation export.

Installation

Add

.package(name: "DocX", url: "https://github.com/shinjukunian/DocX.git", .branch("master"))

to dependencies in your Package.swift file. This requires Swift 5.3, which shipped with Xcode12. Alternatively, add DocX in Xcode via File->Swift Packages->Add Package Dependency, paste https://github.com/shinjukunian/DocX.git as URL and specify master as branch.

Usage

let string = NSAttributedString(string: "This is a string", attributes: [.font: UIFont.systemFont(ofSize: UIFont.systemFontSize), .backgroundColor: UIColor.blue])
let url = URL(fileURLWithPath:"myPath")
try? string.writeDocX(to: url)

Starting from iOS 15 / macOS 12, you can use the new AttributedString.

var att=AttributedString("Lorem ipsum dolor sit amet")
att.font = NSFont(name: "Helvetica", size: 12)
att[att.range(of: "Lorem")!].backgroundColor = .blue
let url = URL(fileURLWithPath:"myPath")
try att.writeDocX(to: url)

Naturally, this works for Markdown as well:

let mD="~~This~~ is a **Markdown** *string*."
let att=try AttributedString(markdown: mD)
try att.writeDocX(to: url)

You can optionally specify MetaData using DocXOptions:

let font = NSFont(name: "Helvetica", size: 13)! //on macOS
let string = NSAttributedString(string: "The Foundation For Law and Government favours Helvetica.", attributes: [.font: font])

var options = DocXOptions()
options.author = "Michael Knight"
options.title = "Helvetica Document"

let url = URL(fileURLWithPath:"myPath")
try string.writeDocX(to: url, options:options)

See the attached sample projects (for iOS and macOS) for usage and limitations. On iOS, DocX also includes a UIActivityItemProvider subclass (DocXActivityItemProvider) for exporting .docx files through UIActivityViewController.

NSAttributedString has no concept of pagination. For manual pagination, use

try DocXWriter.write(pages:[NSAttributedString], to url:URL)

to render each NSAttributedString as a separate page.

A sample output on macOS opened in Word365.

A sample output on macOS with an embedded image (via NSTextAttachment). in the macOS sample application (which is a simple NSTextView), this can be achieved using drag&drop. Note that there is little control over the placement of the image, the image will be inline with text.

A sample output on iOS opened in Word for iOS. Furigana annotations are preserved. The link is clickable. Please note that Quicklook (on both platforms) only supports a limited number of attributes.

Supported Attributes

• most things in NSAttributedString.Key (fonts, colors, underline, indents etc.) except
  • NSAttributedString.Key.expansion
  • NSAttributedString.Key.kern
  • NSAttributedString.Key.ligature
  • NSAttributedString.Key.obliqueness
  • NSAttributedString.Key.superscript (macOS only, doesnt really work for most fonts anyway)
  • NSAttributedString.Key.textEffect
• CTRubyAnnotation for furigana (ruby) annotations in CoreText
• NSTextAttachment embedded images (inline with text)

For AttributedString, DocX supports the attributes present in NSAttributedString, i.e. most attributes in AttributeScopes.AppKitAttributes or AttributeScopes.UIKitAttributes (see above for omissions). For AttributedStrings initialized from Markdown (AttributeScopes.FoundationAttributes), DocX supports links (AttributeScopes.FoundationAttributes.LinkAttribute), bold, italic, and strikethrough (AttributeScopes.FoundationAttributes.InlinePresentationIntentAttribute), and inline images (AttributeScopes.FoundationAttributes.ImageURLAttribute). Please note that images are not rendered by SwiftUI's Text.

Some attributes don't have a direct correspondence. For example NSAttributedString does (typically) not have the concept of a page size.

Dependencies

• my fork of AEXML, many thanks to the original author tadija
• ZipFoundation

Alternatives

• WKDocReader to read old-school .doc files
• BSDocxRipperZipper which supports both reading and writing, but only for a subset of attributes.

References

• OfficeOpenXML Specification
• this blog post on Ruby annotations
• this article on the file structure of the .docx format

Licence

MIT