Class: Uniword::Docx::PartLoader::RawPartLoader

Inherits:
Object
  • Object
show all
Defined in:
lib/uniword/docx/part_loader/raw_part_loader.rb

Overview

Fallback loader: preserves ZIP entries no registry definition claims as raw parts (RawPart) on the package, so unmodelled parts (docProps/meta.xml, glossary documents, VBA projects, unmodelled .rels sidecars, ...) round-trip byte-for-byte with their content types and relationships intact.

Runs last (load_priority 130 on the :raw_parts definition), so the registry claims every path it models first and this loader carries only the remainder — no double emission. A path is claimed when any PartDefinition matches it (loadable or not), when it is the resolved main document part (or its rels sidecar) at a non-standard location, or when it is a customXml item rels sidecar consumed by CustomXmlLoader.

Bytes are re-read from the original ZIP when available (the extracted hash may carry corrupted UTF-8 binary); content types come from the loaded [Content_Types].xml model.

Constant Summary collapse

CUSTOM_XML_ITEM_RELS =

customXml item rels sidecars are consumed by CustomXmlLoader together with their item part (captured group: item index).

%r{\AcustomXml/_rels/item(\d+)\.xml\.rels\z}

Instance Method Summary collapse

Instance Method Details

#load(context, _definition) ⇒ void

This method returns an undefined value.

Parameters:

  • context (LoadContext) —

    shared load state

  • _definition (Ooxml::PartDefinition) —

    :raw_parts (strategy interface; the loader claims by remainder, not by definition matching)



34
35
36
37
38
39
40
41
42
43
44
45
46
 
# File 'lib/uniword/docx/part_loader/raw_part_loader.rb', line 34

def load(context, _definition)
  paths = unclaimed_paths(context)
  return if paths.empty?

  bytes = raw_bytes(context, paths)
  paths.each do |path|
    next unless bytes[path]

    context.package.raw_parts[path] = build_part(
      context.package, path, bytes[path]
    )
  end
end