feat: 完成 DOCX 通用结构映射与严格收口

packages/docx-engine/assets/docx-media-filter.lua

@@ -1,18 +1,29 @@
 local map_path = os.getenv("MD_TO_PDF_DOCX_MEDIA_MAP")
+local structure_plan_path =
+  os.getenv("MD_TO_PDF_DOCX_STRUCTURE_PLAN")
 
 if map_path == nil or map_path == "" then
   error("DOCX media map path is missing")
 end
-
-local map_file, open_error = io.open(map_path, "rb")
-if map_file == nil then
-  error("DOCX media map cannot be opened: " .. tostring(open_error))
+if structure_plan_path == nil or structure_plan_path == "" then
+  error("DOCX structure plan path is missing")
 end
 
-local map_content = map_file:read("*a")
-map_file:close()
+local function read_json(path, description)
+  local file, open_error = io.open(path, "rb")
+  if file == nil then
+    error(
+      description .. " cannot be opened: " .. tostring(open_error)
+    )
+  end
+  local content = file:read("*a")
+  file:close()
+  return pandoc.json.decode(content, false)
+end
 
-local media_map = pandoc.json.decode(map_content, false)
+local media_map = read_json(map_path, "DOCX media map")
+local structure_plan =
+  read_json(structure_plan_path, "DOCX structure plan")
 local counters = {
   image = 0,
   mermaid = 0,
@@ -86,6 +97,69 @@ local function replace_code_block(block)
   return pandoc.Para { image }
 end
 
+local function append_inlines(target, value)
+  target:extend(pandoc.Inlines(value))
+end
+
+local function structure_paragraph(block)
+  local inlines = pandoc.Inlines {}
+  for index, segment in ipairs(block.segments) do
+    if index > 1 then
+      if block.separator == "tab" then
+        inlines:insert(
+          pandoc.RawInline(
+            "openxml",
+            "<w:r><w:tab/></w:r>"
+          )
+        )
+      else
+        inlines:insert(pandoc.Space())
+      end
+    end
+    append_inlines(inlines, segment)
+  end
+  local paragraph = pandoc.Para(inlines)
+  if block.styleId == nil or block.styleId == "" then
+    return paragraph
+  end
+  return pandoc.Div(
+    { paragraph },
+    pandoc.Attr("", {}, { ["custom-style"] = block.styleId })
+  )
+end
+
+local function marker_paragraph(marker)
+  return pandoc.Para { pandoc.Str(marker) }
+end
+
+local function structure_blocks(blocks)
+  local result = pandoc.Blocks {}
+  for _, block in ipairs(blocks) do
+    if block.kind == "paragraph" then
+      result:insert(structure_paragraph(block))
+    elseif block.kind == "container" then
+      local children = structure_blocks(block.blocks)
+      result:insert(marker_paragraph(block.startMarker))
+      result:extend(children)
+      result:insert(marker_paragraph(block.endMarker))
+    elseif block.kind == "section-break" then
+      result:insert(marker_paragraph(block.marker))
+    else
+      error("Unsupported DOCX structure block: " .. tostring(block.kind))
+    end
+  end
+  return result
+end
+
+local function suppress_first_body_heading(blocks)
+  for index, block in ipairs(blocks) do
+    if block.t == "Header" and block.level == 1 then
+      blocks:remove(index)
+      return
+    end
+  end
+end
+
 local function validate_counts()
   for _, kind in ipairs({ "image", "mermaid", "echarts" }) do
     if counters[kind] ~= #media_map[kind] then
@@ -104,6 +178,16 @@ return {
       Image = replace_image,
       CodeBlock = replace_code_block
     }
+    if structure_plan.titlePolicy.metadataTitle == "suppress" then
+      transformed.meta.title = nil
+    end
+    if structure_plan.titlePolicy.firstBodyHeading == "suppress" then
+      suppress_first_body_heading(transformed.blocks)
+    end
+    local blocks = structure_blocks(structure_plan.prefix)
+    blocks:extend(transformed.blocks)
+    blocks:extend(structure_blocks(structure_plan.suffix))
+    transformed.blocks = blocks
     validate_counts()
     return transformed
   end