Authors and affiliations schema exposed to Quarto Lua API

[mattwarkentin]

Hi,

Does Quarto expose the normalized and de-normalized authors and affiliations to Lua such that it is usable by Lua filters? Based on https://quarto.org/docs/journals/authors.html it seems like those authors and affiliation schema are only usable for LaTeX.

I am trying to write a Lua filter to create author lists and affiliation lists for Word document format, and having access to by-author and by-affiliation would be amazing.

Thoughts?

[dragonstyle]

Yes it does - they are available in LUA filters and in any template (e.g. they are used in our HTML title blocks as well). Please give it a try and let us know if you run into issue - the goal is very much to support scenario like you're describing (allowing you to write author and affiliation processing for output without having to worry about the input structure)!

[mattwarkentin]

Wow, this is great. Sorry, I'm still a little new with understanding the ways in which Quarto, Pandoc, and Lua interact. Can you only access the Metadata when it is passed into the Meta function? Is there a reason you can't access the meta data anywhere in the Lua script through the Quarto Lua API, such as with something like quarto.meta['by-author']?

[dragonstyle]

No worries it is a good question - is currently available as Metadata only (so pretty straightforward to access in a filter and passed directly to a shortcode as well). Metadata in the pandoc/Quarto world is just more data in the AST (which is very nice should you choose to mutate it - those changes will persist in the AST and be accessible to other filters and templates downstream).

If you do want to capture their values just to emit other elements into the AST, you can do so something like this (which will force two passes, the first through metadata, then the second through divs). Didn't actually run this, but:

local authors
return {
  { 
    Meta = function(meta)
    authors = meta['by-author']
    end
  }, 
  { 
    Div = function(divEl)
      -- do something with authors variable here
    end
  }  
}

[mattwarkentin]

Thanks for the reply. I am stumbling out of the gates, it seems. The following code dumps nil to the console. Is there something obviously wrong such that meta['by-author'] is nil?

return {
  {
    Meta = function(meta)
      local byAuthor = meta['by-author']
      quarto.utils.dump(byAuthor)
      return meta
    end
  }
}

[dragonstyle]

Is it possible it is something as simple as not having any authors in the document front matter? Here is a simple test case that should show author data:

test.zip

If this isn't working, let me know and we can debug further!

[mattwarkentin]

Okay, your example works. I seem to have found the issue. I don't know why, but meta['by-author'] doesn't work when I specify format: native in the YAML front matter. Specifying any other format works. Is this expected? Are some steps in the processing pipeline skipped when the native format is the output.

For example, doesn't work:

---
title: "Amazing manuscript about really important scientific topic"
author: 
  name: Norah Jones
  affiliation: Blue Note Records
filters:
  - test.lua
format: native
---

pandoc 
  to: native
  output-file: delete.native
  standalone: true
  default-image-extension: png
  
metadata
  title: Amazing manuscript about really important scientific topic
  author:
    name: Norah Jones
    affiliation: Blue Note Records
  
IN META
Output created: delete.native

Works:

---
title: "Amazing manuscript about really important scientific topic"
author: 
  name: Norah Jones
  affiliation: Blue Note Records
filters:
  - test.lua
---

pandoc 
  to: html
  output-file: delete.html
  standalone: true
  section-divs: true
  html-math-method: mathjax
  wrap: none
  default-image-extension: png
  
metadata
  document-css: false
  link-citations: true
  date-format: long
  lang: en
  title: Amazing manuscript about really important scientific topic
  author:
    name: Norah Jones
    affiliation: Blue Note Records
  
IN META
List {
  [1] {
    affiliations: List {
      [1] {
        id: Inlines {
          [1] Str "aff-1"
        }
        metadata: {}
        name: Inlines {
          [1] Str "Blue"
          [2] Space
          [3] Str "Note"
          [4] Space
          [5] Str "Records"
        }
        number: "1"
      }
    }
    id: "1"
    metadata: {}
    name: {
      family: Inlines {
        [1] Str "Jones"
      }
      given: Inlines {
        [1] Str "Norah"
      }
      literal: Inlines {
        [1] Str "Norah"
        [2] Space
        [3] Str "Jones"
      }
    }
    number: "1"
  }
}
Output created: delete.html

[dragonstyle]

You are completely correct and great detective work - we are currently filtering on output type (which I'm not sure is a good idea any longer, especially as we continue to make our format output more flexible and extensible). I'm taking a look at changing this now so we process author metadata without regard to the specified format (with the exception of markdown output, which I think we should continue to leave untransformed).

[mattwarkentin]

In case anyone stumbles upon this Discussion because they too are trying to have authors and affiliations work consistently across Word and PDF academic manuscripts destined for peer-review, this is my start on a Lua filter for exactly that. I'm sure it could be improved in many ways as I am new to programming in Lua. But for now, this works. I took inspiration from: https://github.com/pandoc/lua-filters/tree/master/author-info-blocks and from the suggestions of @dragonstyle.

local List = require 'pandoc.List'
local utils = require 'pandoc.utils'
local stringify = utils.stringify
local byAuthor
local byAffiliation
local Authors = {}
local Affiliations = {}
local Corresponding
local equalCont
local authorHoriz

local function make_correspondance(name, email)
  correspondance = List:new{
    pandoc.Str('* Corresponding Author: '),
    pandoc.Str(name),
    pandoc.Str(' ('),
    pandoc.Link(email, 'mailto:' .. email),
    pandoc.Str(')')
  }
  Corresponding = List:new{pandoc.Span(correspondance)}
end

local function make_equal_contributor()
  eq_statement = pandoc.Str('† These authors contributed equally to this work.')
  equalCont = List:new{pandoc.Span(eq_statement)}
end

local function create_author_list(byAuthor)
  local authors = {}
  for i,author in ipairs(byAuthor) do
    local sups = {}

    for j,aff in ipairs(author.affiliations) do
      table.insert(sups, aff.number)
    end
    sups_str = table.concat(sups, ',')

    local name = stringify(author.name.literal)

    if author.attributes ~= nil then
      if author.attributes.equal ~= nil and author.attributes.equal then
        sups_str = sups_str .. '†'
        make_equal_contributor()
      end
      if author.attributes.corresponding ~= nil and author.attributes.corresponding then
        local email = stringify(author.email)
        sups_str = sups_str .. '*'
        make_correspondance(name, email)
      end
    end

    local authorEntry = List:new{
      pandoc.Str(name),
      pandoc.Superscript(pandoc.Str(sups_str))
    }

    if authorHoriz then
      authorEntry:extend({pandoc.Str(', ')})
    end

    table.insert(authors, pandoc.Span(authorEntry))
  end

  if authorHoriz then
    Authors = {pandoc.Para(authors)}
  else
    Authors = authors
  end
end

local function create_affiliation_list(byAffiliation)
  for i,aff in ipairs(byAffiliation) do
    local number = stringify(aff.number)
    local name = stringify(aff.name)
    local address = stringify(aff.address)
    local city = stringify(aff.city)
    local region = stringify(aff.region)
    local country = stringify(aff.country)
    local full_aff = {name, address, city, region, country}
    local entry = table.concat(full_aff, ', ')
    entry = number .. '. ' .. entry
    table.insert(Affiliations, pandoc.Span(pandoc.Str(entry)))
  end
end

local function remove_author_meta(meta)
  meta.author = nil
  meta.authors = nil
  meta.affiliations = nil
  meta['by-author'] = nil
  meta['by-affiliation'] = nil
  return meta
end

return {
  {
    Meta = function(meta)
      byAuthor = meta['by-author']
      byAffiliation = meta['by-affiliation']
      
      if meta['author-horizontal'] ~= nil then
        authorHoriz = meta['author-horizontal']
      else
        authorHoriz = false
      end

      create_author_list(byAuthor)
      create_affiliation_list(byAffiliation)
      return meta
    end
  },
  {
    Pandoc = function(doc)
      local meta = doc.meta
      local body = List:new{}
      body:extend(Authors)
      body:extend(Affiliations)
      body:extend(equalCont)
      body:extend(Corresponding)
      body:extend(doc.blocks)
      meta = remove_author_meta(meta)
      return pandoc.Pandoc(body, meta)
    end
  }
}

[cscheid]

Hey this looks great! One small suggestion. I think you don't need to split your filters in two, like you did:

{
  {
    Meta = ...
  },
  {
    Pandoc = ...
  }
}

and could instead just do

{
  {
    Meta = function(...)
    end,
    Pandoc = ...
  }
}

This change is 99.9% aesthetics, but it's probably worth doing for future readers of your code. You'd use that two-step process to tell Pandoc (and your readers) that you have two filters, one for Meta and one for Pandoc, and they need to be run sequentially. With the latter syntax, you are telling Pandoc to run a single filter, which is ok because Meta calls always happen before Pandoc calls in lua filters: https://pandoc.org/lua-filters.html#typewise-traversal. They're almost exactly equivalent in this case, of course, but the latter is probably more idiomatic.

[mattwarkentin]

Thanks for the suggestion, @cscheid! This is good to know. I'm still learning how to program with Quarto, Lua, and Pandoc, so I will take all suggestions on following best practices.

[mingsu]

It is exactly what I need.

Is it possible to adapt this lua filter work with the following author-affiliations structure?

cat test.qmd

---
title: Test Document
author: 
  - name: Norah Jones
    email: [email protected]
    affiliations:
      - ref: aff1
      - ref: aff2
affiliations:
  - id: aff1
    name: Blue Note Records
  - id: aff2
    name: Green Note Records
filters:
  - test.lua
format: 
  docx: default
  html: default
---

## This is a test

Currently, with the code above, I got error as shown below.

pandoc 
  to: docx
  output-file: test.docx
  default-image-extension: png
  
metadata
  title: Test Document
  author:
    - name: Norah Jones
      email: [email protected]
      affiliations:
        - ref: aff1
        - ref: aff2
  affiliations:
    - id: aff1
      name: Blue Note Records
    - id: aff2
      name: Green Note Records
  
Error running filter test.lua:
PandocLuaError "Cannot get Attr from TypeNil"
stack traceback:
	test.lua:75: in upvalue 'create_affiliation_list'
	test.lua:108: in function <test.lua:97>

Here is the test.lua copied from above.

local List = require 'pandoc.List'
local utils = require 'pandoc.utils'
local stringify = utils.stringify
local byAuthor
local byAffiliation
local Authors = {}
local Affiliations = {}
local Corresponding
local equalCont
local authorHoriz

local function make_correspondance(name, email)
  correspondance = List:new{
    pandoc.Str('* Corresponding Author: '),
    pandoc.Str(name),
    pandoc.Str(' ('),
    pandoc.Link(email, 'mailto:' .. email),
    pandoc.Str(')')
  }
  Corresponding = List:new{pandoc.Span(correspondance)}
end

local function make_equal_contributor()
  eq_statement = pandoc.Str('† These authors contributed equally to this work.')
  equalCont = List:new{pandoc.Span(eq_statement)}
end

local function create_author_list(byAuthor)
  local authors = {}
  for i,author in ipairs(byAuthor) do
    local sups = {}

    for j,aff in ipairs(author.affiliations) do
      table.insert(sups, aff.number)
    end
    sups_str = table.concat(sups, ',')

    local name = stringify(author.name.literal)

    if author.attributes ~= nil then
      if author.attributes.equal ~= nil and author.attributes.equal then
        sups_str = sups_str .. '†'
        make_equal_contributor()
      end
      if author.attributes.corresponding ~= nil and author.attributes.corresponding then
        local email = stringify(author.email)
        sups_str = sups_str .. '*'
        make_correspondance(name, email)
      end
    end

    local authorEntry = List:new{
      pandoc.Str(name),
      pandoc.Superscript(pandoc.Str(sups_str))
    }

    if authorHoriz then
      authorEntry:extend({pandoc.Str(', ')})
    end

    table.insert(authors, pandoc.Span(authorEntry))
  end

  if authorHoriz then
    Authors = {pandoc.Para(authors)}
  else
    Authors = authors
  end
end

local function create_affiliation_list(byAffiliation)
  for i,aff in ipairs(byAffiliation) do
    local number = stringify(aff.number)
    local name = stringify(aff.name)
    local address = stringify(aff.address)
    local city = stringify(aff.city)
    local region = stringify(aff.region)
    local country = stringify(aff.country)
    local full_aff = {name, address, city, region, country}
    local entry = table.concat(full_aff, ', ')
    entry = number .. '. ' .. entry
    table.insert(Affiliations, pandoc.Span(pandoc.Str(entry)))
  end
end

local function remove_author_meta(meta)
  meta.author = nil
  meta.authors = nil
  meta.affiliations = nil
  meta['by-author'] = nil
  meta['by-affiliation'] = nil
  return meta
end

return {
  {
    Meta = function(meta)
      byAuthor = meta['by-author']
      byAffiliation = meta['by-affiliation']
      
      if meta['author-horizontal'] ~= nil then
        authorHoriz = meta['author-horizontal']
      else
        authorHoriz = false
      end

      create_author_list(byAuthor)
      create_affiliation_list(byAffiliation)
      return meta
    end
  },
  {
    Pandoc = function(doc)
      local meta = doc.meta
      local body = List:new{}
      body:extend(Authors)
      body:extend(Affiliations)
      body:extend(equalCont)
      body:extend(Corresponding)
      body:extend(doc.blocks)
      meta = remove_author_meta(meta)
      return pandoc.Pandoc(body, meta)
    end
  }
}

[dragonstyle]

Here is a pretty functional version (at least in my testing briefly) with keywords:

local List = require("pandoc.List")
local utils = require("pandoc.utils")
local stringify = utils.stringify
local byAuthor
local byAffiliation
local Authors = {}
local Affiliations = {}


local authorHoriz


local Corresponding = nil
local function make_correspondance(name, email)
  correspondance = List:new({
    pandoc.Str("* Corresponding Author: "),
    pandoc.Str(name),
    pandoc.Str(" ("),
    pandoc.Link(email, "mailto:" .. email),
    pandoc.Str(")"),
  })
  Corresponding = List:new({ pandoc.Span(correspondance) })
end

local equalCont
local function make_equal_contributor()
  eq_statement = pandoc.Str("† These authors contributed equally to this work.")
  equalCont = List:new({ pandoc.Span(eq_statement) })
end


local Keywords = nil
local function create_keyword_list(kw)
  Keywords = {}
  table.insert(Keywords, pandoc.Str("Keywords: "))
  quarto.log.output(kw)
  for i, keyword in ipairs(kw) do
      table.insert(Keywords, keyword)
  end
end

local function create_author_list(byAuthor)
  local authors = {}
  for i, author in ipairs(byAuthor) do
    local sups = {}

    for j, aff in ipairs(author.affiliations) do
      table.insert(sups, aff.number)
    end
    sups_str = table.concat(sups, ",")

    local name = stringify(author.name.literal)

    if author.attributes ~= nil then
      if author.attributes['equal-contributor'] ~= nil and author.attributes['equal-contributor'] then
        sups_str = sups_str .. "†"
        make_equal_contributor()
      end
      if author.attributes.corresponding ~= nil and author.attributes.corresponding then
        local email = stringify(author.email)
        sups_str = sups_str .. ",*"
        make_correspondance(name, email)
      end
    end

    local authorEntry = List:new({
      pandoc.Str(name),
      pandoc.Superscript(pandoc.Str(sups_str)),
    })

    if authorHoriz and i < #byAuthor then
      authorEntry:extend({ pandoc.Str(", ") })
    end

    table.insert(authors, pandoc.Span(authorEntry))
  end

  if authorHoriz then
    Authors = { pandoc.Para(authors) }
  else
    Authors = authors
  end
end

local function create_affiliation_list(byAffiliation)
  for i, aff in ipairs(byAffiliation) do
    local full_aff = pandoc.List({})
    if aff.name then
      full_aff:insert(stringify(aff.name))
    end

    if aff.address then
      full_aff:insert(stringify(aff.address))
    end

    if aff.city then
      full_aff:insert(stringify(aff.city))
    end

    if aff.region then
      full_aff:insert(stringify(aff.region))
    end

    if aff.country then
      full_aff:insert(stringify(aff.country))
    end

    local entry = table.concat(full_aff, ", ")
    -- TODO
    -- Don't know how to add a linebreak after each affiliation
    entry = aff.number .. ". " .. entry
    table.insert(Affiliations, pandoc.Span(pandoc.Str(entry)))
  end
end

local function remove_author_meta(meta)
  meta.author = nil
  meta.authors = nil
  meta.affiliations = nil
  meta["by-author"] = nil
  meta["by-affiliation"] = nil
  return meta
end

return {
  {
    Meta = function(meta)
      byAuthor = meta["by-author"]
      byAffiliation = meta["by-affiliation"]
      if meta["author-horizontal"] ~= nil then
        authorHoriz = meta["author-horizontal"]
      else
        authorHoriz = true
      end
      create_author_list(byAuthor)
      create_affiliation_list(byAffiliation)
      if meta["keywords"] ~= nil then
        create_keyword_list(meta["keywords"])
      end
      return meta
    end,
  },
  {
    Pandoc = function(doc)
      local meta = doc.meta
      local body = List:new({})
      body:extend(Authors)
      body:extend(Affiliations)
      if equalCont ~= nil then
        body:extend(equalCont)
      end
      if Corresponding ~= nil then
        body:extend(Corresponding)
      end
      if Keywords then
        body:extend(Keywords)
      end
      body:extend(doc.blocks)
      meta = remove_author_meta(meta)
      return pandoc.Pandoc(body, meta)
    end,
  },
}

[mingsu]

Thanks a ton, finally make it work as expected.

I have made some slight modifications, to put all keywords in one line, and also add the support for abstract.

local List = require("pandoc.List")
local utils = require("pandoc.utils")
local stringify = utils.stringify
local byAuthor
local byAffiliation
local Authors = {}
local Affiliations = {}

local authorHoriz

local Corresponding = nil
local function make_correspondance(name, email)
	correspondance = List:new({
		pandoc.Str("* Corresponding Author: "),
		pandoc.Str(name),
		pandoc.Str(" ("),
		pandoc.Link(email, "mailto:" .. email),
		pandoc.Str(")"),
	})
	Corresponding = List:new({ pandoc.Para(correspondance) })
end

local equalCont
local function make_equal_contributor()
	eq_statement = pandoc.Str("† These authors contributed equally to this work.")
	equalCont = List:new({ pandoc.Para(eq_statement) })
end

local function create_author_list(byAuthor)
	local authors = {}
	for i, author in ipairs(byAuthor) do
		local sups = {}

		for j, aff in ipairs(author.affiliations) do
			table.insert(sups, aff.number)
		end
		sups_str = table.concat(sups, ",")

		local name = stringify(author.name.literal)

		if author.attributes ~= nil then
			if author.attributes["equal-contributor"] ~= nil and author.attributes["equal-contributor"] then
				sups_str = sups_str .. ",†"
				make_equal_contributor()
			end
			if author.attributes.corresponding ~= nil and author.attributes.corresponding then
				local email = stringify(author.email)
				sups_str = sups_str .. ",*"
				make_correspondance(name, email)
			end
		end

		local authorEntry = List:new({
			pandoc.Str(name),
			pandoc.Superscript(pandoc.Str(sups_str)),
		})

		if authorHoriz and i < #byAuthor then
			authorEntry:extend({ pandoc.Str(", ") })
		end

		table.insert(authors, pandoc.Span(authorEntry))
	end

	if authorHoriz then
		Authors = { pandoc.Para(authors) }
	else
		Authors = authors
	end
end

local function create_affiliation_list(byAffiliation)
	for i, aff in ipairs(byAffiliation) do
		local full_aff = pandoc.List({})
		if aff.name then
			full_aff:insert(stringify(aff.name))
		end

		if aff.address then
			full_aff:insert(stringify(aff.address))
		end

		if aff.city then
			full_aff:insert(stringify(aff.city))
		end

		if aff["postal-code"] then
			full_aff:insert(stringify(aff["postal-code"]))
		end

		if aff.region then
			full_aff:insert(stringify(aff.region))
		end

		if aff.country then
			full_aff:insert(stringify(aff.country))
		end

		local entry = table.concat(full_aff, ", ")
		entry = aff.number .. ". " .. entry .. "."
		table.insert(Affiliations, pandoc.Para(pandoc.Str(entry)))
	end
end

local Abstract = nil
local function create_abstract(ab)
	Abstract = {}
	table.insert(Abstract, pandoc.Header(1, "Abstract"))
	table.insert(Abstract, pandoc.Para(ab))
end

local Keywords = nil
local function create_keyword_list(kw)
	Keywords = {}
	-- quarto.log.output(kw)
	local kws = pandoc.List({})
	for i, keyword in ipairs(kw) do
		kws:insert(stringify(keyword))
	end
	local kwentry = table.concat(kws, "; ")
	kwentry = "Keywords: " .. kwentry .. "."
	table.insert(Keywords, pandoc.Para(pandoc.Str(kwentry)))
end

local function remove_author_meta(meta)
	meta.author = nil
	meta.authors = nil
	meta.affiliations = nil
	meta["by-author"] = nil
	meta["by-affiliation"] = nil
	meta["abstract"] = nil
	return meta
end

return {
	{
		Meta = function(meta)
			byAuthor = meta["by-author"]
			byAffiliation = meta["by-affiliation"]
			if meta["author-horizontal"] ~= nil then
				authorHoriz = meta["author-horizontal"]
			else
				authorHoriz = true
			end
			create_author_list(byAuthor)
			create_affiliation_list(byAffiliation)
			if meta["abstract"] ~= nil then
				create_abstract(meta["abstract"])
			end
			if meta["keywords"] ~= nil then
				create_keyword_list(meta["keywords"])
			end
			return meta
		end,
	},
	{
		Pandoc = function(doc)
			local meta = doc.meta
			local body = List:new({})
			body:extend(Authors)
			body:extend(Affiliations)
			if equalCont ~= nil then
				body:extend(equalCont)
			end
			if Corresponding ~= nil then
				body:extend(Corresponding)
			end
			if Abstract then
				body:extend(Abstract)
			end
			if Keywords then
				body:extend(Keywords)
			end
			body:extend(doc.blocks)
			meta = remove_author_meta(meta)
			return pandoc.Pandoc(body, meta)
		end,
	},
}

[jmlondon]

Thank you very much, @mattwarkentin, @dragonstyle, and @mingsu for sharing all of this information and code. I was able to follow the Quarto docs for Author & Affiliations and use the lua code shared here with ease.

That said, I'm noticing that for both pdf and docx outputs, the author & affiliation information is not appearing after document title. Instead, the author/affiliation info is inserted after the date section (or, if toc, lof, lot are true, after those elements). For HTML, the author/affiliation information shows up after the date section instead of after the title but the impact is less disruptive because the toc is off in a sidebar.

Also of note, the author/affiliation information is not tagged/labeled with the 'Author' style in MS Word for consistency with any reference docx provided.

I've spent some time trying to get up to speed on lua, pandoc filters, etc. but I'm not quite grasping how I would approach solving these issues.

[mattwarkentin]

Hi @jmlondon,

Glad it was helpful. It is hard to offer specific advice without knowing more about the structure of your documents. Here is a repo that has the Lua filter described in this thread, in a working manuscript format for PDF and Word. Maybe it will be helpful (https://github.com/oncology-outcomes/quarto-manuscript).

In addition to the authors-affiliations.lua filter, I also setup partials (for PDF) which helps to make sure the content is placed in the right location. Quarto uses a default PDF template which orders things in a certain way - sometimes you have to make changes to customize the order.

The Lua filter above places the authors and affiliations as the first thing in the body and before the rest of the existing content (doc.blocks), shown here:

    Pandoc = function(doc)
      local meta = doc.meta
      local body = List:new({})
      body:extend(Authors)
      body:extend(Affiliations)
      if equalCont ~= nil then
        body:extend(equalCont)
      end
      if Corresponding ~= nil then
        body:extend(Corresponding)
      end
      body:extend(doc.blocks)
      meta = remove_author_meta(meta)
      return pandoc.Pandoc(body, meta)
    end

However, I also set partials/title.tex to be empty (this stops the default behaviour), and I used before-body.tex to control where the title gets placed (before the body, so just before the Authors/Affiliations). ToC are a little tricker if you want them after the Authors/Affiliations but before "normal" content, because the TeX to PDF conversion occurs after the Lua stuff is done.

Perhaps you can modify the Lua filter to inject some raw LaTeX (\tableofcontents) before the line body:extend(doc.blocks) in the above filter?

[mattwarkentin]

Maybe body:extend({ pandoc.RawBlock('tex', '\tableofcontents') }) ??

[jmlondon]

Hi @mattwarkentin,

To provide a simple reprex and demo of the issue, I'll use your quarto-manuscript extension.

first, create an empty Quarto project and then install the extension with:

quarto add oncology-outcomes/quarto-manuscript

Then, copy your template.qmd over and edit the YAML to include generation of a table of contents like so:

---
title: "This is an example manuscript"
toc: true
format: 
  manuscript-docx: default
  manuscript-pdf: default
...

When I render to docx, I get the following (docx on left, pdf on right):

Expectations would be for the author/affiliation information to appear immediately after the Title and before the Table of Contents. Additionally, the MS Word style associated with line 10 ("Matthew T. Warkentin") is 'First Paragraph' and the remaining author/affiliation information is 'Body Text'. This means any MS Word style specified for 'Author' in the reference/template docx isn't applied to the author/affiliation text. And, any style information specified for 'First Paragraph' might cause issues.

I think you might be on to something with your statement that Lua filter above places the authors and affiliations as the first thing in the body and before the rest of the existing content (doc.blocks). This is exactly what's happening. So the Title and Table of Contents must be contained outside the 'body' and in some other section that represents the title block.

Seems there must be a way to access the title block so the author/affiliation data is inserted in that section instead of in the body. Similarly, seems we'd want to make sure style tag/labels are properly associated with the content.

[jmlondon]

Following after some additional searching, it seems the papaja package might have sorted out a possible solution/approach.

See their parse_metadata.lua file at https://github.com/crsh/papaja/blob/main/inst/lua/parse_metadata.lua. It's all too advanced for me to grasp and sort out at this point in my lua skills. But, it seems their approach is to actually modify meta.author, meta.affiliation, etc instead of parsing and adding to the beginning of body.

I was able to confirm that a docx document from papaja does work with table of contents and also maintains the style associations.