File: usfm2html.cpp

package info (click to toggle)
bibledit-cloud 5.1.036-1
links: PTS, VCS
area: main
in suites: forky, sid
size: 250,636 kB
sloc: xml: 915,934; ansic: 261,349; cpp: 92,628; javascript: 32,542; sh: 4,915; makefile: 586; php: 69
file content (1103 lines) | stat: -rw-r--r-- 40,307 bytes
parent folder | download | duplicates (2)
/*
 Copyright (©) 2003-2025 Teus Benschop.
 
 This program is free software; you can redistribute it and/or modify
 it under the terms of the GNU General Public License as published by
 the Free Software Foundation; either version 3 of the License, or
 (at your option) any later version.
 
 This program is distributed in the hope that it will be useful,
 but WITHOUT ANY WARRANTY; without even the implied warranty of
 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License
 along with this program; if not, write to the Free Software
 Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
 */


#include <editor/usfm2html.h>
#include <filter/string.h>
#include <filter/url.h>
#include <filter/usfm.h>
#include <locale/translate.h>
#include <styles/logic.h>
#include <database/logs.h>
#include <filter/quill.h>


void Editor_Usfm2Html::load (std::string usfm)
{
  // Clean the USFM up.
  usfm = filter::strings::trim (std::move(usfm));
  usfm.append ("\n");
  // Separate the USFM into markers and text fragments.
  // Load it into the object.
  m_markers_and_text = filter::usfm::get_markers_and_text (usfm);
  m_markers_and_text_pointer = 0;
}


void Editor_Usfm2Html::stylesheet (const std::string& stylesheet)
{
  m_stylesheet = stylesheet;
  // Load style information into the object.
  for (const stylesv2::Style& style : database::styles::get_styles(stylesheet)) {
    m_note_citations.evaluate_style(style);
  }
}


void Editor_Usfm2Html::run ()
{
  preprocess ();
  process ();
}


std::string Editor_Usfm2Html::get ()
{
  close_paragraph ();
  
  // If there are any notes, move the notes <p> after everything else.
  // (It has the <hr> or <br> as a child).
  {
    const long count = std::distance (m_notes_node.begin (), m_notes_node.end ());
    if (count > 1) {
      m_body_node.append_move (m_notes_node);
    }
  }

  // A Quill-based editor does not work with embedded <p> elements.
  // Move the notes out of their parent and append them to the end of the main body.
  while (pugi::xml_node note = m_notes_node.first_child ().next_sibling ()) {
    m_body_node.append_move (note);
  }

  // If there are word-level attributes, move the word-level attributes <p> after everything else.
  {
    const long count = std::distance (m_word_level_attributes_node.begin (), m_word_level_attributes_node.end ());
    if (count > 1) {
      m_body_node.append_move (m_word_level_attributes_node);
    }
  }

  // A Quill-based editor does not work with embedded <p> elements.
  // Move the word-level attributes out of their parent and append them to the end of the main body.
  while (pugi::xml_node word_level_attribute = m_word_level_attributes_node.first_child ().next_sibling ()) {
    m_body_node.append_move (word_level_attribute);
  }

  // If there are milestone attributes, move the milestone attributes <p> after everything else.
  {
    const long count = std::distance (m_milestone_attributes_node.begin (), m_milestone_attributes_node.end ());
    if (count > 1) {
      m_body_node.append_move (m_milestone_attributes_node);
    }
  }

  // A Quill-based editor does not work with embedded <p> elements.
  // Move the milestone attributes out of their parent and append them to the end of the main body.
  while (pugi::xml_node milestone_attribute = m_milestone_attributes_node.first_child ().next_sibling ()) {
    m_body_node.append_move (milestone_attribute);
  }

  // Get the html code, including body, without head.
  std::stringstream output {};
  m_body_node.print (output, "", pugi::format_raw);
  std::string html = output.str ();
  
  // Remain with the stuff within the <body> elements.
  constexpr const std::string_view body {"<body>"};
  if (const size_t pos1 = html.find (body); pos1 != std::string::npos) {
    html.erase (0, pos1 + body.length());
    if (const size_t pos2 = html.find ("</body>"); pos2 != std::string::npos) {
      html.erase (pos2);
    }
  }
  
  // Currently the XML library produces hexadecimal character entities.
  // This is unwanted behaviour: Convert them to normal characters.
  html = filter::strings::convert_xml_character_entities_to_characters (std::move(html));
  
  // Result.
  return html;
}


void Editor_Usfm2Html::preprocess ()
{
  m_current_paragraph_style.clear ();
  m_current_paragraph_content.clear ();
  m_current_text_styles.clear();
  m_note_count = 0;
  m_current_note_text_styles.clear();
  m_text_tength = 0;
  m_verse_start_offsets = { std::pair (0, 0) };
  m_current_p_open = false;
  m_note_p_open = false;
  
  m_body_node = m_document.append_child ("body");
  
  // Create the xml node for the notes container.
  // It comes at the start of the document.
  // (Later, it will either be deleted, or moved to the end).
  {
    m_notes_node = m_document.append_child ("p");
    std::string notes_class = quill::notes_class;
    notes_class.insert (0, quill::class_prefix_block);
    m_notes_node.append_attribute ("class") = notes_class.c_str ();
    m_notes_node.text().set(filter::strings::non_breaking_space_u00A0().c_str());
  }
  
  // Create the xml node for the word-level attributes container.
  // It comes near the start of the document.
  // (Later, it will either be deleted, or moved to the end).
  {
    m_word_level_attributes_node = m_document.append_child ("p");
    std::string word_level_attributes_class = quill::word_level_attributes_class;
    word_level_attributes_class.insert (0, quill::class_prefix_block);
    m_word_level_attributes_node.append_attribute ("class") = word_level_attributes_class.c_str ();
    m_word_level_attributes_node.text().set(filter::strings::non_breaking_space_u00A0().c_str());
  }
  
  // Create the xml node for the milestone attributes container.
  // It comes near the start of the document.
  // (Later, it will either be deleted, or moved to the end).
  {
    m_milestone_attributes_node = m_document.append_child ("p");
    std::string milestone_attributes_class = quill::milestone_attributes_class;
    milestone_attributes_class.insert (0, quill::class_prefix_block);
    m_milestone_attributes_node.append_attribute ("class") = milestone_attributes_class.c_str ();
    m_milestone_attributes_node.text().set(filter::strings::non_breaking_space_u00A0().c_str());
  }
}


void Editor_Usfm2Html::process ()
{
  m_markers_and_text_pointer = 0;
  for (m_markers_and_text_pointer = 0; m_markers_and_text_pointer < m_markers_and_text.size(); m_markers_and_text_pointer++) {
    const std::string current_item = m_markers_and_text[m_markers_and_text_pointer];
    if (filter::usfm::is_usfm_marker (current_item))
    {
      // Store indicator whether the marker is an opening marker.
      const bool is_opening_marker = filter::usfm::is_opening_marker (current_item);
      const bool is_embedded_marker = filter::usfm::is_embedded_marker (current_item);
      // Clean up the marker, so we remain with the basic version, e.g. 'id'.
      const std::string marker = filter::usfm::get_marker (current_item);
      // Handle preview mode: Strip word-level attributes.
      // Likely this can be removed from preview since word-level ttributes get extracted.
      if (m_preview)
        if (is_opening_marker) {
          filter::usfm::remove_word_level_attributes (marker, m_markers_and_text, m_markers_and_text_pointer);
        }
      if (const stylesv2::Style* style {database::styles::get_marker_data (m_stylesheet, marker)}; style)
      {
        switch (style->type) {
          case stylesv2::Type::starting_boundary:
          case stylesv2::Type::none:
          case stylesv2::Type::book_id:
          case stylesv2::Type::usfm_version:
          case stylesv2::Type::file_encoding:
          case stylesv2::Type::remark:
          case stylesv2::Type::running_header:
          case stylesv2::Type::long_toc_text:
          case stylesv2::Type::short_toc_text:
          case stylesv2::Type::book_abbrev:
          {
            // Output as plain text.
            close_text_style (false);
            output_as_is (marker, is_opening_marker);
            break;
          }
          case stylesv2::Type::introduction_end:
          {
            // Output as plain text.
            close_text_style (false);
            output_as_is (marker, is_opening_marker);
            break;
          }
          case stylesv2::Type::title:
          case stylesv2::Type::heading:
          case stylesv2::Type::paragraph:
          {
            // Output as a paragraph.
            close_text_style (false);
            close_paragraph ();
            if (is_opening_marker)
              new_paragraph (marker);
            break;
          }
          case stylesv2::Type::chapter:
          {
            close_text_style (false);
            close_paragraph ();
            new_paragraph (marker);
            break;
          }
          case stylesv2::Type::chapter_label:
          case stylesv2::Type::published_chapter_marker:
          case stylesv2::Type::alternate_chapter_number:
          {
            // Output as plain text.
            close_text_style (false);
            output_as_is (marker, is_opening_marker);
            break;
          }
          case stylesv2::Type::table_row:
          {
            close_text_style (false);
            output_as_is (marker, is_opening_marker);
            break;
          }
          case stylesv2::Type::table_heading:
          case stylesv2::Type::table_cell:
          {
            close_text_style (false);
            open_text_style (style->marker, false);
            break;
          }
          case stylesv2::Type::verse:
          {
            // Close any existing text style.
            close_text_style (false);
            // Output the space before the verse number in case the paragraph already has some text.
            if (!m_current_paragraph_content.empty()) {
              add_text (" ");
            }
            // Open verse style, record verse/length, add verse number, close style again, and add a space.
            open_text_style (style->marker, false);
            std::string text_following_marker = filter::usfm::get_text_following_marker (m_markers_and_text, m_markers_and_text_pointer);
            const std::string number = filter::usfm::peek_verse_number (text_following_marker);
            m_verse_start_offsets [filter::strings::convert_to_int (number)] = static_cast<int>(m_text_tength);
            add_text (number);
            close_text_style (false);
            add_text (" ");
            // If there was any text following the \v marker, remove the verse number,
            // put the remainder back into the object, and update the pointer.
            if (!text_following_marker.empty()) {
              if (const size_t pos = text_following_marker.find (number);
                  pos != std::string::npos) {
                text_following_marker = text_following_marker.substr (pos + number.length());
              }
              text_following_marker = filter::strings::ltrim (text_following_marker);
              m_markers_and_text [m_markers_and_text_pointer] = text_following_marker;
              m_markers_and_text_pointer--;
            }
            break;
          }
          case stylesv2::Type::published_verse_marker:
          case stylesv2::Type::alternate_verse_marker:
          {
            // Treat the \vp ...\vp* marker as inline text.
            // Same for \va ...\va*.
            if (is_opening_marker) {
              open_text_style (style->marker, is_embedded_marker);
            } else {
              close_text_style (is_embedded_marker);
            }
            break;
          }
          case stylesv2::Type::footnote_wrapper:
          case stylesv2::Type::endnote_wrapper:
          {
            close_text_style (false);
            if (is_opening_marker) {
              m_last_citation = m_note_citations.get (style->marker, "+");
              add_note (m_last_citation, marker);
            } else {
              close_current_note ();
            }
            break;
          }
          case stylesv2::Type::note_standard_content:
          case stylesv2::Type::note_content:
          case stylesv2::Type::note_content_with_endmarker:
          {
            if (is_opening_marker) {
              open_text_style (style->marker, is_embedded_marker);
            } else {
              close_text_style (is_embedded_marker);
            }
            break;
          }
          case stylesv2::Type::note_paragraph:
          {
            // This should start a new paragraph, but it is implemented like a character style.
            if (is_opening_marker) {
              open_text_style (style->marker, is_embedded_marker);
            } else {
              close_text_style (is_embedded_marker);
            }
            break;
          }
          case stylesv2::Type::crossreference_wrapper:
          {
            close_text_style (false);
            if (is_opening_marker) {
              const std::string caller = m_note_citations.get (style->marker, "+");
              add_note (caller, marker);
            } else {
              close_current_note ();
            }
            break;
          }
          case stylesv2::Type::crossreference_standard_content:
          case stylesv2::Type::crossreference_content:
          case stylesv2::Type::crossreference_content_with_endmarker:
          {
            if (is_opening_marker) {
              open_text_style (style->marker, is_embedded_marker);
              // Deal in a special way with possible word-level attributes.
              // Note open | Extract word-level attributes
              //  yes      |  no
              //  no       |  yes
              //-----------------
              // In other words, if a note is open, which is the normal case,
              // then don't extract the word-level attributes, but output them as they are.
              // But if a note is not open, it is assumed to be a link reference in the main text body.
              // In such a case, extract the word-level attributes as usual.
              if (!m_note_opened)
                extract_word_level_attributes();
            } else {
              close_text_style (is_embedded_marker);
            }
            break;
          }
          case stylesv2::Type::character:
          {
            if (is_opening_marker) {
              // Be sure the road ahead is clear.
              if (road_is_clear ()) {
                open_text_style (style->marker, is_embedded_marker);
                extract_word_level_attributes();
              } else {
                add_text (filter::usfm::get_opening_usfm (marker));
              }
            } else {
              close_text_style (is_embedded_marker);
            }
            break;
          }
          case stylesv2::Type::page_break:
          {
            close_text_style (false);
            output_as_is (marker, is_opening_marker);
            break;
          }
          case stylesv2::Type::figure:
          {
            close_text_style (false);
            output_as_is (marker, is_opening_marker);
            break;
          }
          case stylesv2::Type::word_list:
          {
            if (is_opening_marker) {
              open_text_style (marker, false);
              extract_word_level_attributes();
            } else {
              close_text_style (false);
            }
            break;
          }
          case stylesv2::Type::sidebar_begin:
          case stylesv2::Type::sidebar_end:
          {
            // Output as plain text.
            close_text_style (false);
            output_as_is (marker, is_opening_marker);
            break;
          }
          case stylesv2::Type::peripheral:
          {
            close_text_style (false);
            output_as_is (marker, is_opening_marker);
            break;
          }
          case stylesv2::Type::milestone:
          {
            if (is_opening_marker) {
              // Check that the road ahead is clear.
              // Check whether extracting the attributes works.
              // If anything fails, the text gets added as-is.
              if (road_is_clear () && extract_milestone_attributes()) {
                const std::string quill_style = quill::hyphen_to_underscore (style->marker);
                constexpr const bool embedded {false};
                open_text_style (quill_style, embedded);
                if (!m_preview)
                  add_text(quill::milestone_emoji);
                close_text_style(embedded);
              } else {
                add_text (filter::usfm::get_opening_usfm (marker));
              }
            }
            break;
          }
          case stylesv2::Type::stopping_boundary:
          default:
            break;
        }
      }
      else {
        // This is a marker unknown in the stylesheet.
        close_text_style (false);
        output_as_is (marker, is_opening_marker);
      }
    }
    else if (current_item == R"(\)") {
      // Just a backslash.
      // In the editor, output this in mono font.
      // See https://github.com/bibledit/cloud/issues/1007
      close_text_style (false);
      output_as_is (std::string(), true);
    }
    else {
      // Here is no marker. Treat it as text.
      if (m_note_opened) {
        add_note_text (current_item);
      } else {
        add_text (current_item);
      }
    }
  }
}


void Editor_Usfm2Html::output_as_is (const std::string& marker, const bool is_opening_marker)
{
  // Output the marker in monospace font.
  if (is_opening_marker) {
    // Add opening marker as it is.
    close_paragraph ();
    new_paragraph ("mono");
    add_text (filter::usfm::get_opening_usfm (marker));
  } else {
    // Add closing marker to existing paragraph.
    add_text (filter::usfm::get_closing_usfm (marker));
  }
}


void Editor_Usfm2Html::new_paragraph (std::string style)
{
  // Handle new paragraph.
  m_current_p_node = m_body_node.append_child ("p");
  m_current_p_open = true;
  if (!style.empty()) {
    std::string quill_style (style);
    quill_style.insert (0, quill::class_prefix_block);
    m_current_p_node.append_attribute ("class") = quill_style.c_str();
  }
  m_current_paragraph_style = style;
  m_current_paragraph_content.clear();
  // A Quill-based editor assigns a length of one to a new line.
  // Skip the first line.
  if (m_first_line_done)
    m_text_tength++;
  m_first_line_done = true;
}


void Editor_Usfm2Html::close_paragraph ()
{
  // Deal with a blank line.
  // If the paragraph is empty, add a <br> to it.
  // This is how the Quill editor naturally represents a new empty line.
  // This makes that empty paragraph to be displayed properly in the Quill editor.
  // This <br> is also needed for live editor updates.
  if (m_current_p_open) {
    if (m_current_paragraph_content.empty()) {
      m_current_p_node.append_child("br");
    }
  }
}


// This opens a text style.
// $style: the array containing the style variables.
// $embed: boolean: Whether to open embedded / nested style.
void Editor_Usfm2Html::open_text_style (const std::string& marker, const bool embed)
{
  if (m_note_opened) {
    if (!embed)
      m_current_note_text_styles.clear();
    m_current_note_text_styles.push_back (marker);
  } else {
    if (!embed)
      m_current_text_styles.clear();
    m_current_text_styles.push_back (marker);
  }
}


// This closes any open text style.
// $embed: boolean: Whether to close embedded character style.
void Editor_Usfm2Html::close_text_style (const bool embed)
{
  if (m_note_opened) {
    if (!m_current_note_text_styles.empty ())
      m_current_note_text_styles.pop_back ();
    if (!embed)
      m_current_note_text_styles.clear ();
  } else {
    if (!m_current_text_styles.empty())
      m_current_text_styles.pop_back();
    if (!embed)
      m_current_text_styles.clear ();
  }
}


// This function adds text to the current paragraph.
// $text: The text to add.
void Editor_Usfm2Html::add_text (const std::string& text)
{
  if (!text.empty()) {
    if (!m_current_p_open) {
      new_paragraph ();
    }
    pugi::xml_node span_dom_element = m_current_p_node.append_child ("span");
    span_dom_element.text ().set (text.c_str());

    const auto assemble_text_style = [this]() -> std::string {
      std::string textstyle {};
      const auto add_style = [&textstyle](const auto& fragment) {
        if (!textstyle.empty ()) {
          // The Quill library is fussy about class names.
          // It accepts class="i-add" but not class="i-add-nd". It fails on that second hyphen.
          // It also does not accept an underscore as part of the class name.
          // That causes the whole class to be removed.
          // If a class is given as "i-add i-nd", then Quill removes the second class and remains with the first.
          // Right now the way to deal with a class with two styles is like this "i-add0nd".
          // It has one hyphen. And a "0" to separate the two styles.
          textstyle.append ("0");
        }
        textstyle.append (fragment);
      };
      if (!m_current_text_styles.empty ()) {
        // Take character style(s) as specified in this object.
        for (const auto& style : m_current_text_styles) {
          add_style(style);
        }
      }
      if (!textstyle.empty())
        textstyle.insert (0, quill::class_prefix_inline);
      // If any word-level attributes were extracted, store them here, and empty the container again.
      // Initially the idea was to have a "data-*" attribute and add this to the current <span> element.
      // This works in html, but fails in the Quill editor.
      // When the Quill editor loads html with "data-*" attributes, it strips and drops them.
      // The word-level attributes are now stored as-is, i.e. in raw format,
      // at the bottom of the html document,
      if (m_pending_word_level_attributes) {
        const std::string id = std::string(quill::word_level_attribute_class_prefix) + std::to_string(get_word_level_attributes_id(true));
        add_style(id);
        add_word_level_attributes(id);
        m_pending_word_level_attributes.reset();
      }
      // Same as the above for milestone attributes.
      if (m_pending_milestone_attributes) {
        const std::string id = std::string(quill::milestone_attribute_class_prefix) + std::to_string(get_milestone_attributes_id(true));
        add_style(id);
        add_milestone_attributes(id);
        m_pending_milestone_attributes.reset();
      }
      // Done.
      return textstyle;
    };
    const std::string textstyle = assemble_text_style();
    if (!textstyle.empty())
      span_dom_element.append_attribute ("class") = textstyle.c_str();
    m_current_paragraph_content.append (text);
  }
  m_text_tength += filter::strings::unicode_string_length (text);
}


// This function adds a note to the current paragraph.
// $citation: The text of the note citation.
// $style: Style name for the paragraph in the note body.
void Editor_Usfm2Html::add_note (const std::string& citation, const std::string& style)
{
  // Be sure that the road ahead is clear.
  if (!road_is_clear ()) {
    add_text (filter::usfm::get_opening_usfm (style));
    return;
  }
  
  // Ensure that a paragraph is open, so that the note can be added to it.
  if (!m_current_p_open) {
    new_paragraph ();
  }
  
  m_note_count++;
  m_note_opened = true;
  
  // Add the link with all relevant data for the note citation.
  add_note_link (m_current_p_node, m_note_count, "call", citation);
  
  // Open a paragraph element for the note body.
  m_note_p_node = m_notes_node.append_child ("p");
  m_note_p_open = true;
  std::string class_value (style);
  class_value.insert (0, quill::class_prefix_block);
  m_note_p_node.append_attribute ("class") = class_value.c_str();
  
  close_text_style (false);
  
  // Add the link with all relevant data for the note body.
  add_note_link (m_note_p_node, m_note_count, "body", citation);
  
  // Add a space.
  add_note_text (" ");
  
  // Update the text length of the text body, excluding the note.
  m_text_tength += filter::strings::unicode_string_length (citation);
}


// This function adds text to the current footnote.
// $text: The text to add.
void Editor_Usfm2Html::add_note_text (const std::string& text)
{
  if (text.empty ())
    return;
  if (!m_note_p_open) {
    add_note ("?", std::string());
  }
  pugi::xml_node span_dom_element = m_note_p_node.append_child ("span");
  span_dom_element.text ().set (text.c_str());
  if (!m_current_note_text_styles.empty()) {
    // Take character style(s) as specified in this object.
    std::string classs;
    classs = filter::strings::implode (m_current_note_text_styles, "0");
    classs.insert (0, quill::class_prefix_inline);
    span_dom_element.append_attribute ("class") = classs.c_str();
  }
}


// This function closes the current footnote.
void Editor_Usfm2Html::close_current_note ()
{
  // If a note was opened, close that, else close the standard text.
  close_text_style (false);
  m_note_p_open = false;
  m_note_opened = false;
}


// This adds a link as a mechanism to connect body text with a note body.
// $domNode: The DOM node where to add the link to.
// $identifier: The link's identifier.
// $style: A style for the note citation, and one for the note body.
// $text: The link's text.
// It also deals with a Quill-based editor, in a slightly different way.
void Editor_Usfm2Html::add_note_link (pugi::xml_node& dom_node, const int identifier,
                                      const std::string& style, const std::string& text)
{
  pugi::xml_node a_dom_element = dom_node.append_child("span");
  const std::string class_value = "i-note" + style + std::to_string(identifier);
  a_dom_element.append_attribute("class") = class_value.c_str();
  a_dom_element.text().set(text.c_str());
}


// Returns true if the road ahead is clear for the current marker.
bool Editor_Usfm2Html::road_is_clear ()
{
  // Call a unit testable function to do the work.
  return ::road_is_clear (m_markers_and_text, m_markers_and_text_pointer, m_stylesheet);
}


void Editor_Usfm2Html::set_preview ()
{
  m_preview = true;
}


int Editor_Usfm2Html::get_word_level_attributes_id(const bool next)
{
  if (next) {
    m_word_level_attributes_id++;
    // The class name for the Quill editor has defined the "0" as class separator.
    // So make sure that the word-level attribute ID never contains any "0".
    while (std::to_string(m_word_level_attributes_id).find("0") != std::string::npos) {
      m_word_level_attributes_id++;
    }
  }
  return m_word_level_attributes_id;
}


void Editor_Usfm2Html::extract_word_level_attributes()
{
  // The current function is expected to be called on an opening marker
  // that also may contain word-level attributes.
  // See USFM 3.x for details.

  // Check the text following this markup whether it contains word-level attribute(s).
  std::string possible_markup = filter::usfm::peek_text_following_marker (m_markers_and_text, m_markers_and_text_pointer);
  
  // The minimum length of the markup is:
  // 1. The canonical word: At least 1 character.
  // 2. The vertical bar: One character.
  // 3. The attribute: At least character for the default attribute.
  // If the markup is too short then bail out.
  if (possible_markup.size() < 3)
    return;
  
  // Look for the vertical bar. If it's not there, bail out.
  const size_t bar_position = possible_markup.find("|");
  if (bar_position == std::string::npos)
    return;
  
  // Remove the fragment and store the remainder back into the object.
  m_markers_and_text [m_markers_and_text_pointer + 1] = possible_markup.substr(0, bar_position);
  
  // Remain with the fragment with the word-level attributes.
  // Store the word-level attributes as-is, as pending item, ready for processing.
  std::string attribute_string = possible_markup.substr(bar_position + 1);
  m_pending_word_level_attributes = std::move(attribute_string);
}


void Editor_Usfm2Html::add_word_level_attributes(const std::string id)
{
  // If there's no attributes, bail out.
  if (!m_pending_word_level_attributes)
    return;
  
  // Open a paragraph element for the word-level attributes body.
  pugi::xml_node p_node = m_word_level_attributes_node.append_child ("p");

  // Set the correct class in the paragraph. This links it to the correct text fragment in the text body.
  const std::string class_value {quill::class_prefix_block + id};
  p_node.append_attribute ("class") = class_value.c_str();

  // Add the word-level attributes as text.
  p_node.text().set (m_pending_word_level_attributes.value().c_str());
}


int Editor_Usfm2Html::get_milestone_attributes_id(const bool next)
{
  if (next) {
    m_milestone_attributes_id++;
    // The class name for the Quill editor has defined the "0" as class separator.
    // So make sure that the milestone attribute ID never contains any "0".
    while (std::to_string(m_milestone_attributes_id).find("0") != std::string::npos) {
      m_milestone_attributes_id++;
    }
  }
  return m_milestone_attributes_id;
}


bool Editor_Usfm2Html::extract_milestone_attributes()	
{
  // The current function is expected to be called on an opening marker of a milestone.
  // See USFM 3 and later for details.

  // Example milestones:
  //   \qt-s |sid="sid" who="who"\*
  //   \qt-e\*

  // Variable indicating how much to increase the pointer to markers and text.
  unsigned pointer_increase {0};
  
  // Check on the location of the milestone end marker.
  // The correct end marker is "\*".
  // It may be on the first or the second position after the current positiion.
  // Store that location in the pointer increase factor.
  for (const auto increase : {static_cast<unsigned>(1), static_cast<unsigned>(2)}) {
    const std::string endmarker = filter::usfm::peek_text_following_marker (m_markers_and_text, m_markers_and_text_pointer + increase - 1);
    if (endmarker == filter::usfm::get_closing_usfm (std::string()))
      pointer_increase = increase;
  }
  
  // If the pointer increase is still 0, it means no correctly formatted milestone was found.
  if (!pointer_increase)
    return false;

  // If the pointer increase is 1, it means there's the endmarker only, without attributes.
  
  // If the pointer increase is 2, it means there's attributes. Handle those.
  if (pointer_increase == 2) {
    std::string attributes = filter::strings::trim(filter::usfm::peek_text_following_marker (m_markers_and_text, m_markers_and_text_pointer));
    // Check whether the first character in the milestone data is the vertical bar "|".
    // If so, remove it.
    if (attributes.at(0) == '|')
      attributes.erase(attributes.cbegin());
    // Remain with the fragment with the milestone attributes.
    // Store the milestone attributes as-is, as pending item, ready for processing.
    // Don't store this in preview mode.
    if (!m_preview)
      m_pending_milestone_attributes = std::move(attributes);
  }

  // Everything has passed: Increase the pointer to the USFM input data.
  m_markers_and_text_pointer += pointer_increase;

  // Everything is OK:
  return true;
}


void Editor_Usfm2Html::add_milestone_attributes(const std::string id)
{
  // If there's no attributes, bail out.
  if (!m_pending_milestone_attributes)
    return;
  
  // Open a paragraph element for the add_milestone attributes body.
  pugi::xml_node p_node = m_milestone_attributes_node.append_child ("p");
  
  // Set the correct class in the paragraph. This links it to the correct text fragment in the text body.
  const std::string class_value {quill::class_prefix_block + id};
  p_node.append_attribute ("class") = class_value.c_str();
  
  // Add the word-level attributes as text.
  p_node.text().set (m_pending_milestone_attributes.value().c_str());
}


// Returns true if the road ahead is clear for the current marker.
bool road_is_clear(const std::vector<std::string>& markers_and_text,
                   const unsigned int markers_and_text_pointer,
                   const std::string& stylesheet)
{
  // Section 1: Functions to find marker properties.
  
  // Function to determine whether it is starting a paragraph.
  const auto starts_paragraph = [](const stylesv2::Style* style) {
    if (style) {
      if (style->type == stylesv2::Type::title)
        return true;
      if (style->type == stylesv2::Type::heading)
        return true;
      if (style->type == stylesv2::Type::paragraph)
        return true;
    }
    return false;
  };
  
  // Function to determine whether it is inline text.
  const auto is_inline_text = [](const stylesv2::Style* style) {
    if (style) {
      if (style->type == stylesv2::Type::character)
        return true;
      if (style->type == stylesv2::Type::milestone)
        return true;
    }
    return false;
  };
  
  // Function to determine whether it is a verse number markup.
  const auto is_verse_number = [](const stylesv2::Style* style) {
    if (style)
      if (style->type == stylesv2::Type::verse)
        return true;
    return false;
  };

  // Function to determine whether the type is a footnote / endnote.
  const auto is_note_type = [](const stylesv2::Style* style) {
    if (style) {
      if (style->type == stylesv2::Type::footnote_wrapper)
        return true;
      if (style->type == stylesv2::Type::endnote_wrapper)
        return true;
      if (style->type == stylesv2::Type::note_standard_content)
        return true;
      if (style->type == stylesv2::Type::note_content)
        return true;
      if (style->type == stylesv2::Type::note_content_with_endmarker)
        return true;
      if (style->type == stylesv2::Type::note_paragraph)
        return true;
    }
    return false;
  };
  
  // Function to determine whether the style is a footnote / endnote wrapper.
  const auto is_footnote_endnote_wrapper = [](const stylesv2::Style* style) {
    if (style) {
      if (style->type == stylesv2::Type::footnote_wrapper)
        return true;
      if (style->type == stylesv2::Type::endnote_wrapper)
        return true;
    }
    return false;
  };

  // Function to determine whether the type is a footnote / endnote.
  const auto is_xref_type = [](const stylesv2::Style* style) {
    if (style) {
      if (style->type == stylesv2::Type::crossreference_wrapper)
        return true;
      if (style->type == stylesv2::Type::crossreference_standard_content)
        return true;
      if (style->type == stylesv2::Type::crossreference_content)
        return true;
      if (style->type == stylesv2::Type::crossreference_content_with_endmarker)
        return true;
    }
    return false;
  };
  
  // Function to determine whether the subtype is a footnote / endnote.
  const auto is_xref_subtype = [](const stylesv2::Style* style) {
    if (style) {
      if (style->type == stylesv2::Type::crossreference_wrapper)
        return true;
    }
    return false;
  };

  // Section 2: Implementation code.

  // If the input is not a marker, the road ahead is clear.
  const std::string input_item = markers_and_text[markers_and_text_pointer];
  if (!filter::usfm::is_usfm_marker (input_item))
    return true;

  // The marker. If not in stylesheet, the road is clear.
  const std::string input_marker {filter::usfm::get_marker (input_item)};
  const stylesv2::Style* input_style {database::styles::get_marker_data (stylesheet, input_marker)};
  if (!input_style)
    return true;

  // Determine the input markup properties.
  const bool input_opener {filter::usfm::is_opening_marker (input_item)};
  const bool input_embedded {filter::usfm::is_embedded_marker (input_item)};

  // If the imput markup is embedded inline text, the road ahead is clear.
  if (is_inline_text(input_style))
    if (input_embedded)
      return true;

  // If the input to check the road for, is a milestone, the road ahead is declared clear.
  if (input_style->type == stylesv2::Type::milestone)
    return true;
  
  // Determine the road ahead that follows the input markup.
  bool verse_boundary_reached {false};
  for (size_t pointer {markers_and_text_pointer + 1}; pointer < markers_and_text.size(); pointer++) {
    
    // If ready, skip the rest of the markup.
    if (verse_boundary_reached)
      continue;
    
    // Gather data about the current marker.
    const std::string& current_item = markers_and_text.at(pointer);
    if (filter::usfm::is_usfm_marker (current_item))
    {
      const std::string marker = filter::usfm::get_marker (current_item);
      const stylesv2::Style* style {database::styles::get_marker_data (stylesheet, marker)};
      if (style)
      {
        const bool opener {filter::usfm::is_opening_marker (current_item)};
        const bool embedded {filter::usfm::is_embedded_marker (current_item)};
        
        // Take the next verse marker in account but don't go beyond it.
        if (is_verse_number(style))
            verse_boundary_reached = true;
        
        // The input is a note opener.
        if (is_note_type(input_style)) {
          if (input_opener) {
            if (is_footnote_endnote_wrapper(input_style)) {
              // Encounters note closer: road is clear.
              // Encounters another note opener: blocker.
              if (is_note_type(style)) {
                if (is_footnote_endnote_wrapper(style)) {
                  if (opener)
                    return false;
                  else
                    return true;
                }
              }
              // Encounters a verse: blocker.
              if (is_verse_number(style))
                return false;
              // Encounters cross reference opener or closer: blocker.
              // Other \x.. markup is allowed.
              if (is_xref_type(style)) {
                if (is_xref_subtype(style)) {
                  return false;
                }
              }
            }
          }
        }

        // The input is a cross reference opener.
        if (is_xref_type(input_style)) {
          if (input_opener) {
            if (is_xref_subtype(input_style)) {
              // Encounters xref closer: road is clear.
              // Encounters another xref opener: blocker.
              if (is_xref_type(style)) {
                if (is_xref_subtype(style)) {
                  if (opener)
                    return false;
                  else
                    return true;
                }
              }
              // Encounters a verse: blocker.
              if (is_verse_number(style))
                return false;
              // Encounters foot- or endnote opener: blocker.
              // Other \f.. markup is allowed.
              if (is_note_type(style)) {
                if (is_footnote_endnote_wrapper(style)) {
                  return false;
                }
              }
            }
          }
        }

        // The input to check the road ahead for is an inline text opener, non-embedded, like "\add ".
        if (is_inline_text(input_style)) {
          if (input_opener && !input_embedded) {
            if (is_inline_text(style)) {
              if (embedded) {
                // An embedded inline marker is OK: Road ahead is clear.
                return true;
              }
              // If it encounters another non-embedded inline opening marker, that's a blocker.
              if (opener)
                return false;
              // If it finds a matching closing marker: OK.
              if (input_marker == marker) {
                if (!opener)
                  return true;
              }
            }
            // The inline text opener encounters a verse: blocker.
            if (is_verse_number(style))
              return false;
            // The inline text opener encounters a paragraph: blocker.
            if (starts_paragraph(style))
              return false;
          }
        }
        
      }
    }
  }
  
  // If it didn't find the verse markup, it follows that the end of the chapter was reached.
  // That is a blocker.
  if (!verse_boundary_reached)
    return false;

  // No blockers found: The road is clear.
  return true;
}