Skip to content

Commit

Permalink
Fixing of markdown processing
Browse files Browse the repository at this point in the history 
- processing our special markdown notation [[ ]] still had numerous issues, like parsed inside code blocks
- some markdown links still not skipped
- processing order of [[ ]] [[ | ]] was bogus
- added handling of a special form of our notation [[title|-]] that can protect the given title from further autolink/tooltip processing
- prepared possible processing of our markdown notation in H2-H6 headings
Signed-off-by: Hofi <[email protected]>

Signed-off-by: Hofi <[email protected]>
  • Loading branch information
@HofiOne
HofiOne committed May 31, 2024
1 parent c724c63 commit ad8eea8
Showing 1 changed file with 73 additions and 23 deletions.
96 changes: 73 additions & 23 deletions _plugins/generate_tooltips.rb
View file
Original file line number Diff line number Diff line change
Expand Up @@ -47,39 +47,60 @@ def prefixed_url(url, base_url)
return url
end

def is_modifiable_markdown_part?(part)
# TODO: Allowing our custom markdown notation in the headings require removing the custom notation during link generation
return part.start_with?('[[') # || part.start_with?('#')
end

def make_tooltip(page, page_links, id, url, match)
match_parts = match.split(/\|/)

# If the text has an '|' it means it comes from our special autolink/tooltip [[text|id]] markdown block
# We have to reparse it a bit and get the id we must use
if match_parts.length > 1
#puts "match_parts: #{match_parts}"
match = match_parts[0]
title = match_parts[0]
if title.length <= 0
puts "Error: Empty title in matching part: '#{match}' -> #{match_parts}"
# nil means, show the original markdown part, instead of a half rendered one
return nil
end
id = match_parts[1]
# This is a special use case [[title|-]] that protects the given title from further processing
if id == '-'
# Just use the original title text
return title
end
link_data = page_links[id]
if link_data != nil
url = link_data["url"]
url = prefixed_url(url, page.site.config["baseurl"])
else
puts "Error: Unknown ID in matching part: #{match_parts}"
return match
puts "Error: Unknown ID in matching part: '#{match}' -> #{match_parts}"
# nil means, show the original markdown part, instead of a half rendered one
return nil
end
else
title = match
end

if id == nil or id.length <= 0
puts "Error: Empty ID in matching part: #{match}"
return match
puts "Error: Empty ID in matching part: '#{match}' -> #{match_parts}"
# nil means, show the original markdown part, instead of a half rendered one
return nil
end
if url == nil or url.length <= 0
puts "Error: Empty URL for ID: #{id} in matching part: #{match}"
return match
puts "Error: Empty URL for ID: '#{id}' in matching part: '#{match}' -> #{match_parts}"
# nil means, show the original markdown part, instead of a half rendered one
return nil
end

# NOTE: Now we treat every link that has protocol prefix part as an external one
# that allows usage of direct links anywhere if needed (not recommended, plz use external_links.yml instead)
# but, at the same time requires e.g. all the really external links to be fully qualified (even in external_links.yml as well)
external_url = is_prefixed_url?(url)
match = save_from_markdownify(match)
replacement_text = '<a href="' + url + '" class="nav-link content-tooltip"' + (external_url ? ' target="_blank"' : '') + '>' + match + '</a>'
title = save_from_markdownify(title)
replacement_text = '<a href="' + url + '" class="nav-link content-tooltip"' + (external_url ? ' target="_blank"' : '') + '>' + title + '</a>'
# puts "replacement_text: " + replacement_text

return replacement_text
Expand All @@ -91,11 +112,15 @@ def process_markdown_part(page, markdown_part, page_links, full_pattern, id, url
left_separator = $1
matched_text = $2
right_separator = $3
#puts "\nmatch: #{match}\nleft_separator: #{left_separator}\nmatched_text: #{matched_text}\nright_separator: #{right_separator}"
# puts "\nmatch: #{match}\nleft_separator: #{left_separator}\nmatched_text: #{matched_text}\nright_separator: #{right_separator}"

replacement_text = make_tooltip(page, page_links, id, url, matched_text)
if add_separator
replacement_text = left_separator + replacement_text + right_separator
if replacement_text != nil
if add_separator
replacement_text = left_separator + replacement_text + right_separator
end
else
replacement_text = markdown_part.gsub(/\|/, "\\\|")
end
replacement_text
end
Expand All @@ -110,10 +135,26 @@ def process_markdown_parts(page, markdown)
# Regular expression pattern to match special Markdown blocks
# Unlike the others this needs grouping as we use do |match| for enumeration
# NOTE: Use multi line matching partially as e.g. code blocks can span to multiple lines
special_markdown_blocks_pattern = /((?m:````.*?````|```.*?```|``.*?``|`.*?`)|\[\[.*?\]\]|\[.*?\]\(.*?\)\{\:.*?\}|\[.*?\]\(.*?\)|\[.*?\]\{.*?\}|^#+\s.*?$)/
markdown_blocks_pattern = /((?m:````.*?````|```.*?```|``.*?``|`.*?`)|\[\[[^\]^\[]*?\]\]|\[[^\]^\[]*?\]\(.*?\)\{\:.*?\}|\[[^\]^\[]*?\]\(.*?\)|\[[^\]^\[]*?\]:.*?$|\[[^\]^\[]*?\]\s*\[.*?\]|^#+\s.*?$)/
# TODO: Always sync the bellow with the one-liner version for readability
# FIXME: Check why the /x version bellow is not working the same way
# markdown_blocks_pattern = /( # Either Code blocks
# (?m: # Even Multiline ones
# ````.*?```` | # Code block with 4 backticks
# ```.*?``` | # Code block with 3 backticks
# ``.*?`` | # Code block with 2 backticks
# `.*?` # Inline code with 1 backtick
# ) | #
# \[\[[^\]^\[]*?\]\] | # or Our special, custom markdown notation
# \[[^\]^\[]*?\]\(.*?\)\{\:.*?\} | # or Link with attribute
# \[[^\]^\[]*?\]\(.*?\) | # Link without attribute
# \[[^\]^\[]*?\]:.*?$ | # Link reference label declaration
# \[[^\]^\[]*?\]\s*\[.*?\] | # Link using reference label
# ^#+\s.*?$ # or Headers
# )/x

# Split the content by special Markdown blocks
markdown_parts = markdown.split(special_markdown_blocks_pattern)
markdown_parts = markdown.split(markdown_blocks_pattern)
#puts markdown_parts
markdown_parts.each_with_index do |markdown_part, markdown_index|
# puts "---------------\nmarkdown_index: " + markdown_index.to_s + "\n" + (markdown_index.even? ? "NOT " : "") + "markdown_part: " + markdown_part
Expand All @@ -134,27 +175,36 @@ def process_markdown_parts(page, markdown)
#puts "searching for #{title} with pattern #{pattern}"

if markdown_index.even?
# Content outside of special Markdown blocks, aka. pure text (NOTE: Also excludes the reqursively self added <a ...>title</a> tooltips/links)
# Content outside of Markdown blocks, aka. pure text

# Search for known link titles
# NOTE: Using multi line matching here will not help either if the pattern itself is in the middle broken/spaned to multiple lines, so
# using whitespace replacements now inside the patter to handle this, see above!
# NOTE: Also excludes the reqursively self added <a ...>title</a> tooltips/links
full_pattern = /(^|[\s.,;:&'"(])(#{pattern})([\s.,;:&'")]|\z)(?![^<]*?<\/a>)/
markdown_part = process_markdown_part(page, markdown_part, page_links, full_pattern, id, url, true)
else
# Content inside of special Markdown blocks
# Content inside of Markdown blocks

# Handle own auto\tooltip links [[title]], but NOT [[title|id]], see bellow why
full_pattern = /(\[\[)(#{pattern})(\]\])/
markdown_part = process_markdown_part(page, markdown_part, page_links, full_pattern, id, url, false)
# Handle our special markdown notation autolink/tooltip links [[title]], but NOT [[title|id]], see bellow why
if is_modifiable_markdown_part?(markdown_part)
full_pattern = /(\[\[)(#{pattern})(\]\])/
markdown_part = process_markdown_part(page, markdown_part, page_links, full_pattern, id, url, false)
end
end
end

# Handle our special markdown notation autolink/tooltip links [[title|id]]
# This must be a separate run, as independent from the given title, if ID is presented it will always override the title, and the title exclusion as well
if markdown_index.odd?
# Handle own auto\tooltip links [[title|id]]
# This must be a separate run, as independent from the given title, if ID is presented it will always override title, and title exclusion as well
full_pattern = /(\[\[)(.+\|.+)(\]\])/
markdown_part = process_markdown_part(page, markdown_part, page_links, full_pattern, nil, nil, false)
# Content inside of Markdown blocks

if is_modifiable_markdown_part?(markdown_part)
# puts "\nmarkdown_index: " + markdown_index.to_s + "\n" + (markdown_index.even? ? "NOT " : "") + "markdown_part: " + markdown_part
# NOTE: The differences in the patter is intentional, allowing empty part on both sides of | allows the same flow inside process_markdown_part
full_pattern = /(\[\[)(.*?\|.*?)(\]\])/
markdown_part = process_markdown_part(page, markdown_part, page_links, full_pattern, nil, nil, false)
end
end

#puts "new markdown_part: " + markdown_part
Expand Down

0 comments on commit ad8eea8

Please sign in to comment.