lib/jekyll-paginate-v2/generator/paginationModel.rb
module Jekyll
module PaginateV2::Generator
#
# The main model for the pagination, handles the orchestration of the pagination and calling all the necessary bits and bobs needed :)
#
class PaginationModel
@debug = false # is debug output enabled?
@logging_lambda = nil # The lambda to use for logging
@page_add_lambda = nil # The lambda used to create pages and add them to the site
@page_remove_lambda = nil # Lambda to remove a page from the site.pages collection
@collection_by_name_lambda = nil # Lambda to get all documents/posts in a particular collection (by name)
# ctor
def initialize(logging_lambda, page_add_lambda, page_remove_lambda, collection_by_name_lambda)
@logging_lambda = logging_lambda
@page_add_lambda = page_add_lambda
@page_remove_lambda = page_remove_lambda
@collection_by_name_lambda = collection_by_name_lambda
end
def run(default_config, site_pages, site_title)
# By default if pagination is enabled we attempt to find all index.html pages in the site
templates = self.discover_paginate_templates(site_pages)
if( templates.size.to_i <= 0 )
@logging_lambda.call "Is enabled, but I couldn't find any pagination page. Skipping pagination. "+
"Pages must have 'pagination: enabled: true' in their front-matter for pagination to work.", "warn"
return
end
# Now for each template page generate the paginator for it
templates.each do |template|
# All pages that should be paginated need to include the pagination config element
if template.data['pagination'].is_a?(Hash)
template_config = Jekyll::Utils.deep_merge_hashes(default_config, template.data['pagination'] || {})
# Handling deprecation of configuration values
self._fix_deprecated_config_features(template_config)
@debug = template_config['debug'] # Is debugging enabled on the page level
self._debug_print_config_info(template_config, template.path)
# Only paginate the template if it is explicitly enabled
# requiring this makes the logic simpler as I don't need to determine which index pages
# were generated automatically and which weren't
if( template_config['enabled'] )
if !@debug
@logging_lambda.call "found page: "+template.path, 'debug'
end
# Request all documents in all collections that the user has requested
all_posts = self.get_docs_in_collections(template_config['collection'])
# Create the necessary indexes for the posts
all_categories = PaginationIndexer.index_posts_by(all_posts, 'categories')
all_categories['posts'] = all_posts; # Populate a category for all posts (this is here for backward compatibility, do not use this as it will be decommissioned 2018-01-01)
# (this is a default and must not be used in the category system)
all_tags = PaginationIndexer.index_posts_by(all_posts, 'tags')
all_locales = PaginationIndexer.index_posts_by(all_posts, 'locale')
# TODO: NOTE!!! This whole request for posts and indexing results could be cached to improve performance, leaving like this for now during testing
# Now construct the pagination data for this template page
self.paginate(template, template_config, site_title, all_posts, all_tags, all_categories, all_locales)
end
end
end #for
# Return the total number of templates found
return templates.size.to_i
end # function run
#
# This function is here to retain the old compatability logic with the jekyll-paginate gem
# no changes should be made to this function and it should be retired and deleted after 2018-01-01
# (REMOVE AFTER 2018-01-01)
#
def run_compatability(legacy_config, site_pages, site_title, all_posts)
# Decomissioning error
if( date = Date.strptime("20180101","%Y%m%d") <= Date.today )
raise ArgumentError.new("Legacy jekyll-paginate configuration compatibility mode has expired. Please upgrade to jekyll-paginate-v2 configuration.")
end
# Two month warning or general notification
if( date = Date.strptime("20171101","%Y%m%d") <= Date.today )
@logging_lambda.call "Legacy pagination logic will stop working on Jan 1st 2018, update your configs before that time.", "warn"
else
@logging_lambda.call "Detected legacy jekyll-paginate logic running. "+
"Please update your configs to use the jekyll-paginate-v2 logic. This compatibility function "+
"will stop working after Jan 1st 2018 and your site build will throw an error.", "warn"
end
if template = CompatibilityUtils.template_page(site_pages, legacy_config['legacy_source'], legacy_config['permalink'])
CompatibilityUtils.paginate(legacy_config, all_posts, template, @page_add_lambda)
else
@logging_lambda.call "Legacy pagination is enabled, but I couldn't find " +
"an index.html page to use as the pagination page. Skipping pagination.", "warn"
end
end # function run_compatability (REMOVE AFTER 2018-01-01)
# Returns the combination of all documents in the collections that are specified
# raw_collection_names can either be a list of collections separated by a ',' or ' ' or a single string
def get_docs_in_collections(raw_collection_names)
if raw_collection_names.is_a?(String)
collection_names = raw_collection_names.split(/;|,|\s/)
else
collection_names = raw_collection_names
end
docs = []
# Now for each of the collections get the docs
collection_names.each do |coll_name|
# Request all the documents for the collection in question, and join it with the total collection
docs += @collection_by_name_lambda.call(coll_name.downcase.strip)
end
# Hidden documents should not not be processed anywhere.
docs = docs.reject { |doc| doc['hidden'] }
return docs
end
def _fix_deprecated_config_features(config)
keys_to_delete = []
# As of v1.5.1 the title_suffix is deprecated and 'title' should be used
# but only if title has not been defined already!
if( !config['title_suffix'].nil? )
if( config['title'].nil? )
config['title'] = ":title" + config['title_suffix'].to_s # Migrate the old key to title
end
keys_to_delete << "title_suffix" # Always remove the depricated key if found
end
# Delete the depricated keys
config.delete_if{ |k,| keys_to_delete.include? k }
end
def _debug_print_config_info(config, page_path)
r = 20
f = "Pagination: ".rjust(20)
# Debug print the config
if @debug
puts f + "----------------------------"
puts f + "Page: "+page_path.to_s
puts f + " Active configuration"
puts f + " Enabled: ".ljust(r) + config['enabled'].to_s
puts f + " Items per page: ".ljust(r) + config['per_page'].to_s
puts f + " Permalink: ".ljust(r) + config['permalink'].to_s
puts f + " Title: ".ljust(r) + config['title'].to_s
puts f + " Limit: ".ljust(r) + config['limit'].to_s
puts f + " Sort by: ".ljust(r) + config['sort_field'].to_s
puts f + " Sort reverse: ".ljust(r) + config['sort_reverse'].to_s
puts f + " Active Filters"
puts f + " Collection: ".ljust(r) + config['collection'].to_s
puts f + " Offset: ".ljust(r) + config['offset'].to_s
puts f + " Combine: ".ljust(r) + config['combine'].to_s
puts f + " Category: ".ljust(r) + (config['category'].nil? || config['category'] == "posts" ? "[Not set]" : config['category'].to_s)
puts f + " Tag: ".ljust(r) + (config['tag'].nil? ? "[Not set]" : config['tag'].to_s)
puts f + " Locale: ".ljust(r) + (config['locale'].nil? ? "[Not set]" : config['locale'].to_s)
if config['legacy']
puts f + " Legacy Paginate Code Enabled"
puts f + " Legacy Paginate: ".ljust(r) + config['per_page'].to_s
puts f + " Legacy Source: ".ljust(r) + config['legacy_source'].to_s
puts f + " Legacy Path: ".ljust(r) + config['paginate_path'].to_s
end
end
end
def _debug_print_filtering_info(filter_name, before_count, after_count)
# Debug print the config
if @debug
puts "Pagination: ".rjust(20) + " Filtering by: "+filter_name.to_s.ljust(9) + " " + before_count.to_s.rjust(3) + " => " + after_count.to_s
end
end
#
# Rolls through all the pages passed in and finds all pages that have pagination enabled on them.
# These pages will be used as templates
#
# site_pages - All pages in the site
#
def discover_paginate_templates(site_pages)
candidates = []
site_pages.select do |page|
# If the page has the enabled config set, supports any type of file name html or md
if page.data['pagination'].is_a?(Hash) && page.data['pagination']['enabled']
candidates << page
end
end
return candidates
end # function discover_paginate_templates
# Paginates the blog's posts. Renders the index.html file into paginated
# directories, e.g.: page2/index.html, page3/index.html, etc and adds more
# site-wide data.
#
# site - The Site.
# template - The index.html Page that requires pagination.
# config - The configuration settings that should be used
#
def paginate(template, config, site_title, all_posts, all_tags, all_categories, all_locales)
# By default paginate on all posts in the site
using_posts = all_posts
should_union = config['combine'] == 'union'
# Now start filtering out any posts that the user doesn't want included in the pagination
before = using_posts.size.to_i
using_posts = PaginationIndexer.read_config_value_and_filter_posts(config, 'category', using_posts, all_categories, should_union)
self._debug_print_filtering_info('Category', before, using_posts.size.to_i)
before = using_posts.size.to_i
using_posts = PaginationIndexer.read_config_value_and_filter_posts(config, 'tag', using_posts, all_tags, should_union)
self._debug_print_filtering_info('Tag', before, using_posts.size.to_i)
before = using_posts.size.to_i
using_posts = PaginationIndexer.read_config_value_and_filter_posts(config, 'locale', using_posts, all_locales, should_union)
self._debug_print_filtering_info('Locale', before, using_posts.size.to_i)
# Apply sorting to the posts if configured, any field for the post is available for sorting
if config['sort_field']
sort_field = config['sort_field'].to_s
# There is an issue in Jekyll related to lazy initialized member variables that causes iterators to
# break when accessing an uninitialized value during iteration. This happens for document.rb when the <=> compaison function
# is called (as this function calls the 'date' field which for drafts are not initialized.)
# So to unblock this common issue for the date field I simply iterate once over every document and initialize the .date field explicitly
if @debug
puts "Pagination: ".rjust(20) + "Rolling through the date fields for all documents"
end
using_posts.each do |u_post|
if u_post.respond_to?('date')
tmp_date = u_post.date
if( !tmp_date || tmp_date.nil? )
if @debug
puts "Pagination: ".rjust(20) + "Explicitly assigning date for doc: #{u_post.data['title']} | #{u_post.path}"
end
u_post.date = File.mtime(u_post.path)
end
end
end
using_posts.sort!{ |a,b| Utils.sort_values(Utils.sort_get_post_data(a.data, sort_field), Utils.sort_get_post_data(b.data, sort_field)) }
# Remove the first x entries
offset_post_count = [0, config['offset'].to_i].max
using_posts.pop(offset_post_count)
if config['sort_reverse']
using_posts.reverse!
end
end
# Calculate the max number of pagination-pages based on the configured per page value
total_pages = Utils.calculate_number_of_pages(using_posts, config['per_page'])
# If a upper limit is set on the number of total pagination pages then impose that now
if config['limit'] && config['limit'].to_i > 0 && config['limit'].to_i < total_pages
total_pages = config['limit'].to_i
end
#### BEFORE STARTING REMOVE THE TEMPLATE PAGE FROM THE SITE LIST!
@page_remove_lambda.call( template )
# list of all newly created pages
newpages = []
# Consider the default index page name and extension
indexPageName = config['indexpage'].nil? ? '' : config['indexpage'].split('.')[0]
indexPageExt = config['extension'].nil? ? '' : Utils.ensure_leading_dot(config['extension'])
indexPageWithExt = indexPageName + indexPageExt
# In case there are no (visible) posts, generate the index file anyway
total_pages = 1 if total_pages.zero?
# Now for each pagination page create it and configure the ranges for the collection
# This .pager member is a built in thing in Jekyll and defines the paginator implementation
# Simpy override to use mine
(1..total_pages).each do |cur_page_nr|
# 1. Create the in-memory page
# External Proc call to create the actual page for us (this is passed in when the pagination is run)
newpage = PaginationPage.new( template, cur_page_nr, total_pages, indexPageWithExt )
# 2. Create the url for the in-memory page (calc permalink etc), construct the title, set all page.data values needed
paginated_page_url = config['permalink']
first_index_page_url = ""
if template.data['permalink']
first_index_page_url = Utils.ensure_trailing_slash(template.data['permalink'])
else
first_index_page_url = Utils.ensure_trailing_slash(template.dir)
end
paginated_page_url = File.join(first_index_page_url, paginated_page_url)
# 3. Create the pager logic for this page, pass in the prev and next page numbers, assign pager to in-memory page
newpage.pager = Paginator.new( config['per_page'], first_index_page_url, paginated_page_url, using_posts, cur_page_nr, total_pages, indexPageName, indexPageExt)
# Create the url for the new page, make sure we prepend any permalinks that are defined in the template page before
if newpage.pager.page_path.end_with? '/'
newpage.set_url(File.join(newpage.pager.page_path, indexPageWithExt))
elsif newpage.pager.page_path.end_with? indexPageExt
# Support for direct .html files
newpage.set_url(newpage.pager.page_path)
else
# Support for extensionless permalinks
newpage.set_url(newpage.pager.page_path+indexPageExt)
end
if( template.data['permalink'] )
newpage.data['permalink'] = newpage.pager.page_path
end
# Transfer the title across to the new page
if( !template.data['title'] )
tmp_title = site_title
else
tmp_title = template.data['title']
end
# If the user specified a title suffix to be added then let's add that to all the pages except the first
if( cur_page_nr > 1 && config.has_key?('title') )
newpage.data['title'] = "#{Utils.format_page_title(config['title'], tmp_title, cur_page_nr, total_pages)}"
else
newpage.data['title'] = tmp_title
end
# Signals that this page is automatically generated by the pagination logic
# (we don't do this for the first page as it is there to mask the one we removed)
if cur_page_nr > 1
newpage.data['autogen'] = "jekyll-paginate-v2"
end
# Add the page to the site
@page_add_lambda.call( newpage )
# Store the page in an internal list for later referencing if we need to generate a pagination number path later on
newpages << newpage
end #each.do total_pages
# Now generate the pagination number path, e.g. so that the users can have a prev 1 2 3 4 5 next structure on their page
# simplest is to include all of the links to the pages preceeding the current one
# (e.g for page 1 you get the list 2, 3, 4.... and for page 2 you get the list 3,4,5...)
if( config['trail'] && !config['trail'].nil? && newpages.size.to_i > 1 )
trail_before = [config['trail']['before'].to_i, 0].max
trail_after = [config['trail']['after'].to_i, 0].max
trail_length = trail_before + trail_after + 1
if( trail_before > 0 || trail_after > 0 )
newpages.select do | npage |
idx_start = [ npage.pager.page - trail_before - 1, 0].max # Selecting the beginning of the trail
idx_end = [idx_start + trail_length, newpages.size.to_i].min # Selecting the end of the trail
# Always attempt to maintain the max total of <trail_length> pages in the trail (it will look better if the trail doesn't shrink)
if( idx_end - idx_start < trail_length )
# Attempt to pad the beginning if we have enough pages
idx_start = [idx_start - ( trail_length - (idx_end - idx_start) ), 0].max # Never go beyond the zero index
end
# Convert the newpages array into a two dimensional array that has [index, page_url] as items
#puts( "Trail created for page #{npage.pager.page} (idx_start:#{idx_start} idx_end:#{idx_end})")
npage.pager.page_trail = newpages[idx_start...idx_end].each_with_index.map {|ipage,idx| PageTrail.new(idx_start+idx+1, ipage.pager.page_path, ipage.data['title'])}
#puts( npage.pager.page_trail )
end #newpages.select
end #if trail_before / trail_after
end # if config['trail']
end # function paginate
end # class PaginationV2
end # module PaginateV2
end # module Jekyll