sutty-migration 0.1.2 → 0.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: ffe46cc7d270c7f30d4a505704cb244e507df588bb380ad2c101c3a01675844b
-  data.tar.gz: d63a9dd7fde09627c61f02f7b714e151f5122d5133e26d8c4d3be7917d35501e
+  metadata.gz: 5486653e0e1eb13f5c4c4f85235c875c782fee7be37a1bee9e4cdd84d5879d0a
+  data.tar.gz: '096ab9a992ad5b4cf36bb765a4eb99bd4ac9f2fc35ec25737906eb7c3abc8fdf'
 SHA512:
-  metadata.gz: 9c9278d28ab6d4b862c5cc615941102c5ef8716c30b674093dd31f5a6dce1c337ff6729fab69e6018263a113b3b82f8faadc9b8aad2a9bc38cfd472d082d711c
-  data.tar.gz: a016f1a5ff26e8c8c5c1c95d0393f6ebe497919b0301282d412440779ec44fbd0c06e7b7ba95a8d30169d3c936dd4eaf5d3f8e59e8fdd6c563504b6027125a02
+  metadata.gz: e94245fd5af90a7411b842e13c44a5f85bbbe2544449de98eaa9c52dbb70f095938754bb65dea382f5275df26b6753c37c5cea24c564ff8f98ad6a0f29406e0e
+  data.tar.gz: f415da3e9c4ebee1ec8a6101676ae17a319d05ee248c8360d834d7f321190e791eb8274eb6fa0c5fcc74be5c1d2e1b77195894cb9179c3e33626bd090636327b

data/README.md

@@ -23,6 +23,7 @@ Add the plugin to your `_config.yml`:
 ```yaml
 plugins:
 - sutty-migration
+array_separator: ','
 ```
 
 Compile a CSV file with the following required fields:
@@ -66,6 +67,103 @@ To start migration just build your site:
 bundle exec jekyll build
 ```
 
+**Tip:** Files can also be JSON, TSV and YAML, since they're all
+supported by Jekyll.
+
+### Wordpress
+
+Instead of requiring you to install and configure MariaDB/MySQL, you can
+convert the database into SQLite3 like this:
+
+```bash
+git clone https://0xacab.org/sutty/mysql2sqlite.git
+cd mysql2sqlite
+./mysql2sqlite /path/to/database/dump.sql |
+  sed -re "s/, 0x([0-9a-f]+),/, X'\1',/i" |
+  sqlite3 wordpress.sqlite3
+```
+
+It will probably show some errors.
+
+Note the `sed` command is required to convert hexadecimal values into
+SQLite syntax, since `mysql2sqlite` doesn't support this yet.
+
+Wordpress websites can include lots of posts and metadata, depending on
+the amount of plugins installed.  We don't have an official way of
+dumping everything into Jekyll, because you will probably want to move
+things around.  You can write a plugin like this:
+
+```ruby
+# _plugins/wordpress.rb
+# frozen_string_literal: true
+
+require 'sutty_migration/wordpress'
+require 'sutty_migration/jekyll/document_creator'
+require 'jekyll-write-and-commit-changes'
+
+Jekyll::Hooks.register :site, :post_read, priority: :low do |site|
+  wp = SuttyMigration::Wordpress.new(site: site, database: 'wordpress.sqlite3', prefix: 'wp_', url: 'https://wordpre.ss')
+
+  # Download all files
+  wp.download_all
+
+  wp.posts(layout: 'post').each do |post|
+    doc = Jekyll::Document.create(site: site, title: post[:post_title], date: post[:post_date], collection: 'posts')
+    doc.content = post[:content]
+    doc.save
+  end
+end
+```
+
+### WordPress XML
+
+If you have the XML dump from a WordPress site, you can migrate content
+by writing a migration plugin.
+
+```ruby
+# frozen_string_literal: true
+
+require 'sutty_migration/jekyll/document_creator'
+require 'sutty_migration/wordpress_xml'
+require 'jekyll-write-and-commit-changes'
+require 'securerandom'
+
+# Run after reading the site
+Jekyll::Hooks.register :site, :post_read do |site|
+  # Put the XML dump at _files/wordpress.xml
+  xml = SuttyMigration::WordpressXml.new site: site, file: '_files/wordpress.xml'
+
+  # Download all files
+  xml.attachments.values.map(&:download)
+
+  # Migrate posts.  You can move metadata around and recover
+  # relationships or any info your theme requires.
+  xml.posts.values.each do |post|
+    # Update documents already migrated.
+    doc = Jekyll::Document.find_or_create(site: site, collection: locale, title: post.title, slug: post.slug, date: post.date)
+    # Don't change the UUIDv4
+    d.data['uuid'] ||= SecureRandom.uuid
+    d.data['draft'] = post.draft?
+    d.data['layout'] = 'post'
+    d.data['last_modified_at'] = post.last_modified_at
+
+    d.data['categories'] = post.categories.map { |c| c[:title] }
+    d.data['tags'] = post.tags.map { |t| t[:title] }
+
+    d.data['author'] = post.author[:email]
+    d.data['description'] = post.description
+    d.content = post.content
+
+    doc.save
+  rescue => e
+    Jekyll.logger.warn "Couldn't migrate #{post.title}"
+  end
+
+  exit # Stop here
+end
+```
+
+
 ## Contributing
 
 Bug reports and pull requests are welcome on 0xacab.org at

data/lib/sutty-migration.rb

@@ -1,71 +1,3 @@
 # frozen_string_literal: true
 
-require 'securerandom'
-require 'fast_blank'
-require 'jekyll-write-and-commit-changes'
-
-Jekyll::Hooks.register :site, :post_read, priority: :low do |site|
-  documents = site.documents
-
-  site.data['layouts']&.each do |name, layout|
-    site.data.dig('migration', name)&.each do |row|
-      row['date'] = Jekyll::Utils.parse_date(row['date']) unless row['date'].blank?
-
-      if row['id']
-        document = documents.find do |doc|
-          doc.data['id'] == row['id']
-        end
-      end
-
-      document ||=
-        begin
-          base = "#{row['date'] || Date.today.to_s}-#{Jekyll::Utils.slugify(row['title'], mode: 'latin')}.markdown"
-          path = File.join(site.source, '_posts', base)
-
-          raise ArgumentError, "Row #{row['id']} duplicates file #{base}" if File.exist? path
-
-          doc = Jekyll::Document.new(path, site: site, collection: site.collections['posts'])
-          site.collections['posts'] << doc
-
-          doc
-        end
-
-      row.each do |attribute, value|
-        row[attribute] =
-          case layout.dig(attribute, 'type')
-            when 'string' then value
-            when 'text' then value
-            when 'tel' then value
-            when 'color' then value # TODO: validar
-            when 'date' then Jekyll::Utils.parse_date(value)
-            when 'email' then value # TODO: validar
-            when 'url' then value # TODO: validar
-            when 'content' then value
-            when 'markdown_content' then value
-            when 'markdown' then value
-            when 'number' then value.to_i
-            when 'order' then value.to_i
-            when 'boolean' then !value.strip.empty?
-            when 'array' then value.split(',').map(&:strip)
-            # TODO: procesar los valores en base a los valores predefinidos
-            when 'predefined_array' then value.split(',').map(&:strip)
-            when 'image' then { 'path' => value, 'description' => '' }
-            when 'file' then { 'path' => value, 'description' => '' }
-            when 'geo' then %w[lat lng].zip(value.split(',', 2).map(&:to_f)).to_h
-            when 'belongs_to' then value
-            when 'has_many' then value.split(',').map(&:strip)
-            when 'has_and_belongs_to_many' then value.split(',').map(&:strip)
-            when 'related_posts' then value.split(',').map(&:strip)
-            when 'locales' then value.split(',').map(&:strip)
-            else value
-          end
-      end
-
-      document.data['uuid'] ||= SecureRandom.uuid
-      document.content = row.delete('content')
-
-      document.data.merge! row
-      document.save
-    end
-  end
-end
+require_relative 'sutty_migration/data'

data/lib/sutty_migration/core_extensions.rb

@@ -0,0 +1,26 @@
+# frozen_string_literal: true
+
+# Expandir String para poder verificar si está vacía
+require 'fast_blank'
+
+# Verificar que los valores nulos estén vacíos
+class NilClass
+  def blank?
+    true
+  end
+
+  def present?
+    false
+  end
+end
+
+# Verificar que una fecha está vacía
+class Time
+  def blank?
+    false
+  end
+
+  def present?
+    true
+  end
+end

data/lib/sutty_migration/data.rb

@@ -0,0 +1,80 @@
+# frozen_string_literal: true
+
+require 'securerandom'
+require_relative 'core_extensions'
+require_relative 'jekyll/document_creator'
+
+# Registers a plugin for converting CSV files into posts following
+# Sutty's layout definition.
+#
+# If jekyll-write-and-commit-changes is enabled, documents will be saved
+# on disk and commited is the build command is run with
+# JEKYLL_ENV=production
+Jekyll::Hooks.register :site, :post_read, priority: :low do |site|
+  documents = site.documents
+
+  array_separator = site.config.fetch('array_separator', ',')
+
+  site.data['layouts']&.each do |name, layout|
+    site.data.dig('migration', name)&.each do |row|
+      row['date'] = Jekyll::Utils.parse_date(row['date']) unless row['date'].blank?
+      row['date'] ||= Time.now
+
+      unless row['id'].blank?
+        document = documents.find do |doc|
+          doc.data['id'] == row['id']
+        end
+      end
+
+      document ||= Jekyll::Document.create(site: site, collection: 'posts',
+                                           **row.slice(*%w[date slug title]).transform_keys(&:to_sym))
+
+      row.each do |attribute, value|
+        next unless value.blank?
+
+        row[attribute] =
+          case layout.dig(attribute, 'type')
+          when 'string' then value
+          when 'text' then value
+          when 'tel' then value
+          # TODO: validate
+          when 'color' then value
+          when 'date' then Jekyll::Utils.parse_date(value)
+          # TODO: validate
+          when 'email' then value
+          # TODO: validate
+          when 'url' then value
+          when 'content' then value
+          when 'markdown_content' then value
+          when 'markdown' then value
+          when 'number' then value.to_i
+          when 'order' then value.to_i
+          when 'boolean' then !value.strip.empty?
+          when 'array' then value.split(array_separator).map(&:strip)
+          # TODO: process values from the default array
+          when 'predefined_array' then value.split(array_separator).map(&:strip)
+          when 'image' then { 'path' => value, 'description' => '' }
+          when 'file' then { 'path' => value, 'description' => '' }
+          when 'geo' then %w[lat lng].zip(value.split(array_separator, 2).map(&:to_f)).to_h
+          when 'belongs_to' then value
+          when 'has_many' then value.split(array_separator).map(&:strip)
+          when 'has_and_belongs_to_many' then value.split(array_separator).map(&:strip)
+          when 'related_posts' then value.split(array_separator).map(&:strip)
+          when 'locales' then value.split(array_separator).map(&:strip)
+          else value
+          end
+      end
+
+      document.data['uuid'] ||= SecureRandom.uuid
+      document.content = row.delete('content')
+
+      document.data.merge! row
+      document.save if document.respond_to? :save
+    end
+  end
+
+  next unless site.respond_to?(:repository)
+  next unless ENV['JEKYLL_ENV'] == 'production'
+
+  site.repository.commit 'CSV Migration'
+end

data/lib/sutty_migration/jekyll/document_creator.rb

@@ -0,0 +1,85 @@
+# frozen_string_literal: true
+
+require 'jekyll/utils'
+require_relative '../core_extensions'
+
+module SuttyMigration
+  module Jekyll
+    module DocumentCreator
+      class DocumentExists < ArgumentError; end
+
+      def self.included(base)
+        base.class_eval do
+          class << self
+            # Creates a new document in a collection or fails if it already
+            # exists.
+            #
+            # @param :site [Jekyll::Site] Jekyll site
+            # @param :date [Time] Post date
+            # @param :title [String] Post title
+            # @param :slug [String] Post slug, slugified title if empty
+            # @param :collection [Jekyll::Collection,String] Collection label or collection
+            # @return [Jekyll::Document] A new document
+            def create(site:, date:, title:, collection:, slug: nil)
+              collection = site.collections[collection] if collection.is_a? String
+              slug = ::Jekyll::Utils.slugify(title, mode: 'latin') if slug.blank?
+              basename = "#{date.strftime('%F')}-#{slug}.markdown"
+              path = File.join(collection.directory, basename)
+
+              raise DocumentExists, "#{path} already exists" if File.exist? path
+
+              ::Jekyll::Document.new(path, site: site, collection: collection).tap do |document|
+                collection.docs << document
+                document.data['title'] = title
+              end
+            end
+
+            # Finds a document by its relative path or creates it if it
+            # doesn't exist.  Helpful for idempotent migrations (create or
+            # update actions)
+            #
+            # @param :site [Jekyll::Site] Jekyll site
+            # @param :date [Time] Post date
+            # @param :title [String] Post title
+            # @param :slug [String] Post slug, slugified title if empty
+            # @param :collection [Jekyll::Collection,String] Collection label or collection
+            # @return [Jekyll::Document] The found document or a new one
+            def find_or_create(site:, date:, title:, collection:, slug: nil)
+              collection = site.collections[collection] if collection.is_a? String
+              slug = ::Jekyll::Utils.slugify(title, mode: 'latin') if slug.blank?
+              basename = "#{date.strftime('%F')}-#{slug}.markdown"
+              path = File.join(collection.relative_directory, basename)
+
+              return find(site: site, relative_path: path) if File.exist?(path)
+
+              create(site: site, date: date, title: title, slug: slug, collection: collection)
+            end
+
+            # Finds a document by its relative path
+            #
+            # @param :site [Jekyll::Site]
+            # @param :relative_path [String]
+            # @return [Jekyll::Document,Nil]
+            def find(site:, relative_path:)
+              indexed_documents_by_relative_path(site)[relative_path]
+            end
+
+            # Index documents by relative path for faster finding
+            #
+            # @param [Jekyll::Site]
+            # @return [Hash]
+            def indexed_documents_by_relative_path(site)
+              @indexed_documents_by_relative_path ||= site.documents.reduce({}) do |idx, doc|
+                idx.tap do |i|
+                  i[doc.relative_path] = doc
+                end
+              end
+            end
+          end
+        end
+      end
+    end
+  end
+end
+
+::Jekyll::Document.include SuttyMigration::Jekyll::DocumentCreator

data/lib/sutty_migration/wordpress.rb

@@ -0,0 +1,282 @@
+# frozen_string_literal: true
+
+require 'time'
+require 'securerandom'
+require 'sequel'
+require 'sqlite3'
+require 'json'
+require 'faraday'
+require 'progressbar'
+require 'jekyll/utils'
+
+module SuttyMigration
+  # Brings posts and attachments from a SQLite3 database.  You can
+  # convert a MySQL/MariaDB dump by using `mysql2sqlite`.
+  #
+  # It doesn't convert them into Jekyll posts but allows you to write a
+  # migration plugin where you can convert data by yourself.  We may add
+  # this feature in the future.
+  class Wordpress
+    attr_reader :site, :prefix, :limit, :url, :wp, :database, :multisite
+
+    # @param :site [Jekyll::Site] Jekyll site
+    # @param :url [String] Wordpress site URL (must be up for downloads)
+    # @param :database [String] Database path, by default `_data/wordpress.sqlite3`
+    # @param :prefix [String] WP table prefix
+    # @param :limit [Integer] Page length
+    # @param :multisite [Boolean] Site is multisite
+    def initialize(site:, url:, database: nil, prefix: 'wp_', limit: 10, multisite: nil)
+      @site = site
+      @prefix = prefix.freeze
+      @limit = limit.freeze
+      @url = url.freeze
+      @database = database || File.join(site.source, '_data', 'wordpress.sqlite3')
+      @multisite = multisite
+    end
+
+    # Generate database connections for a multisite WP
+    #
+    # @return [Hash] { "ID" => SuttyMigration::Wordpress }
+    def blogs
+      @blogs ||= wp["select * from #{prefix}blogs"].to_a.map do |blog|
+        url   = "https://#{blog[:domain]}#{blog[:path]}"
+        pfx   = "#{prefix}#{blog[:blog_id]}_" if blog[:blog_id] > 1
+        pfx ||= prefix
+
+        [blog[:blog_id],
+         blog.merge(db: self.class.new(site: site, url: url, prefix: pfx, database: database, limit: limit,
+                                       multisite: self))]
+      end.to_h
+    end
+
+    def options
+      @options ||= wp["select option_name, option_value from #{prefix}options"].to_a.map(&:values).to_h.transform_keys(&:to_sym)
+    end
+
+    # Open the database.
+    #
+    # @return [Sequel::SQLite::Database]
+    def wp
+      @wp ||= Sequel.sqlite(database).tap do |db|
+        db.extension :pagination
+      end
+    end
+
+    # Download all attachments.  Adds the local path to them.
+    #
+    # @param :progress [Boolean] Toggle progress bar
+    # @return [Nil]
+    def download_all(progress: true)
+      posts(layout: 'attachment').each do |attachment|
+        attachment[:front_matter]['file_path'] = download(url: attachment[:guid], progress: progress)
+      end
+    end
+
+    # Downloads a file if needed, optionally showing a progress bar.
+    #
+    # @param :url [String] File URL
+    # @param :progress [Boolean] Toggle progress bar
+    # @return [String] File local path
+    def download(url:, progress: true)
+      uri = URI(url)
+      dest = uri.path.sub(%r{\A/}, '')
+      full = File.join(site.source, dest)
+
+      return dest if File.exist? full
+
+      ::Jekyll.logger.info "Downloading #{dest}"
+
+      FileUtils.mkdir_p File.dirname(full)
+
+      File.open(full, 'w') do |f|
+        if progress
+          head = Faraday.head(url)
+          content_length = head.headers['content-length'].to_i
+          progress = ProgressBar.create(title: File.basename(dest), total: content_length, output: $stderr)
+        end
+
+        Faraday.get(url) do |req|
+          req.options.on_data = proc do |chunk, downloaded_bytes|
+            f.write chunk
+
+            if progress
+              progress.progress = downloaded_bytes > content_length ? content_length : downloaded_bytes
+            end
+          end
+        end
+      end
+
+      dest
+    end
+
+    # List post types
+    #
+    # @return [Array]
+    def layouts
+      @layouts ||= wp["select distinct post_type from #{prefix}posts"].to_a.map(&:values).flatten
+    end
+
+    # Finds all posts optionally filtering by post type.  This is not
+    # the official Sequel syntax, but it retrieves metadata as objects
+    # with a single query (and a sub-query).
+    #
+    # @param :layout [String] Layout name, one of #layouts
+    # @param :with_meta [Boolean] Toggle metadata pulling and conversion
+    # @return [Enumerator]
+    def posts(**options)
+      unless options[:layout].blank? || layouts.include?(options[:layout])
+        raise ArgumentError, "#{options[:layout]} must be one of #{layouts.join(', ')}"
+      end
+
+      wp[post_query(**options)].each_page(limit).to_a.map(&:to_a).flatten.tap do |p|
+        p.map do |post|
+          # Sequel parses dates on localtime
+          post[:date] = ::Jekyll::Utils.parse_date(post[:date]) unless post[:date].blank?
+          unless post[:last_modified_at].blank?
+            post[:last_modified_at] =
+              ::Jekyll::Utils.parse_date(post[:last_modified_at])
+          end
+
+          post[:front_matter] =
+            begin
+              unless post[:front_matter].blank?
+                JSON.parse(post[:front_matter]).transform_keys(&:to_sym).transform_values do |v|
+                  v.size == 1 ? v.first : v
+                end
+              end
+            rescue JSON::ParserError
+              {}
+            end
+          post[:terms] =
+            begin
+              unless post[:terms].blank?
+                JSON.parse(post[:terms]).transform_keys(&:to_sym).transform_values do |v|
+                  v.size == 1 ? v.first : v
+                end
+              end
+            rescue JSON::ParserError
+              {}
+            end
+        end
+      end
+    end
+
+    # Brings all users.
+    #
+    # @param :with_meta [Boolean] include metadata
+    # @return [Array]
+    def users(**options)
+      options[:with_meta] = true unless options.key? :with_meta
+
+      wp[user_query(**options)].each_page(limit).to_a.map(&:to_a).flatten.tap do |u|
+        next unless options[:with_meta]
+
+        u.map do |user|
+          user[:meta] = JSON.parse(user[:meta]).transform_keys(&:to_sym) unless user[:meta].blank?
+        end
+      end
+    end
+
+    private
+
+    # Finds all users.  If it's a multisite WP, we need to check the
+    # main table.
+    #
+    # @param :with_meta [Boolean] include metadata
+    # @return [String]
+    def user_query(with_meta: true)
+      pfx = multisite&.prefix || prefix
+
+      <<~EOQ
+        select
+          u.*
+          #{', json_group_object(m.meta_key, m.meta_value) as meta' if with_meta}
+        from #{pfx}users as u
+        #{"left join #{pfx}usermeta as m on m.user_id = u.id" if with_meta}
+        group by u.id
+      EOQ
+    end
+
+    # Query for posts, optionally bringing metadata as JSON objects.
+    #
+    # @param :layout [String] Layout name
+    # @param :with_meta [Boolean] Query metadata
+    # @return [String]
+    def post_query(layout: nil, with_meta: true)
+      <<~EOQ
+        select
+          p.ID as id,
+          strftime('%Y-%m-%d %H:%M:%S UTC', p.post_date_gmt) as date,
+          strftime('%Y-%m-%d %H:%M:%S UTC', p.post_modified_gmt) as last_modified_at,
+          p.post_author as author,
+          p.post_type as layout,
+          p.post_name as slug,
+          p.post_title as title,
+          p.post_content as content,
+          p.post_excerpt as excerpt,
+          p.post_status as status,
+          p.comment_status as comment_status,
+          p.ping_status as ping_status,
+          p.post_password as password,
+          p.to_ping as to_ping,
+          p.pinged as pinged,
+          p.post_content_filtered as content_filtered,
+          p.post_parent as parent,
+          p.guid as guid,
+          p.menu_order as menu_order,
+          p.post_mime_type as mime_type,
+          p.comment_count as comment_count
+          #{', f.front_matter as front_matter' if with_meta}
+          #{', t.terms as terms' if with_meta}
+        from #{prefix}posts as p
+        #{"left join (#{meta_query}) as f on f.post_id = p.ID" if with_meta}
+        #{"left join (#{terms_query}) as t on t.post_id = p.ID" if with_meta}
+        #{"where p.post_type = '#{layout}'" if layout}
+        group by p.ID
+      EOQ
+    end
+
+    # Recover the post meta as a JSON object with multiple values
+    # converted to arrays
+    #
+    # @return [String]
+    def meta_query
+      <<~EOQ
+        select
+          post_id,
+          json_group_object(meta_key, json(meta_values)) as front_matter
+        from (
+          select
+            post_id,
+            meta_key,
+            json_group_array(meta_value) as meta_values
+          from #{prefix}postmeta
+          group by post_id, meta_key
+        )
+        group by post_id
+      EOQ
+    end
+
+    # Term taxonomy query
+    #
+    # @param :layout [String] Layout name
+    # @return [String]
+    def terms_query
+      <<~EOQ
+        select
+          post_id,
+          json_group_object(taxonomy, json(terms)) as terms
+        from (
+          select
+            r.object_id as post_id,
+            tt.taxonomy,
+            json_group_array(t.name) as terms
+          from #{prefix}term_relationships as r
+          left join #{prefix}term_taxonomy as tt on tt.term_taxonomy_id = r.term_taxonomy_id
+          left join #{prefix}terms as t on t.term_id = tt.term_id
+          group by r.object_id)
+        group by post_id
+      EOQ
+    end
+  end
+end

data/lib/sutty_migration/wordpress_xml.rb

@@ -0,0 +1,154 @@
+# frozen_string_literal: true
+
+require 'nokogiri'
+require_relative 'wordpress_xml/post'
+require_relative 'wordpress_xml/attachment'
+
+module SuttyMigration
+  # Understands the XML dump generated by Wordpress and creates
+  # Jekyll::Documents
+  class WordpressXml
+    attr_reader :site, :file, :xml
+
+    # @param :site [Jekyll::Site] Jekyll site
+    # @param :file [String] File path
+    def initialize(site:, file:)
+      @site = site
+      @file = file
+      @xml  = Nokogiri::XML File.read(file)
+
+      # Make things easier by removing namespaces.
+      xml.remove_namespaces!
+    end
+
+    def inspect
+      '#<SuttyMigration::WordpressXml>'
+    end
+
+    # Site URL
+    #
+    # @return [String]
+    def url
+      @url ||= attribute_value(xml, 'channel > link')
+    end
+
+    # Site title
+    #
+    # @return [String]
+    def title
+      @title ||= attribute_value(xml, 'channel > title')
+    end
+
+    # Description
+    #
+    # @return [String]
+    def description
+      @description ||= attribute_value(xml, 'channel > description')
+    end
+
+    # Language
+    #
+    # TODO: Migrate multilanguage sites.
+    #
+    # @return [String]
+    def language
+      @language ||= attribute_value(xml, 'channel > language')
+    end
+
+    # Authors with attributes, indexed by author email.
+    #
+    # @return [Hash]
+    def authors
+      @authors ||= xml.css('channel > author').map do |author|
+        {
+          attribute_value(author, 'author_email') => {
+            id: attribute_value(author, 'author_id').to_i,
+            display_name: attribute_value(author, 'author_display_name'),
+            first_name: attribute_value(author, 'author_first_name'),
+            last_name: attribute_value(author, 'author_last_name'),
+            email: attribute_value(author, 'author_email')
+
+          }
+        }
+      end.reduce(&:merge)
+    end
+
+    # Categories with attributes, indexed by slug ("nicename")
+    #
+    # @return [Hash]
+    def categories
+      @categories ||= xml.css('channel > category').map do |category|
+        {
+          attribute_value(category, 'category_nicename') => {
+            id: attribute_value(category, 'term_id').to_i,
+            title: attribute_value(category, 'cat_name'),
+            parent: attribute_value(category, 'category_parent'),
+            slug: attribute_value(category, 'category_nicename')
+          }
+        }
+      end.reduce(&:merge)
+    end
+
+    # Tags with attributes, indexed by slug
+    #
+    # @return [Hash]
+    def tags
+      @tags ||= xml.css('channel > tag').map do |tag|
+        {
+          attribute_value(tag, 'tag_slug') => {
+            id: attribute_value(tag, 'term_id').to_i,
+            title: attribute_value(tag, 'tag_name'),
+            slug: attribute_value(tag, 'tag_slug')
+          }
+        }
+      end.reduce(&:merge)
+    end
+
+    # Posts, indexed by ID
+    #
+    # @return [Hash]
+    def posts
+      @posts ||= items_find_by('post_type', 'post').map do |post|
+        { attribute_value(post, 'post_id').to_i => Post.new(wordpress: self, item: post) }
+      end.reduce(&:merge)
+    end
+
+    # Pages, indexed by ID
+    #
+    # @return [Hash]
+    def pages
+      @pages ||= items_find_by('post_type', 'page').map do |page|
+        { attribute_value(page, 'post_id').to_i => Post.new(wordpress: self, item: page) }
+      end.reduce(&:merge)
+    end
+
+    # Attachments, indexed by ID
+    #
+    # @return [Hash]
+    def attachments
+      @attachments ||= items_find_by('post_type', 'attachment').map do |attachment|
+        { attribute_value(attachment, 'post_id').to_i => Attachment.new(wordpress: self, item: attachment) }
+      end.reduce(&:merge)
+    end
+
+    # Find items by attribute and value
+    #
+    # @param [String] Attribute name
+    # @param [String] Attribute value
+    # @return [Nokogiri::NodeSet]
+    def items_find_by(attribute, value)
+      xml.css('channel > item').select do |item|
+        attribute_value(item, attribute) == value
+      end
+    end
+
+    # Get element's attribute value
+    #
+    # @param [Nokogiri::XML::Element]
+    # @param [String]
+    # @return [String]
+    def attribute_value(element, attribute)
+      element.at_css(attribute).text
+    end
+  end
+end

data/lib/sutty_migration/wordpress_xml/attachment.rb

@@ -0,0 +1,69 @@
+# frozen_string_literal: true
+
+require_relative 'post'
+require 'php-serialize'
+require 'faraday'
+require 'progressbar'
+
+module SuttyMigration
+  class WordpressXml
+    # Represents an attachment or uploaded file.
+    class Attachment < Post
+      # File URL
+      #
+      # @return [String]
+      def attachment_url
+        @attachment_url ||= attribute_value 'attachment_url'
+      end
+
+      # File destination
+      #
+      # @return [String]
+      def dest
+        @dest ||= URI(attachment_url).path.sub(%r{\A/}, '')
+      end
+
+      # Metadata, with file information as a Hash
+      #
+      # @return [Hash]
+      def meta
+        super.tap do |m|
+          m['_wp_attachment_metadata'] = PHP.unserialize m['_wp_attachment_metadata']
+        end
+      end
+
+      # Download the file if it doesn't exist.  Optionally show a
+      # progress bar.
+      #
+      # @param :progress [Boolean]
+      # @return [Boolean]
+      def download(progress: true)
+        return true if File.exist? dest
+
+        ::Jekyll.logger.info "Downloading #{dest}"
+
+        FileUtils.mkdir_p File.dirname(dest)
+
+        File.open(dest, 'w') do |f|
+          if progress
+            head = Faraday.head(attachment_url)
+            content_length = head.headers['content-length'].to_i
+            progress = ProgressBar.create(title: File.basename(dest), total: content_length, output: $stderr)
+          end
+
+          Faraday.get(attachment_url) do |req|
+            req.options.on_data = proc do |chunk, downloaded_bytes|
+              f.write chunk
+
+              if progress
+                progress.progress = downloaded_bytes > content_length ? content_length : downloaded_bytes
+              end
+            end
+          end
+        end
+
+        File.exist? dest
+      end
+    end
+  end
+end

data/lib/sutty_migration/wordpress_xml/post.rb

@@ -0,0 +1,171 @@
+# frozen_string_literal: true
+
+require 'wordpress_formatting/wpautop'
+require 'jekyll/utils'
+
+module SuttyMigration
+  class WordpressXml
+    # Represents a WordPress post
+    class Post
+      attr_reader :wordpress, :item
+
+      # @param :wordpress [SuttyMigration::WordpressXml]
+      # @param :item [Nokogiri::XML::Element]
+      def initialize(wordpress:, item:)
+        @wordpress = wordpress
+        @item = item
+      end
+
+      def inspect
+        "#<SuttyMigration::WordpressXml::Post title=\"#{title}\">"
+      end
+
+      # Post ID
+      #
+      # @return [Integer]
+      def id
+        @id ||= attribute_value('post_id').to_i
+      end
+
+      # Permalink. Absolute URL to the post.
+      #
+      # @return [String]
+      def permalink
+        @permalink ||= attribute_value('link').sub(wordpress.url, '')
+      end
+
+      # Title
+      #
+      # @return [String]
+      def title
+        @title ||= attribute_value('title')
+      end
+
+      # Description
+      #
+      # @return [String]
+      def description
+        @description ||= attribute_value('description')
+      end
+
+      # Slug ("post name")
+      #
+      # @return [String]
+      def slug
+        @slug ||= attribute_value('post_name')
+      end
+
+      # Publication date.
+      #
+      # WordPress can store this date in three different fields and
+      # sometimes they come empty or invalid.
+      #
+      # @return [Time]
+      def date
+        @date ||= %w[pubDate post_date_gmt post_date].map do |date_attr|
+          ::Jekyll::Utils.parse_date attribute_value(date_attr)
+        rescue StandardError
+        end.compact.first
+      end
+
+      # Modification date.
+      #
+      # @return [Time]
+      def last_modified_at
+        @last_modified_at ||= ::Jekyll::Utils.parse_date attribute_value('post_modified_gmt')
+      end
+
+      # Content as HTML, with site URL removed.
+      #
+      # @return [String]
+      def content
+        @content ||= WordpressFormatting::Wpautop.wpautop(attribute_value('encoded')).gsub(
+          / (href|src)="#{wordpress.url}/, ' \\1="'
+        )
+      end
+
+      # Author attributes.
+      #
+      # @return [Hash]
+      def author
+        @author ||= wordpress.authors[attribute_value('creator')]
+      end
+
+      # Post password.  Use with jekyll-crypto.
+      #
+      # @return [String]
+      def password
+        @password ||= attribute_value 'post_password'
+      end
+
+      # Tags with attributes.
+      #
+      # @return [Hash]
+      def tags
+        @tags ||= item.css('category').select do |c|
+          c[:domain] == 'post_tag'
+        end.map do |c|
+          wordpress.tags[c[:nicename]]
+        end
+      end
+
+      # Categories with attributes.
+      #
+      # @return [Hash]
+      def categories
+        @categories ||= item.css('category').select do |c|
+          c[:domain] == 'category'
+        end.map do |c|
+          wordpress.categories[c[:nicename]]
+        end
+      end
+
+      # Metadata.  Plugins store useful information here.  Duplicated
+      # keys are returned as an Array of values.
+      #
+      # @return [Hash]
+      def meta
+        @meta ||= {}.tap do |meta|
+          item.css('postmeta').each do |m|
+            key = m.css('meta_key').text
+            value = m.css('meta_value').text
+
+            case meta[key]
+            when nil then meta[key] = value
+            when String then meta[key] = [meta[key], value]
+            when Array then meta[key] << value
+            end
+          end
+        end
+      end
+
+      # Order.  Higher are sorted on top by jekyll-order.
+      #
+      # @return [Integer]
+      def order
+        @order ||= attribute_value 'is_sticky'
+      end
+
+      # Publication status
+      #
+      # @return [Boolean]
+      def published?
+        @published ||= attribute_value('status') == 'publish'
+      end
+
+      # Publication status
+      #
+      # @return [Boolean]
+      def draft?
+        @draft ||= attribute_value('status') == 'draft'
+      end
+
+      # Get a value from the attribute
+      #
+      # @return [String]
+      def attribute_value(key)
+        item.at_css(key).text
+      end
+    end
+  end
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: sutty-migration
 version: !ruby/object:Gem::Version
-  version: 0.1.2
+  version: 0.3.0
 platform: ruby
 authors:
 - f
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2021-05-28 00:00:00.000000000 Z
+date: 2021-08-10 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: jekyll
@@ -52,6 +52,118 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '1.0'
+- !ruby/object:Gem::Dependency
+  name: faraday
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.4'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.4'
+- !ruby/object:Gem::Dependency
+  name: progressbar
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.11'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.11'
+- !ruby/object:Gem::Dependency
+  name: sqlite3
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.4'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.4'
+- !ruby/object:Gem::Dependency
+  name: sequel
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '5.45'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '5.45'
+- !ruby/object:Gem::Dependency
+  name: wordpress-formatting
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.1.0
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.1.0
+- !ruby/object:Gem::Dependency
+  name: nokogiri
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 1.12.0
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 1.12.0
+- !ruby/object:Gem::Dependency
+  name: php-serialize
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 1.3.0
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 1.3.0
+- !ruby/object:Gem::Dependency
+  name: pry
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
 description: Takes datafiles and converts them into posts
 email:
 - f@sutty.nl
@@ -64,7 +176,13 @@ files:
 - LICENSE.txt
 - README.md
 - lib/sutty-migration.rb
-- lib/wordpress.rb
+- lib/sutty_migration/core_extensions.rb
+- lib/sutty_migration/data.rb
+- lib/sutty_migration/jekyll/document_creator.rb
+- lib/sutty_migration/wordpress.rb
+- lib/sutty_migration/wordpress_xml.rb
+- lib/sutty_migration/wordpress_xml/attachment.rb
+- lib/sutty_migration/wordpress_xml/post.rb
 homepage: https://0xacab.org/sutty/jekyll/sutty-migration
 licenses:
 - GPL-3.0

data/lib/wordpress.rb

@@ -1,174 +0,0 @@
-# frozen_string_literal: true
-
-# Debug
-require 'pry'
-# Generar UUIDs
-require 'securerandom'
-# Traer resultados de la base de datos
-require 'sequel'
-require 'sqlite3'
-require 'json'
-# Limpieza de contenido
-require 'loofah'
-require 'rails/html/scrubbers'
-require 'rails/html/sanitizer'
-require 'reverse_markdown'
-# Descargar archivos
-require 'faraday'
-
-class Wordpress
-  attr_reader :site, :prefix, :limit, :url
-
-  def initialize(site:, url:, prefix: 'wp_', limit: 10)
-    @site = site
-    @prefix = prefix.freeze
-    @limit = limit.freeze
-    @url = url.freeze
-
-    # Conectarse a la base de datos
-    @wp = Sequel.sqlite(File.join(site.source, '_data', 'wordpress', 'post.sqlite3'))
-    # Las funciones de JSON usan mucha CPU, vamos a traer de a pocos
-    # registros.
-    @wp.extension :pagination
-  end
-
-  def download(file)
-    dest = 'wp-content/uploads/' + file
-    full = File.join(site.source, dest)
-
-    return dest if File.exist? full
-
-    Jekyll.logger.info "Downloading #{dest}"
-
-    FileUtils.mkdir_p File.dirname(full)
-
-    File.open(full, 'w') do |f|
-      Faraday.get(url + '/' + dest) do |req|
-        req.options.on_data = Proc.new do |chunk, _|
-          f.write chunk
-        end
-      end
-    end
-
-    dest
-  end
-
-  # Obtiene todos los posts opcionalmente filtrando por tipo de post.
-  # No es la forma oficial de Sequel pero no tenemos tiempo de
-  # aprenderla específicamente y además tenemos las opciones en formato
-  # JSON que no estarían soportadas.
-  def posts(layout: nil)
-    query  = post_query.dup
-    query += " where post_type = '#{layout}'" if layout
-    query += ' group by posts.ID'
-
-    @wp[query].each_page(limit)
-  end
-
-  def meta(id:)
-    @wp[meta_query(id: id)].to_a
-  end
-
-  private
-
-  # Obtener todos los posts, json_objectagg requiere mariadb 10.5
-  def post_query
-    @post_query ||= <<~EOQ
-      select ID as id,
-        post_title as title,
-        post_name as slug,
-        post_type as layout,
-        strftime('%Y-%m-%d', post_date) as date,
-        post_status as status,
-        post_content as content,
-        json_group_object(meta_key, meta_value) as data
-      from #{prefix}posts as posts
-      left join #{prefix}postmeta as frontmatter
-      on posts.ID = frontmatter.post_id
-    EOQ
-  end
-
-  def meta_query(id:)
-    <<~EOQ
-      SELECT
-        terms.name AS `name`,
-        ttax.taxonomy AS `type`,
-        ttax.parent AS `parent`,
-        ttax.term_id AS `id`
-      FROM
-        #{prefix}terms AS `terms`,
-        #{prefix}term_relationships AS `trels`,
-        #{prefix}term_taxonomy AS `ttax`
-      WHERE
-        trels.object_id = '#{id}' AND
-        trels.term_taxonomy_id = ttax.term_taxonomy_id AND
-        terms.term_id = ttax.term_id
-    EOQ
-  end
-end
-
-# Antes de generar el sitio vamos a leer todos los artículos desde la
-# base de datos y generarlos localmente.
-Jekyll::Hooks.register :site, :post_read do |site|
-  wp = Wordpress.new(site: site,
-                     url: site.config.dig('wordpress', 'url'),
-                     prefix: site.config.dig('wordpress', 'prefix'))
-
-  collection = site.collections['posts']
-  ascii_re = Regexp.new("\P{ASCII}").freeze
-  sanitizer = Rails::Html::SafeListSanitizer.new
-
-  # Traer todas las imágenes cargadas y descargarlas
-  attachments = wp.posts(layout: 'attachment').map do |page|
-    page.map do |attachment|
-      attachment[:data] = JSON.parse(attachment[:data]) unless attachment[:data].nil?
-      file = attachment.dig(:data, '_wp_attached_file')
-
-      next unless file
-
-      dest = wp.download(file)
-
-      # Tener un mapa de IDs y archivos destino
-      [ attachment[:id], dest ]
-    end
-  end.compact.flatten(1).to_h
-
-  %w[post page].each do |type|
-    wp.posts(layout: type).each do |page|
-      page.each do |post|
-        # Convertir los datos extra en un Hash
-        post[:data] = JSON.parse(post[:data]) unless post[:data].nil?
-        post[:slug] = Jekyll::Utils.slugify(post[:title], mode: 'latin') if post[:slug].empty?
-        post[:meta] = wp.meta id: post[:id]
-
-        path = File.join(site.source, '_posts', post.slice(:date, :slug).values.join('-') + '.markdown')
-
-        if File.exist? path
-          Jekyll.logger.info "#{path} ya fue migrado, actualizando"
-
-          doc = site.documents.find do |d|
-            d['id'] == post[:id]
-          end
-        else
-          # Crear un post nuevo y agregarlo a la colección
-          collection.docs << doc = Jekyll::Document.new(path, site: site, collection: collection)
-          doc.data['uuid'] = SecureRandom.uuid
-        end
-
-        thumbnail = post.dig(:data, '_thumbnail_id')&.to_i
-
-        doc.data['layout'] = type
-        doc.data['title'] = post[:title]
-        doc.data['draft'] = post[:status] != 'publish'
-        doc.data['id'] = post[:id]
-        doc.data['date'] = Jekyll::Utils.parse_date(post[:date])
-        doc.data['tags'] = post[:meta].select { |k| k[:type] == 'post_tag' }.map { |k| k[:name] }
-        doc.data['categories'] = post[:meta].select { |k| k[:type] == 'category' }.map { |k| k[:name] }
-        doc.data['image'] = attachments[thumbnail] if thumbnail
-
-        doc.content = ReverseMarkdown.convert(sanitizer.sanitize(post[:content]))
-        doc.save
-      end
-    end
-  end
-end