simplec 0.6.0 → 0.7.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 3e3189399a08bf6dc14a452311b79f7f48eb6a16
-  data.tar.gz: 179e5ac787ed6456ac46282808591748ba10e560
+  metadata.gz: 96f4e9c1f6b7ebca90065eaafc12867bd2402760
+  data.tar.gz: bdbd5d66111b3afa442c3cb9a1d84cd25bc7387a
 SHA512:
-  metadata.gz: 61dcc253ee7661aaea00164e559700fedd731d5932049fb9f21e285a65fd4840e35a3c1dbc84f433cd79bf61611cf75b0bffe7abd6876a4db7ce63774dba7f72
-  data.tar.gz: 8d0115591036074ab888821f17b04fdfdd2692bb941273a8329fa61d25104763ea1630215a06b5fac7ed714bd1ea9bfff5b1a44f93f05a1d1589f825cb51fa1e
+  metadata.gz: 2e60a95279a8c1aebbb8cdbfd2a1c17f8f66f8d6598a6d6fd307342a17597f95431a60bfccadf100f0c5df526bb0bc9d8b930741e92218791bcf0617f816ba31
+  data.tar.gz: 3a267018544e4d5c9c5990beee28b1d694ce24055d70f7156481830a2c0de3a825d58ccebfbef14f787e3dc4a33b429b94703b8c83bb640393b05e067dbc9ca2

data/app/models/simplec/document.rb

@@ -25,6 +25,7 @@ module Simplec
     end
 
     def validate_slug!
+      return if self.slug.nil?
       similar = self.class.where(slug: self.slug).where.not(id: self.id)
       return if similar.size == 0
       return if (

data/app/models/simplec/page.rb

@@ -55,6 +55,7 @@ module Simplec
     before_validation :match_parent_subdomain
     before_validation :build_path
     after_save :link_embedded_images!
+    after_save :index!
 
     # @!attribute slug
     #   The value is normalized to a string starting without a leading slash
@@ -86,6 +87,37 @@ module Simplec
     #   know what you are doing.
     #   @return [JSON]
 
+    # @!method search(term, options={})
+    #
+    # If the term is nil or blank, all results will be returned.
+    #
+    # @param term [String]
+    # @param options [Hash]
+    # @option options [Class, Array, String] :types
+    #   A Class or Array of Classes (of page types, typically `Page::Home`) to
+    #   limit results to.
+    # @option options [Symbol, String] all other options
+    #   All other options are matched to the `query` JSONB field. These are
+    #   all direct matches on the indexed `query` field. If you want to do
+    #   anything more complicated, append it to the scope returned.
+    #
+    # @return [ActiveRecord::Relation] a relation for the query
+    # @!scope class
+    scope :search, ->(term, options={}) {
+      _types      = Array(options.delete(:types))
+
+      query = all
+      query = query.where(type: _types) if _types.any?
+      options.each { |k,v| query = query.where("query->>:k = :v", k: k, v: v) }
+
+      if term.blank?
+        query
+      else
+        tsq = tsquery term
+        query.where("tsv @@ #{tsq}").order("ts_rank_cd(tsv, #{tsq}) DESC")
+      end
+    }
+
     # Define a field on the page
     #
     # There is as template for each type for customization located in:
@@ -140,12 +172,86 @@ module Simplec
       _fields = case type
                 when :file
                   fields.select {|k, v| FILE_FIELDS.member?(v[:type])}
+                when :textual
+                  fields.select {|k, v| !FILE_FIELDS.member?(v[:type])}
                 else
                   fields
                 end
       _fields.keys
     end
 
+    # Set extra attributes on the record for querying.
+    #
+    # @example set attributes
+    #   class Page::Home < Page
+    #     has_many :tags
+    #
+    #     field :category
+    #
+    #     # Where category is a Simplec::Page::field and tags is a defined
+    #     # method.
+    #     search_query_attributes! :category, :tags
+    #
+    #     def tags
+    #       self.tags.pluck(:name)
+    #     end
+    #   end
+    #
+    #   # Built-in matching
+    #   Page.search('foo', category: 'how-to')
+    #
+    #   # Manual matching
+    #   Page.search('bar').where("query->>'tags' IN ('home', 'garden')")
+    #
+    def self.search_query_attributes!(*args)
+      @_search_query_attrs = args.map(&:to_sym)
+    end
+
+    # Get extra attributes on the record for querying.
+    #
+    # See #search_query_attributes! for more information.
+    #
+    # @return [Array] of attributes
+    def self.search_query_attributes
+      @_search_query_attrs = Set.new(@_search_query_attrs).add(:id).to_a
+    end
+
+    # Index every record.
+    #
+    # Internally this method iterates over all pages in batches of 3.
+    #
+    # @return [NilClass]
+    def self.index!
+      find_each(batch_size: 3) { |page| page.index! }
+    end
+
+    # Create a to_tsquery statement.
+    #
+    # Mainly used internally, but could be used in custom queries.
+    #
+    # @param input [String] string to be queried
+    # @param options [Hash] optional
+    # @option options [String] :language defaults to 'english'
+    #   This is really a future addition, all of the tsvector fields are set to
+    #   'english'.
+    #
+    # @return [String] a to_tsquery statement
+    def self.tsquery(input, options={})
+      options[:language] ||= 'english'
+      value = input.to_s.strip
+      value  = value.
+        gsub('(', '').
+        gsub(')', '').
+        gsub(%q('), '').
+        gsub(' ', '\\ ').
+        gsub(':', '').
+        gsub("\t", '').
+        gsub("!", '')
+      value << ':*'
+      query = "to_tsquery(?, ?)"
+      sanitize_sql_array([query, options[:language], value])
+    end
+
     # Return field options for building forms.
     #
     def field_options
@@ -156,7 +262,6 @@ module Simplec
     #
     # This is a recursive, expensive call.
     #
-    # @return [Array] of parent Pages
     def parents
       page, parents = self, Array.new
       while page.parent
@@ -225,9 +330,64 @@ module Simplec
       @layouts ||= Subdomain.new.layouts
     end
 
+    # Index this record for search.
+    #
+    # Internally, this method uses update_columns so it can be used in
+    # `after_save` callbacks, etc.
+    #
+    # @return [Boolean] success
+    def index!
+      set_search_text!
+      set_query_attributes!
+      update_columns text: self.text, query: self.query
+    end
+
+    # Extract text out of HTML or plain strings. Basically removes html
+    # formatting.
+    #
+    # @param attributes [Symbol, String]
+    #   variable list of attributes or methods to be extracted for search
+    #
+    # @return [String] content of each attribute separated by new lines
+    def extract_search_text(*attributes)
+      Array(attributes).map { |meth|
+        Nokogiri::HTML(self.send(meth)).xpath("//text()").
+          map {|node| text = node.text; text.try(:strip!); text}.join(" ")
+      }.reject(&:blank?).join("\n")
+    end
+
+    # Set the text which will be index.
+    #
+    # a title, meta_description
+    # b slug, path (non-printable, add tags, added terms)
+    # c textual fields
+    # d (reserved for sub-records, etc)
+    #
+    # 'a' correlates to 'A' priority in Postgresql. For more information:
+    # https://www.postgresql.org/docs/9.6/static/functions-textsearch.html
+    #
+    def set_search_text!
+      self.text['a'] = extract_search_text :title, :meta_description
+      self.text['b'] = extract_search_text :slug, :path
+      self.text['c'] = extract_search_text *self.class.field_names(:textual)
+      self.text['d'] = nil
+    end
+
+    # Build query attribute hash.
+    #
+    # Internally stored as JSONB.
+    #
+    # @return [Hash] to be set for query attribute
+    def set_query_attributes!
+      attr_names = self.class.search_query_attributes.map(&:to_s)
+      self.query = attr_names.inject({}) { |memo, attr|
+        memo[attr] = self.send(attr)
+        memo
+      }
+    end
+
     # @!visibility private
     module Normalizers
-
       def slug=(val)
         val = val ? val.to_s.split('/').reject(&:blank?).join('/') : val
         super val

data/db/migrate/20170917144923_add_search_to_simplec_pages.rb

@@ -0,0 +1,58 @@
+class AddSearchToSimplecPages < ActiveRecord::Migration[5.0]
+  def change
+    add_column :simplec_pages, :tsv, :tsvector
+    add_column :simplec_pages, :query, :jsonb
+    add_column :simplec_pages, :text, :jsonb
+
+    add_index :simplec_pages, :tsv, using: :gin
+
+    reversible do |dir|
+      dir.up {
+        execute <<-SQL.gsub(/\s+/, " ").strip
+ALTER TABLE simplec_pages ALTER COLUMN query
+  SET DEFAULT '{}'::JSONB;
+ALTER TABLE simplec_pages ALTER COLUMN text
+  SET DEFAULT '{"a": null, "b": null, "c": null, "d": null}'::JSONB;
+
+UPDATE simplec_pages
+  SET
+    query = '{}'::JSONB,
+    text = '{"a": null, "b": null, "c": null, "d": null}'::JSONB;
+
+CREATE INDEX simplec_pages_query ON simplec_pages
+  USING GIN (query jsonb_path_ops);
+
+CREATE FUNCTION simplec_pages_search_tsvector_update_trigger() RETURNS trigger AS $$
+begin
+  new.tsv :=
+    setweight(
+      to_tsvector('pg_catalog.english', coalesce(new.text ->> 'a','')), 'A'
+    ) ||
+    setweight(
+      to_tsvector('pg_catalog.english', coalesce(new.text ->> 'b','')), 'B'
+    ) ||
+    setweight(
+      to_tsvector('pg_catalog.english', coalesce(new.text ->> 'c','')), 'C'
+    ) ||
+    setweight(
+      to_tsvector('pg_catalog.english', coalesce(new.text ->> 'd','')), 'D'
+    );
+  return new;
+end
+
+$$ LANGUAGE plpgsql;
+CREATE TRIGGER simplec_pages_search_tsvector_update BEFORE INSERT OR UPDATE
+  ON simplec_pages FOR EACH ROW
+  EXECUTE PROCEDURE simplec_pages_search_tsvector_update_trigger();
+        SQL
+      }
+      dir.down {
+        execute <<-SQL.gsub(/\s+/, " ").strip
+DROP TRIGGER IF EXISTS simplec_pages_search_tsvector_update ON simplec_pages;
+DROP FUNCTION IF EXISTS simplec_pages_search_tsvector_update_trigger();
+DROP INDEX simplec_pages_query;
+        SQL
+      }
+    end
+  end
+end

data/lib/simplec/version.rb

@@ -1,3 +1,3 @@
 module Simplec
-  VERSION = '0.6.0'
+  VERSION = '0.7.0'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: simplec
 version: !ruby/object:Gem::Version
-  version: 0.6.0
+  version: 0.7.0
 platform: ruby
 authors:
 - Matt Smith
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2017-09-16 00:00:00.000000000 Z
+date: 2017-09-17 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rails
@@ -153,6 +153,7 @@ files:
 - db/migrate/20170809210304_create_simplec_embedded_images.rb
 - db/migrate/20170814211816_create_simplec_documents.rb
 - db/migrate/20170814211929_create_simplec_document_sets.rb
+- db/migrate/20170917144923_add_search_to_simplec_pages.rb
 - lib/simplec.rb
 - lib/simplec/action_controller/extensions.rb
 - lib/simplec/action_view/helper.rb