Module: PaperTrail::VersionConcern::ClassMethods

Defined in:
lib/paper_trail/version_concern.rb

Overview

:nodoc:

Instance Method Summary collapse

Instance Method Details

#between(start_time, end_time) ⇒ Object 



52
53
54
55
56
57
 
# File 'lib/paper_trail/version_concern.rb', line 52

def between(start_time, end_time)
  where(
    arel_table[:created_at].gt(start_time).
    and(arel_table[:created_at].lt(end_time))
  ).order(timestamp_sort_order)
end

#creates ⇒ Object 



36
37
38
 
# File 'lib/paper_trail/version_concern.rb', line 36

def creates
  where event: "create"
end

#destroys ⇒ Object 



44
45
46
 
# File 'lib/paper_trail/version_concern.rb', line 44

def destroys
  where event: "destroy"
end

#item_subtype_column_present? ⇒ Boolean

Returns:

  • (Boolean) 


28
29
30
 
# File 'lib/paper_trail/version_concern.rb', line 28

def item_subtype_column_present?
  column_names.include?("item_subtype")
end

#not_creates ⇒ Object 



48
49
50
 
# File 'lib/paper_trail/version_concern.rb', line 48

def not_creates
  where "event <> ?", "create"
end

#object_changes_col_is_json? ⇒ Boolean

Returns whether the ‘object_changes` column is using the `json` type supported by PostgreSQL.

Returns:

  • (Boolean) 


137
138
139
 
# File 'lib/paper_trail/version_concern.rb', line 137

def object_changes_col_is_json?
  %i[json jsonb].include?(columns_hash["object_changes"].try(:type))
end

#object_col_is_json? ⇒ Boolean

Returns whether the ‘object` column is using the `json` type supported by PostgreSQL.

Returns:

  • (Boolean) 


131
132
133
 
# File 'lib/paper_trail/version_concern.rb', line 131

def object_col_is_json?
  %i[json jsonb].include?(columns_hash["object"].type)
end

#preceding(obj, timestamp_arg = false) ⇒ Object

Returns versions before ‘obj`.

Parameters:

  • obj —

    • a ‘Version` or a timestamp

  • timestamp_arg (defaults to: false) —

    • boolean - When true, ‘obj` is a timestamp.

    Default: false.

Returns:

  • ‘ActiveRecord::Relation`



148
149
150
151
152
153
154
 
# File 'lib/paper_trail/version_concern.rb', line 148

def preceding(obj, timestamp_arg = false)
  if timestamp_arg != true && primary_key_is_int?
    preceding_by_id(obj)
  else
    preceding_by_timestamp(obj)
  end
end

#primary_key_is_int? ⇒ Boolean

Returns:

  • (Boolean) 


123
124
125
126
127
 
# File 'lib/paper_trail/version_concern.rb', line 123

def primary_key_is_int?
  @primary_key_is_int ||= columns_hash[primary_key].type == :integer
rescue StandardError # TODO: Rescue something more specific
  true
end

#subsequent(obj, timestamp_arg = false) ⇒ Object

Returns versions after ‘obj`.

Parameters:

  • obj —

    • a ‘Version` or a timestamp

  • timestamp_arg (defaults to: false) —

    • boolean - When true, ‘obj` is a timestamp.

    Default: false.

Returns:

  • ‘ActiveRecord::Relation`



163
164
165
166
167
168
169
 
# File 'lib/paper_trail/version_concern.rb', line 163

def subsequent(obj, timestamp_arg = false)
  if timestamp_arg != true && primary_key_is_int?
    subsequent_by_id(obj)
  else
    subsequent_by_timestamp(obj)
  end
end

#timestamp_sort_order(direction = "asc") ⇒ Object

Defaults to using the primary key as the secondary sort order if possible.



61
62
63
64
65
 
# File 'lib/paper_trail/version_concern.rb', line 61

def timestamp_sort_order(direction = "asc")
  [arel_table[:created_at].send(direction.downcase)].tap do |array|
    array << arel_table[primary_key].send(direction.downcase) if primary_key_is_int?
  end
end

#updates ⇒ Object 



40
41
42
 
# File 'lib/paper_trail/version_concern.rb', line 40

def updates
  where event: "update"
end

#where_object(args = {}) ⇒ Object

Given a hash of attributes like ‘name: ’Joan’‘, query the `versions.objects` column.

“‘ SELECT “versions”.* FROM “versions” WHERE (“versions”.“object” LIKE ’% name: Joan %‘) “`

This is useful for finding versions where a given attribute had a given value. Imagine, in the example above, that Joan had changed her name and we wanted to find the versions before that change.

Based on the data type of the ‘object` column, the appropriate SQL operator is used. For example, a text column will use `like`, and a jsonb column will use `@>`.

Raises:

  • (ArgumentError) 


87
88
89
90
 
# File 'lib/paper_trail/version_concern.rb', line 87

def where_object(args = {})
  raise ArgumentError, "expected to receive a Hash" unless args.is_a?(Hash)
  Queries::Versions::WhereObject.new(self, args).execute
end

#where_object_changes(args = {}) ⇒ Object

Given a hash of attributes like ‘name: ’Joan’‘, query the `versions.objects_changes` column.

“‘ SELECT “versions”.* FROM “versions” WHERE .. (“versions”.“object_changes” LIKE ’% name:

  • Joan

%‘ OR “versions”.“object_changes” LIKE ’% name: -%

  • Joan

%‘) “`

This is useful for finding versions immediately before and after a given attribute had a given value. Imagine, in the example above, that someone changed their name to Joan and we wanted to find the versions immediately before and after that change.

Based on the data type of the ‘object` column, the appropriate SQL operator is used. For example, a text column will use `like`, and a jsonb column will use `@>`.

Raises:

  • (ArgumentError) 


118
119
120
121
 
# File 'lib/paper_trail/version_concern.rb', line 118

def where_object_changes(args = {})
  raise ArgumentError, "expected to receive a Hash" unless args.is_a?(Hash)
  Queries::Versions::WhereObjectChanges.new(self, args).execute
end

#with_item_keys(item_type, item_id) ⇒ Object 



32
33
34
 
# File 'lib/paper_trail/version_concern.rb', line 32

def with_item_keys(item_type, item_id)
  where item_type: item_type, item_id: item_id
end