-
Notifications
You must be signed in to change notification settings - Fork 639
/
Copy pathannotate_models.rb
988 lines (836 loc) · 35.6 KB
/
annotate_models.rb
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
# rubocop:disable Metrics/ModuleLength
require 'bigdecimal'
require 'annotate/constants'
require_relative 'annotate_models/file_patterns'
module AnnotateModels
# Annotate Models plugin use this header
COMPAT_PREFIX = '== Schema Info'.freeze
COMPAT_PREFIX_MD = '## Schema Info'.freeze
PREFIX = '== Schema Information'.freeze
PREFIX_MD = '## Schema Information'.freeze
END_MARK = '== Schema Information End'.freeze
SKIP_ANNOTATION_PREFIX = '# -\*- SkipSchemaAnnotations'.freeze
MATCHED_TYPES = %w(test fixture factory serializer scaffold controller helper).freeze
# Don't show limit (#) on these column types
# Example: show "integer" instead of "integer(4)"
NO_LIMIT_COL_TYPES = %w(integer bigint boolean).freeze
# Don't show default value for these column types
NO_DEFAULT_COL_TYPES = %w(json jsonb hstore).freeze
INDEX_CLAUSES = {
unique: {
default: 'UNIQUE',
markdown: '_unique_'
},
where: {
default: 'WHERE',
markdown: '_where_'
},
using: {
default: 'USING',
markdown: '_using_'
}
}.freeze
MAGIC_COMMENT_MATCHER = Regexp.new(/(^#\s*encoding:.*(?:\n|r\n))|(^# coding:.*(?:\n|\r\n))|(^# -\*- coding:.*(?:\n|\r\n))|(^# -\*- encoding\s?:.*(?:\n|\r\n))|(^#\s*frozen_string_literal:.+(?:\n|\r\n))|(^# -\*- frozen_string_literal\s*:.+-\*-(?:\n|\r\n))/).freeze
class << self
def annotate_pattern(options = {})
if options[:wrapper_open]
return /(?:^(\n|\r\n)?# (?:#{options[:wrapper_open]}).*(\n|\r\n)?# (?:#{COMPAT_PREFIX}|#{COMPAT_PREFIX_MD}).*?(\n|\r\n)(#.*(\n|\r\n))*(\n|\r\n)*)|^(\n|\r\n)?# (?:#{COMPAT_PREFIX}|#{COMPAT_PREFIX_MD}).*?(\n|\r\n)(#.*(\n|\r\n))*(\n|\r\n)*/
end
/^(\n|\r\n)?# (?:#{COMPAT_PREFIX}|#{COMPAT_PREFIX_MD}).*?(\n|\r\n)(#.*(\n|\r\n))*(\n|\r\n)*/
end
def model_dir
@model_dir.is_a?(Array) ? @model_dir : [@model_dir || 'app/models']
end
attr_writer :model_dir
def root_dir
if @root_dir.blank?
['']
elsif @root_dir.is_a?(String)
@root_dir.split(',')
else
@root_dir
end
end
attr_writer :root_dir
def skip_subdirectory_model_load
# This option is set in options[:skip_subdirectory_model_load]
# and stops the get_loaded_model method from loading a model from a subdir
if @skip_subdirectory_model_load.blank?
false
else
@skip_subdirectory_model_load
end
end
attr_writer :skip_subdirectory_model_load
def get_patterns(options, pattern_types = [])
current_patterns = []
root_dir.each do |root_directory|
Array(pattern_types).each do |pattern_type|
patterns = FilePatterns.generate(root_directory, pattern_type, options)
current_patterns += if pattern_type.to_sym == :additional_file_patterns
patterns
else
patterns.map { |p| p.sub(/^[\/]*/, '') }
end
end
end
current_patterns
end
# Simple quoting for the default column value
def quote(value)
case value
when NilClass then 'NULL'
when TrueClass then 'TRUE'
when FalseClass then 'FALSE'
when Float, Integer then value.to_s
# BigDecimals need to be output in a non-normalized form and quoted.
when BigDecimal then value.to_s('F')
when Array then value.map { |v| quote(v) }
else
value.inspect
end
end
def schema_default(klass, column)
quote(klass.column_defaults[column.name])
end
def retrieve_indexes_from_table(klass)
table_name = klass.table_name
return [] unless table_name
indexes = klass.connection.indexes(table_name)
return indexes if indexes.any? || !klass.table_name_prefix
# Try to search the table without prefix
table_name_without_prefix = table_name.to_s.sub(klass.table_name_prefix, '')
if klass.connection.table_exists?(table_name_without_prefix)
klass.connection.indexes(table_name_without_prefix)
else
[]
end
end
# Use the column information in an ActiveRecord class
# to create a comment block containing a line for
# each column. The line contains the column name,
# the type (and length), and any optional attributes
def get_schema_info(klass, header, options = {}) # rubocop:disable Metrics/MethodLength
info = "# #{header}\n"
info << get_schema_header_text(klass, options)
max_size = max_schema_info_width(klass, options)
md_names_overhead = 6
md_type_allowance = 18
bare_type_allowance = 16
if options[:format_markdown]
info << sprintf( "# %-#{max_size + md_names_overhead}.#{max_size + md_names_overhead}s | %-#{md_type_allowance}.#{md_type_allowance}s | %s\n", 'Name', 'Type', 'Attributes' )
info << "# #{ '-' * ( max_size + md_names_overhead ) } | #{'-' * md_type_allowance} | #{ '-' * 27 }\n"
end
cols = columns(klass, options)
with_comments = with_comments?(klass, options)
with_comments_column = with_comments_column?(klass, options)
# Precalculate Values
cols_meta = cols.map do |col|
col_comment = with_comments || with_comments_column ? col.comment&.gsub(/\n/, "\\n") : nil
col_type = get_col_type(col)
attrs = get_attributes(col, col_type, klass, options)
col_name = if with_comments && col_comment
"#{col.name}(#{col_comment})"
else
col.name
end
simple_formatted_attrs = attrs.join(", ")
[col.name, { col_type: col_type, attrs: attrs, col_name: col_name, simple_formatted_attrs: simple_formatted_attrs, col_comment: col_comment }]
end.to_h
# Output annotation
bare_max_attrs_length = cols_meta.map { |_, m| m[:simple_formatted_attrs].length }.max
cols.each do |col|
col_type = cols_meta[col.name][:col_type]
attrs = cols_meta[col.name][:attrs]
col_name = cols_meta[col.name][:col_name]
simple_formatted_attrs = cols_meta[col.name][:simple_formatted_attrs]
col_comment = cols_meta[col.name][:col_comment]
if options[:format_rdoc]
info << sprintf("# %-#{max_size}.#{max_size}s<tt>%s</tt>", "*#{col_name}*::", attrs.unshift(col_type).join(", ")).rstrip + "\n"
elsif options[:format_yard]
info << sprintf("# @!attribute #{col_name}") + "\n"
ruby_class = col.respond_to?(:array) && col.array ? "Array<#{map_col_type_to_ruby_classes(col_type)}>": map_col_type_to_ruby_classes(col_type)
info << sprintf("# @return [#{ruby_class}]") + "\n"
elsif options[:format_markdown]
name_remainder = max_size - col_name.length - non_ascii_length(col_name)
type_remainder = (md_type_allowance - 2) - col_type.length
info << (sprintf("# **`%s`**%#{name_remainder}s | `%s`%#{type_remainder}s | `%s`", col_name, " ", col_type, " ", attrs.join(", ").rstrip)).gsub('``', ' ').rstrip + "\n"
elsif with_comments_column
info << format_default(col_name, max_size, col_type, bare_type_allowance, simple_formatted_attrs, bare_max_attrs_length, col_comment)
else
info << format_default(col_name, max_size, col_type, bare_type_allowance, simple_formatted_attrs)
end
end
if options[:show_indexes] && klass.table_exists?
info << get_index_info(klass, options)
end
if options[:show_foreign_keys] && klass.table_exists?
info << get_foreign_key_info(klass, options)
end
if options[:show_check_constraints] && klass.table_exists?
info << get_check_constraint_info(klass, options)
end
info << get_schema_footer_text(klass, options)
end
def get_schema_header_text(klass, options = {})
info = "#\n"
if options[:format_markdown]
info << "# Table name: `#{klass.table_name}`\n"
info << "#\n"
info << "# ### Columns\n"
else
info << "# Table name: #{klass.table_name}\n"
end
info << "#\n"
end
def get_schema_footer_text(_klass, options = {})
info = ''
if options[:format_rdoc]
info << "#--\n"
info << "# #{END_MARK}\n"
info << "#++\n"
else
info << "#\n"
end
end
def get_index_info(klass, options = {})
index_info = if options[:format_markdown]
"#\n# ### Indexes\n#\n"
else
"#\n# Indexes\n#\n"
end
indexes = retrieve_indexes_from_table(klass)
return '' if indexes.empty?
max_size = indexes.collect{|index| index.name.size}.max + 1
indexes.sort_by(&:name).each do |index|
index_info << if options[:format_markdown]
final_index_string_in_markdown(index)
else
final_index_string(index, max_size)
end
end
index_info
end
def get_col_type(col)
if (col.respond_to?(:bigint?) && col.bigint?) || /\Abigint\b/ =~ col.sql_type
'bigint'
else
(col.type || col.sql_type).to_s
end.dup
end
def index_columns_info(index)
Array(index.columns).map do |col|
if index.try(:orders) && index.orders[col.to_s]
"#{col} #{index.orders[col.to_s].upcase}"
else
col.to_s.gsub("\r", '\r').gsub("\n", '\n')
end
end
end
def index_unique_info(index, format = :default)
index.unique ? " #{INDEX_CLAUSES[:unique][format]}" : ''
end
def index_where_info(index, format = :default)
value = index.try(:where).try(:to_s)
if value.blank?
''
else
" #{INDEX_CLAUSES[:where][format]} #{value}"
end
end
def index_using_info(index, format = :default)
value = index.try(:using) && index.using.try(:to_sym)
if !value.blank? && value != :btree
" #{INDEX_CLAUSES[:using][format]} #{value}"
else
''
end
end
def final_index_string_in_markdown(index)
details = sprintf(
"%s%s%s",
index_unique_info(index, :markdown),
index_where_info(index, :markdown),
index_using_info(index, :markdown)
).strip
details = " (#{details})" unless details.blank?
sprintf(
"# * `%s`%s:\n# * **`%s`**\n",
index.name,
details,
index_columns_info(index).join("`**\n# * **`")
)
end
def final_index_string(index, max_size)
sprintf(
"# %-#{max_size}.#{max_size}s %s%s%s%s",
index.name,
"(#{index_columns_info(index).join(',')})",
index_unique_info(index),
index_where_info(index),
index_using_info(index)
).rstrip + "\n"
end
def hide_limit?(col_type, options)
excludes =
if options[:hide_limit_column_types].blank?
NO_LIMIT_COL_TYPES
else
options[:hide_limit_column_types].split(',')
end
excludes.include?(col_type)
end
def hide_default?(col_type, options)
excludes =
if options[:hide_default_column_types].blank?
NO_DEFAULT_COL_TYPES
else
options[:hide_default_column_types].split(',')
end
excludes.include?(col_type)
end
def get_foreign_key_info(klass, options = {})
fk_info = if options[:format_markdown]
"#\n# ### Foreign Keys\n#\n"
else
"#\n# Foreign Keys\n#\n"
end
return '' unless klass.connection.respond_to?(:supports_foreign_keys?) &&
klass.connection.supports_foreign_keys? && klass.connection.respond_to?(:foreign_keys)
foreign_keys = klass.connection.foreign_keys(klass.table_name)
return '' if foreign_keys.empty?
format_name = lambda do |fk|
return fk.options[:column] if fk.name.blank?
options[:show_complete_foreign_keys] ? fk.name : fk.name.gsub(/(?<=^fk_rails_)[0-9a-f]{10}$/, '...')
end
max_size = foreign_keys.map(&format_name).map(&:size).max + 1
foreign_keys.sort_by {|fk| [format_name.call(fk), fk.column]}.each do |fk|
ref_info = "#{fk.column} => #{fk.to_table}.#{fk.primary_key}"
constraints_info = ''
constraints_info += "ON DELETE => #{fk.on_delete} " if fk.on_delete
constraints_info += "ON UPDATE => #{fk.on_update} " if fk.on_update
constraints_info.strip!
fk_info << if options[:format_markdown]
sprintf("# * `%s`%s:\n# * **`%s`**\n", format_name.call(fk), constraints_info.blank? ? '' : " (_#{constraints_info}_)", ref_info)
else
sprintf("# %-#{max_size}.#{max_size}s %s %s", format_name.call(fk), "(#{ref_info})", constraints_info).rstrip + "\n"
end
end
fk_info
end
def get_check_constraint_info(klass, options = {})
cc_info = if options[:format_markdown]
"#\n# ### Check Constraints\n#\n"
else
"#\n# Check Constraints\n#\n"
end
return '' unless klass.connection.respond_to?(:supports_check_constraints?) &&
klass.connection.supports_check_constraints? && klass.connection.respond_to?(:check_constraints)
check_constraints = klass.connection.check_constraints(klass.table_name)
return '' if check_constraints.empty?
max_size = check_constraints.map { |check_constraint| check_constraint.name.size }.max + 1
check_constraints.sort_by(&:name).each do |check_constraint|
expression = check_constraint.expression ? "(#{check_constraint.expression.squish})" : nil
cc_info << if options[:format_markdown]
cc_info_markdown = sprintf("# * `%s`", check_constraint.name)
cc_info_markdown << sprintf(": `%s`", expression) if expression
cc_info_markdown << "\n"
else
sprintf("# %-#{max_size}.#{max_size}s %s", check_constraint.name, expression).rstrip + "\n"
end
end
cc_info
end
# Add a schema block to a file. If the file already contains
# a schema info block (a comment starting with "== Schema Information"),
# check if it matches the block that is already there. If so, leave it be.
# If not, remove the old info block and write a new one.
#
# == Returns:
# true or false depending on whether the file was modified.
#
# === Options (opts)
# :force<Symbol>:: whether to update the file even if it doesn't seem to need it.
# :position_in_*<Symbol>:: where to place the annotated section in fixture or model file,
# :before, :top, :after or :bottom. Default is :before.
#
def annotate_one_file(file_name, info_block, position, options = {})
return false unless File.exist?(file_name)
old_content = File.read(file_name)
return false if old_content =~ /#{SKIP_ANNOTATION_PREFIX}.*\n/
# Ignore the Schema version line because it changes with each migration
header_pattern = /(^# Table name:.*?\n(#.*[\r]?\n)*[\r]?)/
old_header = old_content.match(header_pattern).to_s
new_header = info_block.match(header_pattern).to_s
column_pattern = /^#[\t ]+[\w\*\.`]+[\t ]+.+$/
old_columns = old_header && old_header.scan(column_pattern).sort
new_columns = new_header && new_header.scan(column_pattern).sort
return false if old_columns == new_columns && !options[:force]
abort "annotate error. #{file_name} needs to be updated, but annotate was run with `--frozen`." if options[:frozen]
# Replace inline the old schema info with the new schema info
wrapper_open = options[:wrapper_open] ? "# #{options[:wrapper_open]}\n" : ""
wrapper_close = options[:wrapper_close] ? "# #{options[:wrapper_close]}\n" : ""
wrapped_info_block = "#{wrapper_open}#{info_block}#{wrapper_close}"
old_annotation = old_content.match(annotate_pattern(options)).to_s
# if there *was* no old schema info or :force was passed, we simply
# need to insert it in correct position
if old_annotation.empty? || options[:force]
magic_comments_block = magic_comments_as_string(old_content)
old_content.gsub!(MAGIC_COMMENT_MATCHER, '')
old_content.sub!(annotate_pattern(options), '')
new_content = if %w(after bottom).include?(options[position].to_s)
magic_comments_block + (old_content.rstrip + "\n\n" + wrapped_info_block)
elsif magic_comments_block.empty?
magic_comments_block + wrapped_info_block + old_content.lstrip
else
magic_comments_block + "\n" + wrapped_info_block + old_content.lstrip
end
else
# replace the old annotation with the new one
# keep the surrounding whitespace the same
space_match = old_annotation.match(/\A(?<start>\s*).*?\n(?<end>\s*)\z/m)
new_annotation = space_match[:start] + wrapped_info_block + space_match[:end]
new_content = old_content.sub(annotate_pattern(options), new_annotation)
end
File.open(file_name, 'wb') { |f| f.puts new_content }
true
end
def magic_comments_as_string(content)
magic_comments = content.scan(MAGIC_COMMENT_MATCHER).flatten.compact
if magic_comments.any?
magic_comments.join
else
''
end
end
def remove_annotation_of_file(file_name, options = {})
if File.exist?(file_name)
content = File.read(file_name)
return false if content =~ /#{SKIP_ANNOTATION_PREFIX}.*\n/
wrapper_open = options[:wrapper_open] ? "# #{options[:wrapper_open]}\n" : ''
content.sub!(/(#{wrapper_open})?#{annotate_pattern(options)}/, '')
File.open(file_name, 'wb') { |f| f.puts content }
true
else
false
end
end
def matched_types(options)
types = MATCHED_TYPES.dup
types << 'admin' if options[:active_admin] =~ Annotate::Constants::TRUE_RE && !types.include?('admin')
types << 'additional_file_patterns' if options[:additional_file_patterns].present?
types
end
# Given the name of an ActiveRecord class, create a schema
# info block (basically a comment containing information
# on the columns and their types) and put it at the front
# of the model and fixture source files.
#
# === Options (opts)
# :position_in_class<Symbol>:: where to place the annotated section in model file
# :position_in_test<Symbol>:: where to place the annotated section in test/spec file(s)
# :position_in_fixture<Symbol>:: where to place the annotated section in fixture file
# :position_in_factory<Symbol>:: where to place the annotated section in factory file
# :position_in_serializer<Symbol>:: where to place the annotated section in serializer file
# :exclude_tests<Symbol>:: whether to skip modification of test/spec files
# :exclude_fixtures<Symbol>:: whether to skip modification of fixture files
# :exclude_factories<Symbol>:: whether to skip modification of factory files
# :exclude_serializers<Symbol>:: whether to skip modification of serializer files
# :exclude_scaffolds<Symbol>:: whether to skip modification of scaffold files
# :exclude_controllers<Symbol>:: whether to skip modification of controller files
# :exclude_helpers<Symbol>:: whether to skip modification of helper files
# :exclude_sti_subclasses<Symbol>:: whether to skip modification of files for STI subclasses
#
# == Returns:
# an array of file names that were annotated.
#
def annotate(klass, file, header, options = {})
begin
klass.reset_column_information
info = get_schema_info(klass, header, options)
model_name = klass.name.underscore
table_name = klass.table_name
model_file_name = File.join(file)
annotated = []
if annotate_one_file(model_file_name, info, :position_in_class, options_with_position(options, :position_in_class))
annotated << model_file_name
end
matched_types(options).each do |key|
exclusion_key = "exclude_#{key.pluralize}".to_sym
position_key = "position_in_#{key}".to_sym
# Same options for active_admin models
if key == 'admin'
exclusion_key = 'exclude_class'.to_sym
position_key = 'position_in_class'.to_sym
end
next if options[exclusion_key]
get_patterns(options, key)
.map { |f| resolve_filename(f, model_name, table_name) }
.map { |f| expand_glob_into_files(f) }
.flatten
.each do |f|
if annotate_one_file(f, info, position_key, options_with_position(options, position_key))
annotated << f
end
end
end
rescue StandardError => e
$stderr.puts "Unable to annotate #{file}: #{e.message}"
$stderr.puts "\t" + e.backtrace.join("\n\t") if options[:trace]
end
annotated
end
# position = :position_in_fixture or :position_in_class
def options_with_position(options, position_in)
options.merge(position: (options[position_in] || options[:position]))
end
# Return a list of the model files to annotate.
# If we have command line arguments, they're assumed to the path
# of model files from root dir. Otherwise we take all the model files
# in the model_dir directory.
def get_model_files(options)
model_files = []
model_files = list_model_files_from_argument unless options[:is_rake]
return model_files unless model_files.empty?
model_dir.each do |dir|
Dir.chdir(dir) do
list = if options[:ignore_model_sub_dir]
Dir["*.rb"].map { |f| [dir, f] }
else
Dir["**/*.rb"].reject { |f| f["concerns/"] }.map { |f| [dir, f] }
end
model_files.concat(list)
end
end
model_files
rescue SystemCallError
$stderr.puts "No models found in directory '#{model_dir.join("', '")}'."
$stderr.puts "Either specify models on the command line, or use the --model-dir option."
$stderr.puts "Call 'annotate --help' for more info."
exit 1
end
def list_model_files_from_argument
return [] if ARGV.empty?
specified_files = ARGV.map { |file| File.expand_path(file) }
model_files = model_dir.flat_map do |dir|
absolute_dir_path = File.expand_path(dir)
specified_files
.find_all { |file| file.start_with?(absolute_dir_path) }
.map { |file| [dir, file.sub("#{absolute_dir_path}/", '')] }
end
if model_files.size != specified_files.size
puts "The specified file could not be found in directory '#{model_dir.join("', '")}'."
puts "Call 'annotate --help' for more info."
exit 1
end
model_files
end
private :list_model_files_from_argument
# Retrieve the classes belonging to the model names we're asked to process
# Check for namespaced models in subdirectories as well as models
# in subdirectories without namespacing.
def get_model_class(file)
model_path = file.gsub(/\.rb$/, '')
model_dir.each { |dir| model_path = model_path.gsub(/^#{dir}/, '').gsub(/^\//, '') }
begin
get_loaded_model(model_path, file) || raise(BadModelFileError.new)
rescue LoadError
# this is for non-rails projects, which don't get Rails auto-require magic
file_path = File.expand_path(file)
if File.file?(file_path) && Kernel.require(file_path)
retry
elsif model_path =~ /\//
model_path = model_path.split('/')[1..-1].join('/').to_s
retry
else
raise
end
end
end
# Retrieve loaded model class
def get_loaded_model(model_path, file)
unless skip_subdirectory_model_load
loaded_model_class = get_loaded_model_by_path(model_path)
return loaded_model_class if loaded_model_class
end
# We cannot get loaded model when `model_path` is loaded by Rails
# auto_load/eager_load paths. Try all possible model paths one by one.
absolute_file = File.expand_path(file)
model_paths =
$LOAD_PATH.map(&:to_s)
.select { |path| absolute_file.include?(path) }
.map { |path| absolute_file.sub(path, '').sub(/\.rb$/, '').sub(/^\//, '') }
model_paths
.map { |path| get_loaded_model_by_path(path) }
.find { |loaded_model| !loaded_model.nil? }
end
# Retrieve loaded model class by path to the file where it's supposed to be defined.
def get_loaded_model_by_path(model_path)
ActiveSupport::Inflector.constantize(ActiveSupport::Inflector.camelize(model_path))
rescue StandardError, LoadError
# Revert to the old way but it is not really robust
ObjectSpace.each_object(::Class)
.select do |c|
Class === c && # note: we use === to avoid a bug in activesupport 2.3.14 OptionMerger vs. is_a?
c.ancestors.respond_to?(:include?) && # to fix FactoryGirl bug, see https://github.com/ctran/annotate_models/pull/82
c.ancestors.include?(ActiveRecord::Base)
end.detect { |c| ActiveSupport::Inflector.underscore(c.to_s) == model_path }
end
def parse_options(options = {})
self.model_dir = split_model_dir(options[:model_dir]) if options[:model_dir]
self.root_dir = options[:root_dir] if options[:root_dir]
self.skip_subdirectory_model_load = options[:skip_subdirectory_model_load].present?
end
def split_model_dir(option_value)
option_value = option_value.is_a?(Array) ? option_value : option_value.split(',')
option_value.map(&:strip).reject(&:empty?)
end
# We're passed a name of things that might be
# ActiveRecord models. If we can find the class, and
# if its a subclass of ActiveRecord::Base,
# then pass it to the associated block
def do_annotations(options = {})
parse_options(options)
header = options[:format_markdown] ? PREFIX_MD.dup : PREFIX.dup
version = ActiveRecord::Migrator.current_version rescue 0
if options[:include_version] && version > 0
header << "\n# Schema version: #{version}"
end
annotated = []
get_model_files(options).each do |path, filename|
annotate_model_file(annotated, File.join(path, filename), header, options)
end
if annotated.empty?
puts 'Model files unchanged.'
else
puts "Annotated (#{annotated.length}): #{annotated.join(', ')}"
end
end
def expand_glob_into_files(glob)
Dir.glob(glob)
end
def annotate_model_file(annotated, file, header, options)
begin
return false if /#{SKIP_ANNOTATION_PREFIX}.*/ =~ (File.exist?(file) ? File.read(file) : '')
klass = get_model_class(file)
do_annotate = klass.is_a?(Class) &&
klass < ActiveRecord::Base &&
(!options[:exclude_sti_subclasses] || !(klass.superclass < ActiveRecord::Base && klass.table_name == klass.superclass.table_name)) &&
!klass.abstract_class? &&
klass.table_exists?
annotated.concat(annotate(klass, file, header, options)) if do_annotate
rescue BadModelFileError => e
unless options[:ignore_unknown_models]
$stderr.puts "Unable to annotate #{file}: #{e.message}"
$stderr.puts "\t" + e.backtrace.join("\n\t") if options[:trace]
end
rescue StandardError => e
$stderr.puts "Unable to annotate #{file}: #{e.message}"
$stderr.puts "\t" + e.backtrace.join("\n\t") if options[:trace]
end
end
def remove_annotations(options = {})
parse_options(options)
deannotated = []
deannotated_klass = false
get_model_files(options).each do |file|
file = File.join(file)
begin
klass = get_model_class(file)
if klass < ActiveRecord::Base && !klass.abstract_class?
model_name = klass.name.underscore
table_name = klass.table_name
model_file_name = file
deannotated_klass = true if remove_annotation_of_file(model_file_name, options)
get_patterns(options, matched_types(options))
.map { |f| resolve_filename(f, model_name, table_name) }
.each do |f|
if File.exist?(f)
remove_annotation_of_file(f, options)
deannotated_klass = true
end
end
end
deannotated << klass if deannotated_klass
rescue StandardError => e
$stderr.puts "Unable to deannotate #{File.join(file)}: #{e.message}"
$stderr.puts "\t" + e.backtrace.join("\n\t") if options[:trace]
end
end
puts "Removed annotations from: #{deannotated.join(', ')}"
end
def resolve_filename(filename_template, model_name, table_name)
filename_template
.gsub('%MODEL_NAME%', model_name)
.gsub('%PLURALIZED_MODEL_NAME%', model_name.pluralize)
.gsub('%TABLE_NAME%', table_name || model_name.pluralize)
end
def classified_sort(cols)
rest_cols = []
timestamps = []
associations = []
id = nil
cols.each do |c|
if c.name.eql?('id')
id = c
elsif c.name.eql?('created_at') || c.name.eql?('updated_at')
timestamps << c
elsif c.name[-3, 3].eql?('_id')
associations << c
else
rest_cols << c
end
end
[rest_cols, timestamps, associations].each { |a| a.sort_by!(&:name) }
([id] << rest_cols << timestamps << associations).flatten.compact
end
private
def with_comments?(klass, options)
options[:with_comment] &&
klass.columns.first.respond_to?(:comment) &&
klass.columns.any? { |col| !col.comment.nil? }
end
def with_comments_column?(klass, options)
options[:with_comment_column] &&
klass.columns.first.respond_to?(:comment) &&
klass.columns.any? { |col| !col.comment.nil? }
end
def max_schema_info_width(klass, options)
cols = columns(klass, options)
if with_comments?(klass, options)
max_size = cols.map do |column|
column.name.size + (column.comment ? width(column.comment) : 0)
end.max || 0
max_size += 2
else
max_size = cols.map(&:name).map(&:size).max
end
max_size += options[:format_rdoc] ? 5 : 1
max_size
end
# rubocop:disable Metrics/ParameterLists
def format_default(col_name, max_size, col_type, bare_type_allowance, simple_formatted_attrs, bare_max_attrs_length = 0, col_comment = nil)
sprintf(
"# %s:%s %s %s",
mb_chars_ljust(col_name, max_size),
mb_chars_ljust(col_type, bare_type_allowance),
mb_chars_ljust(simple_formatted_attrs, bare_max_attrs_length),
col_comment
).rstrip + "\n"
end
def width(string)
string.chars.inject(0) { |acc, elem| acc + (elem.bytesize == 3 ? 2 : 1) }
end
def mb_chars_ljust(string, length)
string = string.to_s
padding = length - width(string)
if padding > 0
string + (' ' * padding)
else
string[0..length-1]
end
end
def non_ascii_length(string)
string.to_s.chars.reject(&:ascii_only?).length
end
def map_col_type_to_ruby_classes(col_type)
case col_type
when 'integer' then Integer.to_s
when 'float' then Float.to_s
when 'decimal' then BigDecimal.to_s
when 'datetime', 'timestamp', 'time' then Time.to_s
when 'date' then Date.to_s
when 'text', 'string', 'binary', 'inet', 'uuid' then String.to_s
when 'json', 'jsonb' then Hash.to_s
when 'boolean' then 'Boolean'
end
end
def columns(klass, options)
cols = klass.columns
cols += translated_columns(klass)
if ignore_columns = options[:ignore_columns]
cols = cols.reject do |col|
col.name.match(/#{ignore_columns}/)
end
end
cols = cols.sort_by(&:name) if options[:sort]
cols = classified_sort(cols) if options[:classified_sort]
cols
end
##
# Add columns managed by the globalize gem if this gem is being used.
def translated_columns(klass)
return [] unless klass.respond_to? :translation_class
ignored_cols = ignored_translation_table_colums(klass)
klass.translation_class.columns.reject do |col|
ignored_cols.include? col.name.to_sym
end
end
##
# These are the columns that the globalize gem needs to work but
# are not necessary for the models to be displayed as annotations.
def ignored_translation_table_colums(klass)
# Construct the foreign column name in the translations table
# eg. Model: Car, foreign column name: car_id
foreign_column_name = [
klass.table_name.to_s.singularize,
'_id'
].join.to_sym
[
:id,
:created_at,
:updated_at,
:locale,
foreign_column_name
]
end
##
# Get the list of attributes that should be included in the annotation for
# a given column.
def get_attributes(column, column_type, klass, options)
attrs = []
attrs << "default(#{schema_default(klass, column)})" unless column.default.nil? || hide_default?(column_type, options)
attrs << 'unsigned' if column.respond_to?(:unsigned?) && column.unsigned?
attrs << 'not null' unless column.null
attrs << 'primary key' if klass.primary_key && (klass.primary_key.is_a?(Array) ? klass.primary_key.collect(&:to_sym).include?(column.name.to_sym) : column.name.to_sym == klass.primary_key.to_sym)
if column_type == 'decimal'
column_type << "(#{column.precision}, #{column.scale})"
elsif !%w[spatial geometry geography].include?(column_type)
if column.limit && !options[:format_yard]
if column.limit.is_a? Array
attrs << "(#{column.limit.join(', ')})"
else
column_type << "(#{column.limit})" unless hide_limit?(column_type, options)
end
end
end
# Check out if we got an array column
attrs << 'is an Array' if column.respond_to?(:array) && column.array
# Check out if we got a geometric column
# and print the type and SRID
if column.respond_to?(:geometry_type)
attrs << [column.geometry_type, column.try(:srid)].compact.join(', ')
elsif column.respond_to?(:geometric_type) && column.geometric_type.present?
attrs << [column.geometric_type.to_s.downcase, column.try(:srid)].compact.join(', ')
end
# Check if the column has indices and print "indexed" if true
# If the index includes another column, print it too.
if options[:simple_indexes] && klass.table_exists?# Check out if this column is indexed
indices = retrieve_indexes_from_table(klass)
if indices = indices.select { |ind| ind.columns.include? column.name }
indices.sort_by(&:name).each do |ind|
next if ind.columns.is_a?(String)
ind = ind.columns.reject! { |i| i == column.name }
attrs << (ind.empty? ? "indexed" : "indexed => [#{ind.join(", ")}]")
end
end
end
attrs
end
end
class BadModelFileError < LoadError
def to_s
"file doesn't contain a valid model class"
end
end
end