Module: GoodData::UserFilterBuilder

Defined in:
lib/gooddata/models/user_filters/user_filter_builder.rb

Class Method Summary collapse

Class Method Details

.collect_labels(data) ⇒ Object

Groups the values by particular label. And passes each group to deduplication

Parameters:

  • options (Hash)

Returns:



126
127
128
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 126

def self.collect_labels(data)
  data.group_by { |x| [x[:label], x[:over], x[:to]] }.map { |l, v| { label: l[0], over: l[1], to: l[2], values: UserFilterBuilder.collect_values(v) } }
end

.collect_values(data) ⇒ Object

Collects specific values and deduplicates if necessary



131
132
133
134
135
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 131

def self.collect_values(data)
  data.mapcat do |e|
    e[:values]
  end.uniq
end

.create_attrs_cache(filters, options = {}) ⇒ Object 



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
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 193

def self.create_attrs_cache(filters, options = {})
  project = options[:project]

  labels = filters.flat_map do |f|
    f[:filters]
  end

  over_cache = labels.reduce({}) do |a, e|
    a[e[:over]] = e[:over]
    a
  end
  to_cache = labels.reduce({}) do |a, e|
    a[e[:to]] = e[:to]
    a
  end
  cache = over_cache.merge(to_cache)
  attr_cache = {}
  cache.each_pair do |k, v|
    begin
      attr_cache[k] = project.attributes(v)
    rescue
      nil
    end
  end
  attr_cache
end

.create_cache(data, key) ⇒ Object 



137
138
139
140
141
142
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 137

def self.create_cache(data, key)
  data.reduce({}) do |a, e|
    a[e.send(key)] = e
    a
  end
end

.create_expression(filter, labels_cache, lookups_cache, attr_cache, options = {}, login) ⇒ Object

Creates a MAQL expression(s) based on the filter defintion. Takes the filter definition looks up any necessary values and provides API executable MAQL rubocop:disable Metrics/ParameterLists

Parameters:

  • labels_cache —

    e.g. { 'label_uri': label_object }

  • lookups_cache —

    e.g. { 'label_uri': { "[email protected]": 'value_uri' }}



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
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 233

def self.create_expression(filter, labels_cache, lookups_cache, attr_cache, options = {}, )
  values = filter[:values]
  # Do not create MUF for label when all its values is NULL (https://jira.intgdc.com/browse/TMA-1361)
  non_null_values = values.select { |value| !value.nil? && value.downcase != 'null' }
  return ['TRUE', []] if non_null_values.empty?

  label = labels_cache[filter[:label]]
  if label.nil?
    err_message = "Unable to apply filter values: #{values} since the project: #{options[:project].pid} doesn't have label: #{filter[:label]} for login: #{}"
    if options[:ignore_missing_values]
      GoodData.logger.warn(err_message)
      return ['TRUE', []]
    else
      fail err_message
    end
  end
  errors = []

  element_uris_by_values = Hash[values.map do |v|
    if lookups_cache.key?(label.uri)
      [v, lookups_cache[label.uri][v]]
    else
      [v, label.find_value_uri(v)]
    end
  end]

  missing_value_errors = element_uris_by_values.select { |_, v| v.nil? }.map do |k, _|
    {
      type: :error,
      label: label.title,
      value: k,
      reason: 'Can not find the value of the attribute referenced in the MUF'
    }
  end
  errors += missing_value_errors unless options[:ignore_missing_values]

  element_uris = element_uris_by_values.values.compact
  # happens when data is not yet loaded in the project
  no_values = element_uris.empty?

  expression = if no_values && options[:restrict_if_missing_all_values]
                 # create a filter that is always false to ensure the user can not see any data
                 # as the proper MUF can not be constructed yet
                 case options[:type]
                 when :muf
                   '1 <> 1'
                 when :variable
                   nil
                 end
               elsif no_values
                 # create a filter that is always true to ensure the user can see all data
                 'TRUE'
               elsif filter[:over] && filter[:to]
                 over = attr_cache[filter[:over]]
                 to = attr_cache[filter[:to]]
                 "([#{label.attribute_uri}] IN (#{element_uris.sort.map { |e| '[' + e + ']' }.join(', ')})) OVER [#{over && over.uri}] TO [#{to && to.uri}]"
               else
                 "[#{label.attribute_uri}] IN (#{element_uris.sort.map { |e| '[' + e + ']' }.join(', ')})"
               end
  [expression, errors]
end

.create_filter(label, values) ⇒ Object 



105
106
107
108
109
110
111
112
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 105

def self.create_filter(label, values)
  {
    :label => label[:label],
    :values => values,
    :over => label[:over],
    :to => label[:to]
  }
end

.create_label_cache(result, options = {}) ⇒ Object 



170
171
172
173
174
175
176
177
178
179
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 170

def self.create_label_cache(result, options = {})
  project = options[:project]

  result.reduce({}) do |a, e|
    e[:filters].map do |filter|
      a[filter[:label]] = project.labels(filter[:label]) unless a.key?(filter[:label])
    end
    a
  end
end

.create_lookups_cache(small_labels) ⇒ Object 



181
182
183
184
185
186
187
188
189
190
191
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 181

def self.create_lookups_cache(small_labels)
  small_labels.reduce({}) do |a, e|
    # The validElements API allow maximum paging with 10000 items
    lookup = e.values(:limit => 10_000).reduce({}) do |a1, e1|
      a1[e1[:value]] = e1[:uri]
      a1
    end
    a[e.uri] = lookup
    a
  end
end

.create_user_filter(expression, related) ⇒ Object

Encapuslates the creation of filter



297
298
299
300
301
302
303
304
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 297

def self.create_user_filter(expression, related)
  {
    related: related,
    level: :user,
    expression: expression,
    type: :filter
  }
end

.create_user_profile_mapping(filters, project_users, options = {}) ⇒ Object 



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
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 306

def self.(filters, project_users, options = {})
  domain = options[:domain]
  found_list = {}
  missing_list = []

  # Get the list of user login from filters
   = filters.flat_map do |filter|
    filter[:login]
  end

  # Then find user login in the users_brick_input
  users_brick_input = options[:users_brick_input]
  if users_brick_input&.any?
    users_brick_input.map do |user|
       << user.with_indifferent_access['login']
    end
  end

  .uniq.flat_map do ||
    user = project_users.find { |u| u. ==  }
    if user
      found_list[] = user.profile_url
    else
      missing_list << 
    end
  end
  # rubocop:disable Metrics/BlockNesting
  unless missing_list.empty? || domain.nil?
    if missing_list.size < 100
      missing_list.each do ||
        user = domain.()
        found_list[] = user.links['self'] if user
      end
    else
      if @all_domain_users[domain.name].nil?
        @mutex.lock
        if @all_domain_users[domain.name].nil?
          domain_users = domain.users
          @all_domain_users[domain.name] = domain_users
          GoodData.logger.info("action=lcm_get_domain_users domain=#{domain.name} number_users=#{domain_users.size} number_missing_users=#{missing_list.size} use_cache=false")
        else
          domain_users = @all_domain_users[domain.name]
          GoodData.logger.info("action=lcm_get_domain_users domain=##{domain.name} number_users=#{domain_users.size} number_missing_users=#{missing_list.size} use_cache=true")
        end
        @mutex.unlock
      else
        domain_users = @all_domain_users[domain.name]
        GoodData.logger.info("action=lcm_get_domain_users domain=##{domain.name} number_users=#{domain_users.size} number_missing_users=#{missing_list.size} use_cache=true")
      end

      missing_list.each do ||
        user = domain_users.find { |u| u. ==  }
        found_list[] = user.links['self'] if user
      end
    end
  end
  # rubocop:enable Metrics/BlockNesting
  found_list
end

.execute_mufs(user_filters, options = {}) ⇒ Object 



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
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 475

def self.execute_mufs(user_filters, options = {})
  client = options[:client]
  project = options[:project]
  ignore_missing_values = options[:ignore_missing_values]
  users_must_exist = options[:users_must_exist] == false ? false : true
  dry_run = options[:dry_run]
  project_log_formatter = GoodData::ProjectLogFormatter.new(project)

  project_users = project.users
  filters = normalize_filters(user_filters)
   = (filters, project_users, options)
  user_filters, errors = maqlify_filters(filters, , options.merge(users_must_exist: users_must_exist, type: :muf))
  if !ignore_missing_values && !errors.empty?
    errors = errors.map do |e|
      e.merge(pid: project.pid)
    end
    fail GoodData::FilterMaqlizationError, errors
  end

  filters = user_filters.map { |data| client.create(MandatoryUserFilter, data, project: project) }
  to_create, to_delete = resolve_user_filters(filters, project.data_permissions)

  to_delete = sanitize_filters_to_delete(to_delete, options[:users_brick_input], ) unless options[:no_sanitize]

  if options[:do_not_touch_filters_that_are_not_mentioned]
    GoodData.logger.warn("Data permissions computed: #{to_create.count} to create")
  else
    GoodData.logger.warn("Data permissions computed: #{to_create.count} to create and #{to_delete.count} to delete")
  end

  if dry_run
    GoodData.logger.warn('Option "dry_run" specified. No user filters will be altered!')
    create_results = to_create.map { |x| { status: 'dry_run', user: x.first, type: 'create' } }
    delete_results = to_delete.map { |x| { status: 'dry_run', user: x.first, type: 'delete' } }
    return { created: {},
             deleted: {},
             results: create_results + delete_results }
  end

  if to_create.empty?
    create_results = []
  else
    create_results = to_create.each_slice(100).flat_map do |batch|
      batch.pmapcat do |related_uri, group|
        group.each(&:save)
        res = client.get("/gdc/md/#{project.pid}/userfilters?users=#{related_uri}")
        items = res['userFilters']['items'].empty? ? [] : res['userFilters']['items'].first['userFilters']

        payload = {
          'userFilters' => {
            'items' => [{
              'user' => related_uri,
              'userFilters' => items.concat(group.map(&:uri))
            }]
          }
        }
        res = client.post("/gdc/md/#{project.pid}/userfilters", payload)

        # turn the errors from hashes into array of hashes
        update_result = res['userFiltersUpdateResult'].flat_map do |k, v|
          v.map { |r| { status: k.to_sym, user: r, type: :create } }
        end

        update_result.map do |result|
          result[:status] == :failed ? result.merge(GoodData::Helpers.symbolize_keys(result[:user])) : result
        end
      end
    end
    project_log_formatter.log_user_filter_results(create_results, to_create)
    create_errors = create_results.select { |r| r[:status] == :failed }
    fail "Creating MUFs resulted in errors: #{create_errors}" if create_errors.any?
  end

  if to_delete.empty?
    delete_results = []
  elsif !options[:do_not_touch_filters_that_are_not_mentioned]
    delete_results = to_delete.each_slice(100).flat_map do |batch|
      batch.flat_map do |related_uri, group|
        results = []
        if related_uri
          res = client.get("/gdc/md/#{project.pid}/userfilters?users=#{related_uri}")
          items = res['userFilters']['items'].empty? ? [] : res['userFilters']['items'].first['userFilters']
          payload = {
            'userFilters' => {
              'items' => [
                {
                  'user' => related_uri,
                  'userFilters' => items - group.map(&:uri)
                }
              ]
            }
          }
          res = client.post("/gdc/md/#{project.pid}/userfilters", payload)
          results.concat(res['userFiltersUpdateResult']
                             .flat_map { |k, v| v.map { |r| { status: k.to_sym, user: r, type: :delete } } }
                             .map { |result| result[:status] == :failed ? result.merge(GoodData::Helpers.symbolize_keys(result[:user])) : result })
        end
        group.peach(&:delete)
        results
      end

      project_log_formatter.log_user_filter_results(delete_results, to_delete)
      delete_errors = delete_results.select { |r| r[:status] == :failed } if delete_results
      fail "Deleting MUFs resulted in errors: #{delete_errors}" if delete_errors&.any?
    end
  end

  { created: to_create, deleted: to_delete, results: create_results + (delete_results || []) }
end

.execute_variables(filters, var, options = {}) ⇒ Array

Executes the update for variables. It resolves what is new and needed to update.

Parameters:

  • filters (Array<Hash>) —

    Filter Definitions

  • filters (Variable) —

    Variable instance to be updated

  • options (Hash) (defaults to: {})

Options Hash (options):

  • :dry_run (Boolean) —

    If dry run is true. No changes to he proejct are made but list of changes is provided

Returns:

  • (Array) —

    list of filters that needs to be created and deleted



456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 456

def self.execute_variables(filters, var, options = {})
  client = options[:client]
  project = options[:project]
  dry_run = options[:dry_run]
  to_create, to_delete = execute(filters, var.user_values, VariableUserFilter, options.merge(type: :variable))
  return [to_create, to_delete] if dry_run

  # TODO: get values that are about to be deleted and created and update them.
  # This will make sure there is no downitme in filter existence
  unless options[:do_not_touch_filters_that_are_not_mentioned]
    to_delete.each { |_, group| group.each(&:delete) }
  end
  data = to_create.values.flatten.map(&:to_hash).map { |var_val| var_val.merge(prompt: var.uri) }
  data.each_slice(200) do |slice|
    client.post("/gdc/md/#{project.obj_id}/variables/user", :variables => slice)
  end
  [to_create, to_delete]
end

.get_filters(file, options = {}) ⇒ Boolean

Main Entry function. Gets values and processes them to get filters that are suitable for other function to process. Values can be read from file or provided inline as an array. The results are then preprocessed. It is possible to provide multiple values for an attribute tries to deduplicate the values if they are not unique. Allows for setting over/to filters and allows for setting up filters from multiple columns. It is specially designed so many aspects of configuration are modifiable so you do have to preprocess the data as little as possible ideally you should be able to use data that came directly from the source system and that are intended for use in other parts of ETL.

Parameters:

  • options (Hash) (defaults to: {})

Returns:

  • (Boolean) 


39
40
41
42
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 39

def self.get_filters(file, options = {})
  values = get_values(file, options)
  reduce_results(values)
end

.get_missing_users(filters, options = {}) ⇒ Object 



144
145
146
147
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 144

def self.get_missing_users(filters, options = {})
  users_cache = options[:users_cache]
  filters.reject { |u| users_cache.key?(u[:login]) }
end

.get_small_labels(labels_cache) ⇒ Object

Walks over provided labels and picks those that have fewer than certain amount of values This tries to balance for speed when working with small datasets (like users) so it precaches the values and still be able to function for larger ones even though that would mean tons of requests



224
225
226
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 224

def self.get_small_labels(labels_cache)
  labels_cache.values.select { |label| label &.values_count &. < 100_000 }
end

.maqlify_filters(filters, user_profile_mapping, options = {}) ⇒ Array

Resolves and creates maql statements from filter definitions. This method does not perform any modifications on API but collects all the information that is needed to do so. Method collects all info from the user and current state in project and compares. Returns suggestion of what should be deleted and what should be created If there is some discrepancies in the data (missing values, nonexistent users) it finishes and collects all the errors at once

Parameters:

  • filters (Array<Hash>) —

    Filters definition

Returns:

  • (Array) —

    first is list of MAQL statements



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
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 376

def self.maqlify_filters(filters, , options = {})
  fail_early = options[:fail_early] == false ? false : true
  users_cache = options[:users_cache]
  labels_cache = create_label_cache(filters, options)
  small_labels = get_small_labels(labels_cache)
  lookups_cache = create_lookups_cache(small_labels)
  attrs_cache = create_attrs_cache(filters, options)
  create_filter_proc = proc do |, f|
    expression, errors = create_expression(f, labels_cache, lookups_cache, attrs_cache, options, )
     = .downcase
    profiles_uri = if options[:type] == :muf
                     [].nil? ? ('/gdc/account/profile/' + ) : []
                   elsif options[:type] == :variable
                     (users_cache[] && users_cache[].uri)
                   else
                     fail 'Unsuported type in maqlify_filters.'
                   end

    if profiles_uri && expression && expression != 'TRUE'
      [create_user_filter(expression, profiles_uri)] + errors
    else
      [] + errors
    end
  end

  # if fail early process until first error
  results = if fail_early
              x = filters.inject([true, []]) do |(enough, a), e|
                 = e[:login]
                if enough
                  y = e[:filters].pmapcat { |f| create_filter_proc.call(, f) }
                  [!y.any? { |r| r[:type] == :error }, a.concat(y)]
                else
                  [false, a]
                end
              end
              x.last
            else
              filters.flat_map do |filter|
                 = filter[:login]
                filter[:filters].pmapcat { |f| create_filter_proc.call(, f) }
              end
            end
  results.group_by { |i| i[:type] }.values_at(:filter, :error).map { |i| i || [] }
end

.process_line(line, options = {}) ⇒ Object

Processes a line from source file. It is processed in 2 formats. First mode is column_based. It means getting all specific columns. These are specified either by index or name. Multiple values are provided by several rows for the same user

Second mode is row based which means there are no headers and number of columns can be variable. Each row specifies multiple values for one user. It is implied that the file provides values for just one label

Parameters:

  • options (Hash) (defaults to: {})

Returns:



93
94
95
96
97
98
99
100
101
102
103
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 93

def self.process_line(line, options = {})
  index = options[:user_column] || 0
   = line[index]

  results = options[:labels].mapcat do |label|
    column = label[:column] || Range.new(1, -1)
    values = column.is_a?(Range) ? line.slice(column) : [line[column]]
    [create_filter(label, values.compact)]
  end
  [, results]
end

.read_data_with_header(file, memo, options) ⇒ Object 



71
72
73
74
75
76
77
78
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 71

def self.read_data_with_header(file, memo, options)
  CSV.foreach(file, headers: true, return_headers: false) do |row|
    key, data = process_line(row, options)
    memo[key] = [] unless memo.key?(key)
    memo[key].concat(data)
  end
  memo
end

.read_data_without_header(file, memo, options) ⇒ Object 



63
64
65
66
67
68
69
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 63

def self.read_data_without_header(file, memo, options)
  CSV.foreach(file, headers: false, return_headers: false) do |row|
    key, data = process_line(row, options)
    memo[key] = [] unless memo.key?(key)
    memo[key].concat(data)
  end
end

.read_file(file, options = {}) ⇒ Object 



53
54
55
56
57
58
59
60
61
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 53

def self.read_file(file, options = {})
  memo = Hash[]
  if row_based?(options)
    read_data_without_header(file, memo, options)
  else
    read_data_with_header(file, memo, options)
  end
  memo
end

.reduce_results(data) ⇒ Array

Processes values in a map reduce way so the result is as readable as possible and poses minimal impact on the API

Parameters:

  • options (Hash)

Returns:

  • (Array) 


119
120
121
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 119

def self.reduce_results(data)
  data.map { |k, v| { login: k, filters: UserFilterBuilder.collect_labels(v) } }
end

.resolve_user_filter(user = [], project = []) ⇒ Object 



422
423
424
425
426
427
428
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 422

def self.resolve_user_filter(user = [], project = [])
  user ||= []
  project ||= []
  to_create = user - project
  to_delete = project - user
  { :create => to_create, :delete => to_delete }
end

.resolve_user_filters(user_filters, vals) ⇒ Object

Gets user defined filters and values from project regardless if they come from Mandatory Filters or Variable filters and tries to resolve what needs to be removed an what needs to be updated



433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 433

def self.resolve_user_filters(user_filters, vals)
  project_vals_lookup = vals.group_by(&:related_uri)
  user_vals_lookup = user_filters.group_by(&:related_uri)

  a = vals.map(&:related_uri)
  b = user_filters.map(&:related_uri)

  users_to_try = (a + b).uniq
  results = users_to_try.map do |user|
    resolve_user_filter(user_vals_lookup[user], project_vals_lookup[user])
  end

  to_create = results.map { |x| x[:create] }.flatten.group_by(&:related_uri)
  to_delete = results.map { |x| x[:delete] }.flatten.group_by(&:related_uri)
  [to_create, to_delete]
end

.row_based?(options = {}) ⇒ Boolean

Function that tells you if the file should be read line_wise. This happens if you have only one label defined and you do not have columns specified

Parameters:

  • options (Hash) (defaults to: {})

Returns:

  • (Boolean) 


49
50
51
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 49

def self.row_based?(options = {})
  options[:labels].count == 1 && !options[:labels].first.key?(:column)
end

.verify_existing_users(filters, options = {}) ⇒ Object 



149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
 
# File 'lib/gooddata/models/user_filters/user_filter_builder.rb', line 149

def self.verify_existing_users(filters, options = {})
  users_must_exist = options[:users_must_exist] == false ? false : true
  users_cache = options[:users_cache]
  domain = options[:domain]

  if users_must_exist
    missing_users = filters.reject do |u|
      next true if users_cache.key?(u[:login])
      domain_user = (domain && domain.(u[:login]))
      users_cache[domain_user.] = domain_user if domain_user
      next true if domain_user
      false
    end
    unless missing_users.empty?
      fail "#{missing_users.count} users are not part of the project and " \
           "variable cannot be resolved since :users_must_exist is set " \
           "to true (#{missing_users.join(', ')})"
    end
  end
end