Module: Sparkql::ParserTools

Included in:
Parser
Defined in:
lib/sparkql/parser_tools.rb

Constant Summary collapse

DATE_TYPES =

Coercible types from highest precision to lowest

[:datetime, :date]
NUMBER_TYPES = 
[:decimal, :integer]
ARITHMETIC_TYPES = 
[:decimal, :integer, :field, :arithmetic]
GROUP = 
'Group'.freeze
NEGATION = 
'Negation'.freeze

Instance Method Summary collapse

Instance Method Details

#add_fold(n1, n2) ⇒ Object 



270
271
272
273
274
275
 
# File 'lib/sparkql/parser_tools.rb', line 270

def add_fold(n1, n2)
  return if arithmetic_error?(n1) || arithmetic_error?(n2)

  value = escape_arithmetic_value(n1) + escape_arithmetic_value(n2)
  { type: arithmetic_type(n1, n2), value: unescape_arithmetic(value) }
end

#arithmetic_error?(side) ⇒ Boolean

Returns:

  • (Boolean) 


254
255
256
257
258
259
260
261
262
 
# File 'lib/sparkql/parser_tools.rb', line 254

def arithmetic_error?(side)
  side_type = side[:type] == :function ? side[:return_type] : side[:type]
  return false unless (!ARITHMETIC_TYPES.include?(side_type) || !ARITHMETIC_TYPES.include?(side_type))

  compile_error(:token => side[:value], :expression => side,
        :message => "Error attempting arithmetic with type: #{side_type}",
        :status => :fatal, :syntax => false, :constraint => true )
  true
end

#arithmetic_field(nested_representation) ⇒ Object 



30
31
32
33
34
35
36
37
38
39
40
41
 
# File 'lib/sparkql/parser_tools.rb', line 30

def arithmetic_field(nested_representation)
  return if nested_representation.nil?

  return nested_representation[:value] if nested_representation[:type] == :field
  return nested_representation[:field] if nested_representation.key?(:field)

  field = arithmetic_field(nested_representation[:lhs])
  return field unless field.nil?

  field = arithmetic_field(nested_representation[:rhs])
  return field unless field.nil?
end

#arithmetic_type(num1, num2) ⇒ Object 



309
310
311
312
313
314
315
 
# File 'lib/sparkql/parser_tools.rb', line 309

def arithmetic_type(num1, num2)
  if (num1[:type] == :decimal || num2[:type] == :decimal)
    :decimal
  else
    :integer
  end
end

#coercible_types(type1, type2) ⇒ Object

If both types support coercion with eachother, always selects the highest precision type to return as a reflection of the two. Any type that doesn’t support coercion with the other type returns nil



399
400
401
402
403
404
405
406
407
 
# File 'lib/sparkql/parser_tools.rb', line 399

def coercible_types type1, type2
  if DATE_TYPES.include?(type1) && DATE_TYPES.include?(type2)
    DATE_TYPES.first
  elsif NUMBER_TYPES.include?(type1) && NUMBER_TYPES.include?(type2)
    NUMBER_TYPES.first
  else
    nil
  end
end

#div_fold(n1, n2) ⇒ Object 



291
292
293
294
295
296
297
298
 
# File 'lib/sparkql/parser_tools.rb', line 291

def div_fold(n1, n2)
  return if arithmetic_error?(n1) ||
    arithmetic_error?(n2) ||
    zero_error?(n2)

  value = escape_arithmetic_value(n1) / escape_arithmetic_value(n2)
  { type: arithmetic_type(n1, n2), value: unescape_arithmetic(value) }
end

#escape_arithmetic_value(expression) ⇒ Object 



317
318
319
320
321
322
323
324
 
# File 'lib/sparkql/parser_tools.rb', line 317

def escape_arithmetic_value(expression)
  case expression[:type]
  when :decimal
    BigDecimal.new(expression[:value])
  else
    escape_value(expression)
  end
end

#group_fold(exp) ⇒ Object 



264
265
266
267
268
 
# File 'lib/sparkql/parser_tools.rb', line 264

def group_fold(exp)
  @lexer.leveldown
  @lexer.block_group_identifier -= 1
  exp
end

#mod_fold(n1, n2) ⇒ Object 



300
301
302
303
304
305
306
307
 
# File 'lib/sparkql/parser_tools.rb', line 300

def mod_fold(n1, n2)
  return if arithmetic_error?(n1) ||
    arithmetic_error?(n2) ||
    zero_error?(n2)

  value = escape_arithmetic_value(n1) % escape_arithmetic_value(n2)
  { type: arithmetic_type(n1, n2), value: unescape_arithmetic(value) }
end

#mul_fold(n1, n2) ⇒ Object 



284
285
286
287
288
289
 
# File 'lib/sparkql/parser_tools.rb', line 284

def mul_fold(n1, n2)
  return if arithmetic_error?(n1) || arithmetic_error?(n2)

  value = escape_arithmetic_value(n1) * escape_arithmetic_value(n2)
  { type: arithmetic_type(n1, n2), value: unescape_arithmetic(value) }
end

#nested_function_depth(expression) ⇒ Object 



409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
 
# File 'lib/sparkql/parser_tools.rb', line 409

def nested_function_depth(expression)
  return 0 unless expression && expression[:type] == :function

  height = 0
  queue = []
  queue.push(expression)

  while true
    count = queue.size
    return height if count == 0

    height += 1

    while count > 0
      node = queue.shift
      node[:args].each do |child|
        queue.push(child) if child[:type] == :function
      end
      count -= 1
    end
  end
end

#next_token ⇒ Object 



22
23
24
25
26
27
28
 
# File 'lib/sparkql/parser_tools.rb', line 22

def next_token
  t = @lexer.shift
  while t[0] == :SPACE or t[0] == :NEWLINE
    t = @lexer.shift
  end
  t
end

#no_field_error(field, operator) ⇒ Object 



43
44
45
46
47
 
# File 'lib/sparkql/parser_tools.rb', line 43

def no_field_error(field, operator)
  tokenizer_error(:token => field,
                  :expression => {operator: operator, conjuction: 'And', conjunction_level: 0, level: @lexer.level},
                  :message => "Each expression must evaluate a field", :status => :fatal )
end

#on_error(error_token_id, error_value, value_stack) ⇒ Object 



342
343
344
345
346
347
348
349
 
# File 'lib/sparkql/parser_tools.rb', line 342

def on_error(error_token_id, error_value, value_stack)
  token_name = token_to_str(error_token_id)
  token_name.downcase!
  tokenizer_error(:token => @lexer.current_token_value, 
                  :message => "Error parsing token #{token_name}",
                  :status => :fatal, 
                  :syntax => true)    
end

#parse(str) ⇒ Object 



13
14
15
16
17
18
19
20
 
# File 'lib/sparkql/parser_tools.rb', line 13

def parse(str)
  @lexer = Sparkql::Lexer.new(str)
  @expression_count = 0
  results = do_parse
  return if results.nil?
  validate_expressions results
  results
end

#sub_fold(n1, n2) ⇒ Object 



277
278
279
280
281
282
 
# File 'lib/sparkql/parser_tools.rb', line 277

def sub_fold(n1, n2)
  return if arithmetic_error?(n1) || arithmetic_error?(n2)

  value = escape_arithmetic_value(n1) - escape_arithmetic_value(n2)
  { type: arithmetic_type(n1, n2), value: unescape_arithmetic(value) }
end

#tokenize_arithmetic(lhs, operator, rhs) ⇒ Object 



240
241
242
243
244
245
246
247
248
249
250
251
252
 
# File 'lib/sparkql/parser_tools.rb', line 240

def tokenize_arithmetic(lhs, operator, rhs)
  lhs = {type: :field, value: lhs} if lhs.is_a?(String)
  rhs = {type: :field, value: rhs} if rhs.is_a?(String)

  arithmetic_error?(lhs)
  arithmetic_error?(rhs)
  {
    type: :arithmetic,
    op: operator,
    lhs: lhs,
    rhs: rhs
  }
end

#tokenize_arithmetic_group(lhs) ⇒ Object 



128
129
130
131
132
133
134
135
136
137
 
# File 'lib/sparkql/parser_tools.rb', line 128

def tokenize_arithmetic_group(lhs)
  @lexer.leveldown
  @lexer.block_group_identifier -= 1
  lhs = {type: :field, value: lhs} if lhs.is_a?(String)
  {
    type: :arithmetic,
    op: GROUP,
    lhs: lhs
  }
end

#tokenize_arithmetic_negation(lhs) ⇒ Object 



139
140
141
142
143
144
145
146
 
# File 'lib/sparkql/parser_tools.rb', line 139

def tokenize_arithmetic_negation(lhs)
  lhs = {type: :field, value: lhs} if lhs.is_a?(String)
  {
    type: :arithmetic,
    op: NEGATION,
    lhs: lhs
  }
end

#tokenize_conjunction(exp1, conj, exp2) ⇒ Object 



99
100
101
102
103
 
# File 'lib/sparkql/parser_tools.rb', line 99

def tokenize_conjunction(exp1, conj, exp2)
  exp2.first[:conjunction] = conj
  exp2.first[:conjunction_level] = @lexer.level
  exp1 + exp2
end

#tokenize_expression(field, op, val) ⇒ Object 



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
 
# File 'lib/sparkql/parser_tools.rb', line 49

def tokenize_expression(field, op, val)
  operator = get_operator(val,op) unless val.nil?

  field_manipulations = nil
  if field.is_a?(Hash) && field[:type] == :function
    function = Sparkql::FunctionResolver::SUPPORTED_FUNCTIONS[field[:function_name].to_sym]
    if function.nil?
      tokenizer_error(:token => field[:function_name],
        :message => "Unsupported function type", :status => :fatal )
    end
    field_manipulations = field
    field = field[:field]
  elsif field.is_a?(Hash) && field[:type] == :arithmetic
    field_manipulations = field
    field = arithmetic_field(field)
    no_field_error(field, operator) if field.nil?
  elsif field.is_a?(Hash)
    no_field_error(field, operator)
  end

  custom_field = !field.nil? && field.is_a?(String) && field.start_with?('"')

  block_group = (@lexer.level == 0) ? 0 : @lexer.block_group_identifier
  expression = {:field => field, :operator => operator, :conjunction => 'And',
    :conjunction_level => 0, :level => @lexer.level,
    :block_group => block_group, :custom_field => custom_field}

  if !field_manipulations.nil?
    # Keeping field_function and field_function_type for backward compatibility with datacon
    expression.merge!(field_manipulations: field_manipulations)

    if field_manipulations[:type] == :function
      expression.merge!(field_function: field_manipulations[:function_name],
                        field_function_type: field_manipulations[:return_type],
                        args: field_manipulations[:function_parameters])
    end
  end

  expression = val.merge(expression) unless val.nil?
  expression[:condition] ||= expression[:value]
  validate_level_depth expression
  validate_field_function_depth(expression[:field_manipulations])
  if operator.nil?
    tokenizer_error(:token => op, :expression => expression,
      :message => "Operator not supported for this type and value string", :status => :fatal )
  end
  @expression_count += 1
  [expression]
end

#tokenize_field_arg(field) ⇒ Object 



203
204
205
206
207
208
209
210
211
212
 
# File 'lib/sparkql/parser_tools.rb', line 203

def tokenize_field_arg(field)
  if field.is_a?(String)
    {
      :type => :field,
      :value => field,
    }
  else
    field
  end
end

#tokenize_function(name, f_args) ⇒ Object 



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
 
# File 'lib/sparkql/parser_tools.rb', line 214

def tokenize_function(name, f_args)
  @lexer.leveldown
  @lexer.block_group_identifier -= 1

  args = f_args.instance_of?(Array) ? f_args : [f_args]
  validate_multiple_arguments args
  condition_list = []
  args.each do |arg|
    condition_list << arg[:value] # Needs to be pure string value
    arg[:value] = escape_value(arg)
  end
  resolver = Sparkql::FunctionResolver.new(name, args)
  
  resolver.validate
  if(resolver.errors?)
    tokenizer_error(:token => @lexer.last_field, 
                    :message => "Error parsing function #{resolver.errors.join(',')}",
                    :status => :fatal, 
                    :syntax => true)    
    return nil
  else
    result = resolver.call()
    result.nil? ? result : result.merge(:condition => "#{name}(#{condition_list.join(',')})")
  end
end

#tokenize_function_args(lit1, lit2) ⇒ Object 



197
198
199
200
201
 
# File 'lib/sparkql/parser_tools.rb', line 197

def tokenize_function_args(lit1, lit2)
  array = lit1.kind_of?(Array) ? lit1 : [lit1]
  array << lit2
  array
end

#tokenize_group(expressions) ⇒ Object 



123
124
125
126
 
# File 'lib/sparkql/parser_tools.rb', line 123

def tokenize_group(expressions)
  @lexer.leveldown
  expressions
end

#tokenize_list(list) ⇒ Object 



148
149
150
151
152
153
 
# File 'lib/sparkql/parser_tools.rb', line 148

def tokenize_list(list)
  return if list.nil?
  validate_multiple_values list[:value]
  list[:condition] ||= list[:value]
  list
end

#tokenize_literal_negation(number_token) ⇒ Object 



155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
 
# File 'lib/sparkql/parser_tools.rb', line 155

def tokenize_literal_negation(number_token)
  old_val = case number_token[:type]
  when :integer
    number_token[:value].to_i
  when :decimal
    number_token[:value].to_f
  else
    tokenizer_error(:token => @lexer.current_token_value,
                    :expression => number_token,
                    :message => "Negation is only allowed for integer and floats",
                    :status => :fatal,
                    :syntax => true)
    return number_token
  end
  number_token[:value] = (-1 * old_val).to_s

  number_token
end

#tokenize_multiple(lit1, lit2) ⇒ Object 



174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
 
# File 'lib/sparkql/parser_tools.rb', line 174

def tokenize_multiple(lit1, lit2)
  final_type = lit1[:type]
  if lit1[:type] != lit2[:type]
    final_type = coercible_types(lit1[:type],lit2[:type])
    if final_type.nil?
      final_type = lit1[:type]
      tokenizer_error(:token => @lexer.last_field, 
                      :message => "Type mismatch in field list.",
                      :status => :fatal, 
                      :syntax => true)
    end
  end
  array = Array(lit1[:value])
  condition = lit1[:condition] || lit1[:value] 
  array << lit2[:value]
  {
    :type => final_type ,
    :value => array,
    :multiple => "true",
    :condition => condition + "," + (lit2[:condition] || lit2[:value])
  }
end

#tokenize_unary_conjunction(conj, exp) ⇒ Object 



105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
 
# File 'lib/sparkql/parser_tools.rb', line 105

def tokenize_unary_conjunction(conj, exp)
  # Handles the case when a SparkQL filter string
  # begins with a unary operator, and is nested, such as:
  #   Not (Not Field Eq 1)
  # In this instance we treat the outer unary as a conjunction. With any other
  # expression this would be the case, so that should make processing 
  # consistent.
  if exp.first[:unary] && @lexer.level == 0
    exp.first[:conjunction] =  conj
    exp.first[:conjunction_level] = @lexer.level
  else
    exp.first[:unary] = conj
    exp.first[:unary_level] = @lexer.level
  end

  exp
end

#unescape_arithmetic(value) ⇒ Object 



326
327
328
329
330
331
332
 
# File 'lib/sparkql/parser_tools.rb', line 326

def unescape_arithmetic(value)
  if value.is_a?(BigDecimal)
    value.round(20).to_s('F')
  else
    value.to_s
  end
end

#validate_expressions(results) ⇒ Object 



367
368
369
370
371
372
373
374
 
# File 'lib/sparkql/parser_tools.rb', line 367

def validate_expressions results
  if results.size > max_expressions 
    compile_error(:token => results[max_expressions][:field], :expression => results[max_expressions],
          :message => "You have exceeded the maximum expression count.  Please limit to no more than #{max_expressions} expressions in a filter.",
          :status => :fatal, :syntax => false, :constraint => true )
    results.slice!(max_expressions..-1)
  end
end

#validate_field_function_depth(expression) ⇒ Object 



359
360
361
362
363
364
365
 
# File 'lib/sparkql/parser_tools.rb', line 359

def validate_field_function_depth(expression)
  if nested_function_depth(expression) > max_function_depth
    compile_error(:token => "(", :expression => expression,
          :message => "You have exceeded the maximum function nesting level.  Please nest no more than #{max_function_depth} levels deep.",
          :status => :fatal, :syntax => false, :constraint => true )
  end
end

#validate_level_depth(expression) ⇒ Object 



351
352
353
354
355
356
357
 
# File 'lib/sparkql/parser_tools.rb', line 351

def validate_level_depth expression
  if @lexer.level > max_level_depth
    compile_error(:token => "(", :expression => expression,
          :message => "You have exceeded the maximum nesting level.  Please nest no more than #{max_level_depth} levels deep.",
          :status => :fatal, :syntax => false, :constraint => true )
  end
end

#validate_multiple_arguments(args) ⇒ Object 



386
387
388
389
390
391
392
393
394
 
# File 'lib/sparkql/parser_tools.rb', line 386

def validate_multiple_arguments args
  args = Array(args)
  if args.size > max_values 
    compile_error(:token => args[max_values],
          :message => "You have exceeded the maximum parameter count.  Please limit to #{max_values} parameters to a single function.",
          :status => :fatal, :syntax => false, :constraint => true )
    args.slice!(max_values..-1)
  end
end

#validate_multiple_values(values) ⇒ Object 



376
377
378
379
380
381
382
383
384
 
# File 'lib/sparkql/parser_tools.rb', line 376

def validate_multiple_values values
  values = Array(values)
  if values.size > max_values 
    compile_error(:token => values[max_values],
          :message => "You have exceeded the maximum value count.  Please limit to #{max_values} values in a single expression.",
          :status => :fatal, :syntax => false, :constraint => true )
    values.slice!(max_values..-1)
  end
end

#zero_error?(number) ⇒ Boolean

Returns:

  • (Boolean) 


334
335
336
337
338
339
340
 
# File 'lib/sparkql/parser_tools.rb', line 334

def zero_error?(number)
  return unless escape_value(number) == 0

  compile_error(:token => "#{number[:value]}", :expression => number,
        :message => "Error attempting to divide by zero",
        :status => :fatal, :syntax => false, :constraint => true )
end