Module: Sparkql::ParserTools

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

Overview

This is the guts of the parser internals and is mixed into the parser for organization.

Constant Summary collapse

DATE_TYPES =

Coercible types from highest precision to lowest

[:datetime, :date]
NUMBER_TYPES = 
[:decimal, :integer]

Instance Method Summary collapse

Instance Method Details

#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



230
231
232
233
234
235
236
237
238
 
# File 'lib/sparkql/parser_tools.rb', line 230

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

#next_token ⇒ Object 



17
18
19
20
21
22
23
 
# File 'lib/sparkql/parser_tools.rb', line 17

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

#on_error(error_token_id, error_value, value_stack) ⇒ Object 



180
181
182
183
184
185
186
187
188
 
# File 'lib/sparkql/parser_tools.rb', line 180

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

#parse(str) ⇒ Object 



8
9
10
11
12
13
14
15
 
# File 'lib/sparkql/parser_tools.rb', line 8

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

#tokenize_conjunction(exp1, conj, exp2) ⇒ Object 



64
65
66
67
68
 
# File 'lib/sparkql/parser_tools.rb', line 64

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 



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
 
# File 'lib/sparkql/parser_tools.rb', line 25

def tokenize_expression(field, op, val)
  operator = get_operator(val,op) unless val.nil?
  field_args = {}
  # Function support for fields is stapled in here. The function information
  # is remapped to the expression
  if field.is_a?(Hash) && field[:type] == :function
    function = Sparkql::FunctionResolver::SUPPORTED_FUNCTIONS[field[:value].to_sym]
    if !function.nil?
      field_args[:field_function] = field[:value]
      field_args[:args] = field[:args]

      if field_args[:field_function] == 'cast'
        field_args[:field_function_type] = field[:args].last.to_sym
      else
        field_args[:field_function_type] = function[:return_type]
      end
    else
      tokenizer_error(:token => field[:value], 
        :message => "Unsupported function type", :status => :fatal )
    end
    field = field[:args].first
  end
  custom_field = !field.nil? && 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}.
    merge!(field_args)
  expression = val.merge(expression) unless val.nil?
  expression[:condition] ||= expression[:value]
  validate_level_depth expression
  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 



147
148
149
150
151
152
 
# File 'lib/sparkql/parser_tools.rb', line 147

def tokenize_field_arg(field)
  {
    :type => :field,
    :value => field,
  }
end

#tokenize_function(name, f_args) ⇒ Object 



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
 
# File 'lib/sparkql/parser_tools.rb', line 154

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 



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

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

#tokenize_group(expressions) ⇒ Object 



88
89
90
91
 
# File 'lib/sparkql/parser_tools.rb', line 88

def tokenize_group(expressions)
  @lexer.leveldown
  expressions
end

#tokenize_list(list) ⇒ Object 



93
94
95
96
97
 
# File 'lib/sparkql/parser_tools.rb', line 93

def tokenize_list(list)
  validate_multiple_values list[:value]
  list[:condition] ||= list[:value]
  list
end

#tokenize_literal_negation(number_token) ⇒ Object 



99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
 
# File 'lib/sparkql/parser_tools.rb', line 99

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 



118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
 
# File 'lib/sparkql/parser_tools.rb', line 118

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 



70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
 
# File 'lib/sparkql/parser_tools.rb', line 70

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

#validate_expressions(results) ⇒ Object 



198
199
200
201
202
203
204
205
 
# File 'lib/sparkql/parser_tools.rb', line 198

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_level_depth(expression) ⇒ Object 



190
191
192
193
194
195
196
 
# File 'lib/sparkql/parser_tools.rb', line 190

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 



217
218
219
220
221
222
223
224
225
 
# File 'lib/sparkql/parser_tools.rb', line 217

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 



207
208
209
210
211
212
213
214
215
 
# File 'lib/sparkql/parser_tools.rb', line 207

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