Class: J2119::Matcher

Inherits:
Object
  • Object
show all
Defined in:
lib/j2119/matcher.rb

Overview

Does the heavy lifting of parsing j2119 files to extract all

the assertions, making egregious use of regular expressions.

This is the kind of thing I actively discourage when other

programmers suggest it.  If I were a real grown-up I'd
implement a proper lexer and bullet-proof parser.

Constant Summary collapse

MUST = 
'(?<modal>MUST|MAY|MUST NOT)'
TYPES = 
[
  'array',
  'object',
  'string',
  'boolean',
  'numeric',
  'integer',
  'float',
  'timestamp',
  'JSONPath',
  'referencePath',
  'URI'
]
RELATIONS = 
[
  '', 'equal to', 'greater than', 'less than',
  'greater than or equal to', 'less than or equal to'
].join('|')
RELATION = 
"((?<relation>#{RELATIONS})\\s+)"
S =

string

'"[^"]*"'
V =

non-string value: number, true, false, null

'\S+'
RELATIONAL = 
"#{RELATION}(?<target>#{S}|#{V})"
CHILD_ROLE = 
';\s+((its\s+(?<child_type>value))|' +
'each\s+(?<child_type>field|element))' +
'\s+is\s+an?\s+' +
'"(?<child_role>[^"]+)"'
@@initialized = 
false

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(root) ⇒ Matcher

Returns a new instance of Matcher.



147
148
149
150
151
152
 
# File 'lib/j2119/matcher.rb', line 147

def initialize(root)
  constants
  @roles = []
  add_role root
  reconstruct
end

Instance Attribute Details

#constraint_match ⇒ Object (readonly)

crutch for RE debugging



26
27
28
 
# File 'lib/j2119/matcher.rb', line 26

def constraint_match
  @constraint_match
end

#eachof_match ⇒ Object (readonly)

actual exports



29
30
31
 
# File 'lib/j2119/matcher.rb', line 29

def eachof_match
  @eachof_match
end

#only_one_match ⇒ Object (readonly)

crutch for RE debugging



26
27
28
 
# File 'lib/j2119/matcher.rb', line 26

def only_one_match
  @only_one_match
end

#role_matcher ⇒ Object (readonly)

actual exports



29
30
31
 
# File 'lib/j2119/matcher.rb', line 29

def role_matcher
  @role_matcher
end

#roledef_match ⇒ Object (readonly)

crutch for RE debugging



26
27
28
 
# File 'lib/j2119/matcher.rb', line 26

def roledef_match
  @roledef_match
end

Class Method Details

.tokenize_strings(s) ⇒ Object 



160
161
162
163
164
165
166
167
168
169
 
# File 'lib/j2119/matcher.rb', line 160

def self.tokenize_strings(s)
  # should be a way to do this with capture groups but I'm not smart enough
  strings = []
  r = Regexp.new '^[^"]*"([^"]*)"'
  while s =~ r
    strings << $1
    s = $'
  end
  strings
end

Instance Method Details

#add_role(role) ⇒ Object 



154
155
156
157
158
 
# File 'lib/j2119/matcher.rb', line 154

def add_role(role)
  @roles << role
  @role_matcher = @roles.join('|')
  reconstruct
end

#build(re, line) ⇒ Object 



205
206
207
208
209
210
211
212
213
214
215
 
# File 'lib/j2119/matcher.rb', line 205

def build(re, line)
  data = {}
  match = re.match(line)
  unless match
    puts "No names for: #{line}"
  end
  match.names.each do |name|
    data[name] = match[name]
  end
  data
end

#build_constraint(line) ⇒ Object 



229
230
231
 
# File 'lib/j2119/matcher.rb', line 229

def build_constraint(line)
  build(@constraint_match, line)
end

#build_only_one(line) ⇒ Object 



217
218
219
 
# File 'lib/j2119/matcher.rb', line 217

def build_only_one(line)
  build(@only_one_match, line)
end

#build_role_def(line) ⇒ Object 



201
202
203
 
# File 'lib/j2119/matcher.rb', line 201

def build_role_def(line)
  build(@roledef_match, line)
end

#constants ⇒ Object

constants that need help from oxford



64
65
66
67
68
69
70
71
72
73
 
# File 'lib/j2119/matcher.rb', line 64

def constants
  if !@@initialized
    @@initialized = true

    @@strings = Oxford.re(S, :capture_name => 'strings')
    enum = "one\s+of\s+#{@@strings}"

    @@predicate = "(#{RELATIONAL}|#{enum})"
  end
end

#is_constraint_line(line) ⇒ Object 



221
222
223
 
# File 'lib/j2119/matcher.rb', line 221

def is_constraint_line(line)
  line =~ @constraint_start
end

#is_only_one_match_line(line) ⇒ Object 



225
226
227
 
# File 'lib/j2119/matcher.rb', line 225

def is_only_one_match_line(line)
  line =~ @only_one_start
end

#is_role_def_line(line) ⇒ Object 



197
198
199
 
# File 'lib/j2119/matcher.rb', line 197

def is_role_def_line(line)
  line =~ %r{is\s+an?\s+"[^"]*"\.\s*$}
end

#make_type_regex ⇒ Object 



175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
 
# File 'lib/j2119/matcher.rb', line 175

def make_type_regex
  
  # add modified numeric types
  types = TYPES.clone
  number_types = [ 'float', 'integer', 'numeric' ]
  number_modifiers = [ 'positive', 'negative', 'nonnegative' ]
  number_types.each do |number_type|
    number_modifiers.each do |number_modifier|
      types << "#{number_modifier}-#{number_type}"
    end
  end
  
  # add array types
  array_types = types.map { |t| "#{t}-array" }
  types |= array_types
  nonempty_array_types = array_types.map { |t| "nonempty-#{t}" }
  types |= nonempty_array_types
  nullable_types = types.map { |t| "nullable-#{t}" }
  types |= nullable_types
  @type_regex = types.join('|')
end

#reconstruct ⇒ Object 



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
 
# File 'lib/j2119/matcher.rb', line 75

def reconstruct
  make_type_regex

  # conditional clause
  excluded_roles = "not\\s+" +
                   Oxford.re(@role_matcher,
                             :capture_name => 'excluded',
                             :use_article => true) +
                   "\\s+"
  conditional = "which\\s+is\\s+" +
                excluded_roles

  # regex for matching constraint lines
  c_start = '^An?\s+' +
            "(?<role>#{@role_matcher})" + '\s+' +
            "(#{conditional})?" +
            MUST + '\s+have\s+an?\s+'
  field_list = "one\\s+of\\s+" +
               Oxford.re('"[^"]+"', :capture_name => 'field_list')
  c_match = c_start + 
            "((?<type>#{@type_regex})\\s+)?" +
            "field\\s+named\\s+" +
            "((\"(?<field_name>[^\"]+)\")|(#{field_list}))" +
            '(\s+whose\s+value\s+MUST\s+be\s+' + @@predicate + ')?' +
            '(' + CHILD_ROLE + ')?' +
            '\.'

  # regexp for matching lines of the form
  #  "An X MUST have only one of "Y", "Z", and "W".
  #  There's a pattern here, building a separate regex rather than
  #  adding more complexity to @constraint_matcher.  Any further
  #  additions should be done this way, and
  #  TODO: Break @constraint_matcher into a bunch of smaller patterns
  #  like this.
  oo_start = '^An?\s+' +
            "(?<role>#{@role_matcher})" + '\s+' +
             MUST + '\s+have\s+only\s+'
  oo_field_list = "one\\s+of\\s+" +
                  Oxford.re('"[^"]+"',
                            :capture_name => 'field_list',
                            :connector => 'and')
  oo_match = oo_start + oo_field_list

  # regex for matching role-def lines
  val_match = "whose\\s+\"(?<fieldtomatch>[^\"]+)\"" +
              "\\s+field's\\s+value\\s+is\\s+" +
              "(?<valtomatch>(\"[^\"]*\")|([^\"\\s]\\S+))\\s+"
  with_a_match = "with\\s+an?\\s+\"(?<with_a_field>[^\"]+)\"\\s+field\\s"

  rd_match = '^An?\s+' +
             "(?<role>#{@role_matcher})" + '\s+' +
             "((?<val_match_present>#{val_match})|(#{with_a_match}))?" +
             "is\\s+an?\\s+" +
             "\"(?<newrole>[^\"]*)\"\\.\\s*$"
  @roledef_match = Regexp.new(rd_match)

  @constraint_start = Regexp.new(c_start)
  @constraint_match = Regexp.new(c_match)

  @only_one_start = Regexp.new(oo_start)
  @only_one_match = Regexp.new(oo_match)
                   
  eo_match = "^Each\\s+of\\s" +
             Oxford.re(@role_matcher,
                       :capture_name => 'each_of',
                       :use_article => true,
                       :connector => 'and') +
             "\\s+(?<trailer>.*)$"

  @eachof_match = Regexp.new(eo_match)
end

#tokenize_values(vals) ⇒ Object 



171
172
173
 
# File 'lib/j2119/matcher.rb', line 171

def tokenize_values(vals)
  vals.gsub(',', ' ').gsub('or', ' ').split(/\s+/)
end