Class: Openphar::Exporters::CoverageValidator

Inherits:
Object
  • Object
show all
Defined in:
lib/openphar/exporters/coverage_validator.rb

Overview

Validates that model coverage is 100% complete

Compares:

  1. All fields in JSON-LD files have corresponding model attributes
  2. All model attributes can be exported to JSON-LD
  3. All model attributes can be exported to Neo4j

Constant Summary collapse

STANDARD_KEYS =

Standard JSON-LD keys that are framework-related, not domain data

%w[
  @context @graph @id @type @vocab @base
].freeze
FIELD_MAPPING =

Mapping from JSON-LD camelCase to model snake_case

{
  "monographId" => "monograph_id",
  "prefLabel" => "pref_label",
  "altLabel" => "alt_label",
  "belongsToEdition" => "belongs_to_edition",
  "effectiveDate" => "effective_date",
  "referencesPreparation" => "references_preparation",
  "referencesTCMProfile" => "references_tcm_profile",
  "referencesAyurvedaProfile" => "references_ayurveda_profile",
  "referencesWesternProfile" => "references_western_profile",
  "referencesPlantSpecies" => "references_plant_species",
  "referencesPlantPart" => "references_plant_part",
  "testSpecification" => "test_specifications",
  "sameSubstanceAs" => "same_substance_as",
  "hasEquivalentIn" => "has_equivalent_in",
  "similarTo" => "similar_to",
  "storageContainer" => "storage_container",
  "storageConditions" => "storage_conditions",
  "botanicalSource" => "botanical_source",
  "macroscopicDescription" => "macroscopic_description",
  "microscopicDescription" => "microscopic_description",
  "molecularFormula" => "molecular_formula",
  "molecularWeight" => "molecular_weight",
  "casNumber" => "cas_number",
  "systematicName" => "systematic_name",
  "inchiKey" => "inchi_key",
  "smiles" => "smiles",
  "chemicalStructure" => "chemical_structure",
  "appearance" => "appearance",
  "solubility" => "solubility",
  "testName" => "test_name",
  "testType" => "test_type",
  "harmonizedMethod" => "harmonized_method",
  "publisherMethod" => "publisher_method",
  "testConditions" => "test_conditions",
  "limitType" => "limit_type",
  "limitValue" => "limit_value",
  "lowerLimit" => "lower_limit",
  "upperLimit" => "upper_limit",
  "limitUnit" => "limit_unit",
  "assayTarget" => "assay_target",
  "assayExpression" => "assay_expression",
  "impurityType" => "impurity_type",
  "physicalProperty" => "physical_property"
}.freeze
REVERSE_MAPPING =

Reverse mapping: snake_case to camelCase

FIELD_MAPPING.invert

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize ⇒ CoverageValidator

Returns a new instance of CoverageValidator.



72
73
74
75
76
 
# File 'lib/openphar/exporters/coverage_validator.rb', line 72

def initialize
  @results = {}
  @missing_fields = []
  @extra_fields = []
end

Instance Attribute Details

#extra_fields ⇒ Object (readonly)

Returns the value of attribute extra_fields.



70
71
72
 
# File 'lib/openphar/exporters/coverage_validator.rb', line 70

def extra_fields
  @extra_fields
end

#missing_fields ⇒ Object (readonly)

Returns the value of attribute missing_fields.



70
71
72
 
# File 'lib/openphar/exporters/coverage_validator.rb', line 70

def missing_fields
  @missing_fields
end

#results ⇒ Object (readonly)

Returns the value of attribute results.



70
71
72
 
# File 'lib/openphar/exporters/coverage_validator.rb', line 70

def results
  @results
end

Instance Method Details

#normalize_key(key) ⇒ Object

Normalize a key from JSON-LD to snake_case



123
124
125
126
127
128
129
 
# File 'lib/openphar/exporters/coverage_validator.rb', line 123

def normalize_key(key)
  # Remove nested path (e.g., "prefLabel.en" -> "prefLabel")
  base_key = key.split(".").first

  # Map to snake_case if known
  FIELD_MAPPING[base_key] || base_key
end

#print_report ⇒ Object

Print detailed report



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
 
# File 'lib/openphar/exporters/coverage_validator.rb', line 172

def print_report
  puts "=" * 60
  puts "COVERAGE VALIDATION REPORT"
  puts "=" * 60

  summary = self.summary

  puts "\nFiles Validated: #{summary[:files_validated]}"
  puts "Total JSON-LD Fields: #{summary[:total_jsonld_fields]}"
  puts "Total Expected Fields: #{summary[:total_expected_fields]}"
  puts "Average Coverage: #{summary[:average_coverage].round(2)}%"

  if summary[:unique_missing_fields].any?
    puts "\n" + "!" * 60
    puts "MISSING FIELDS (in JSON-LD but not in model):"
    puts "!" * 60
    summary[:unique_missing_fields].each do |field|
      puts "  - #{field}"
    end
  end

  if summary[:unique_extra_fields].any?
    puts "\n" + "+" * 60
    puts "EXTRA FIELDS (in model but not in JSON-LD):"
    puts "+" * 60
    summary[:unique_extra_fields].each do |field|
      puts "  + #{field}"
    end
  end

  puts "\n" + "=" * 60
  puts "COVERAGE BY FILE:"
  puts "=" * 60

  @results.each do |file, result|
    status = result[:coverage] >= 100 ? "✓" : "✗"
    puts "#{status} #{File.basename(file)}: #{result[:coverage].round(2)}%"
  end
end

#summary ⇒ Hash

Generate report of all unique missing fields across all files

Returns:

  • (Hash) —

    Report of coverage



147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
 
# File 'lib/openphar/exporters/coverage_validator.rb', line 147

def summary
  unique_missing = @missing_fields.uniq
  unique_extra = @extra_fields.uniq

  total_jsonld_keys = @results.values.sum { |r| r[:jsonld_keys] }
  total_expected = @results.values.sum { |r| r[:expected] }

  avg_coverage = if @results.any?
                   @results.values.map { |r| r[:coverage] }.sum / @results.size
                 else
                   100
                 end

  {
    files_validated: @results.size,
    total_jsonld_fields: total_jsonld_keys,
    total_expected_fields: total_expected,
    unique_missing_fields: unique_missing,
    unique_extra_fields: unique_extra,
    average_coverage: avg_coverage,
    details: @results
  }
end

#validate_directory(directory, expected_fields = nil) ⇒ Object

Validate all JSON-LD files in a directory

Parameters:

  • directory (String) —

    Directory path

  • expected_fields (Array<Symbol>) (defaults to: nil) —

    Expected model fields



135
136
137
138
139
140
141
142
 
# File 'lib/openphar/exporters/coverage_validator.rb', line 135

def validate_directory(directory, expected_fields = nil)
  Dir.glob(File.join(directory, "**", "*.jsonld")).each do |file|
    puts "Validating: #{file}"
    validate_file(file, expected_fields)
  end

  summary
end

#validate_file(jsonld_file, expected_fields = nil) ⇒ Object

Analyze a JSON-LD file and compare to expected model attributes

Parameters:

  • jsonld_file (String) —

    Path to JSON-LD file

  • expected_fields (Array<Symbol>) (defaults to: nil) —

    Expected field names from model



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
 
# File 'lib/openphar/exporters/coverage_validator.rb', line 82

def validate_file(jsonld_file, expected_fields = nil)
  data = JSON.parse(File.read(jsonld_file))

  # Extract all keys from JSON-LD
  jsonld_keys = extract_all_keys(data).to_a
  domain_keys = jsonld_keys - STANDARD_KEYS

  # Normalize keys to snake_case for comparison
  normalized_jsonld_keys = domain_keys.map { |k| normalize_key(k) }

  # Get expected fields from model if not provided
  expected ||= expected_fields || model_expected_fields
  expected_normalized = expected.map(&:to_s)

  # Find missing fields (in JSON-LD but not in model)
  missing = normalized_jsonld_keys - expected_normalized

  # Find extra fields (in model but not in JSON-LD)
  extra = expected_normalized - normalized_jsonld_keys

  @missing_fields.concat(missing)
  @extra_fields.concat(extra)

  coverage = if normalized_jsonld_keys.any?
               (1 - missing.size.to_f / normalized_jsonld_keys.size) * 100
             else
               100
             end

  @results[jsonld_file] = {
    jsonld_keys: domain_keys.size,
    expected: expected.size,
    missing: missing,
    extra: extra,
    coverage: coverage
  }

  { missing: missing, extra: extra, coverage: coverage }
end