Class: RosettAi::Adopter::RuleAdopter

Inherits:
Object
  • Object
show all
Defined in:
lib/rosett_ai/adopter/rule_adopter.rb

Overview

Analyzes compiled markdown rule files for inconsistencies, conflicts, harmful content, duplicates, and other issues.

Supports four layers of data privacy protection:

  1. Opt-in per file — sensitive: true in YAML excludes from API analysis
  2. Redaction — regex patterns replace matches before sending to API
  3. Configurable endpoint — ANTHROPIC_API_BASE_URL for proxy/Bedrock/Vertex
  4. Local-only mode — structural checks without API calls

Constant Summary collapse

GENERATED_MARKER =

Returns Marker string identifying generated files.

Returns:

  • (String) —

    Marker string identifying generated files.

'<!-- rosett-ai-'

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(rules_dir:, cache_path:, redactions_path:, engine: 'claude') ⇒ RuleAdopter

Returns a new instance of RuleAdopter.



28
29
30
31
32
33
 
# File 'lib/rosett_ai/adopter/rule_adopter.rb', line 28

def initialize(rules_dir:, cache_path:, redactions_path:, engine: 'claude')
  @rules_dir = Pathname.new(rules_dir)
  @cache_path = Pathname.new(cache_path)
  @redactions_path = Pathname.new(redactions_path)
  @engine = engine.to_s
end

Instance Attribute Details

#cache_path ⇒ Object (readonly)

Returns the value of attribute cache_path.



26
27
28
 
# File 'lib/rosett_ai/adopter/rule_adopter.rb', line 26

def cache_path
  @cache_path
end

#engine ⇒ Object (readonly)

Returns the value of attribute engine.



26
27
28
 
# File 'lib/rosett_ai/adopter/rule_adopter.rb', line 26

def engine
  @engine
end

#redactions_path ⇒ Object (readonly)

Returns the value of attribute redactions_path.



26
27
28
 
# File 'lib/rosett_ai/adopter/rule_adopter.rb', line 26

def redactions_path
  @redactions_path
end

#rules_dir ⇒ Object (readonly)

Returns the value of attribute rules_dir.



26
27
28
 
# File 'lib/rosett_ai/adopter/rule_adopter.rb', line 26

def rules_dir
  @rules_dir
end

Instance Method Details

#analyze(files) ⇒ Hash

Perform analysis on the given files.

Parameters:

  • files (Object) —

    the files

Returns:

  • (Hash) 


133
134
135
136
137
138
139
 
# File 'lib/rosett_ai/adopter/rule_adopter.rb', line 133

def analyze(files)
  return empty_result if files.empty?

  prompt = build_prompt(files)
  executor = resolve_executor
  executor.analyze(prompt)
end

#analyze_local(files) ⇒ Hash

Perform local-only structural analysis.

Parameters:

  • files (Object) —

    the files

Returns:

  • (Hash) 


145
146
147
148
 
# File 'lib/rosett_ai/adopter/rule_adopter.rb', line 145

def analyze_local(files)
  collector = LocalAnalysisCollector.new
  collector.analyze(files)
end

#build_prompt(files) ⇒ String

Build the analysis prompt from file contents.

Parameters:

  • files (Object) —

    the files

Returns:

  • (String) 


107
108
109
110
111
 
# File 'lib/rosett_ai/adopter/rule_adopter.rb', line 107

def build_prompt(files)
  parts = [prompt_header]
  append_file_contents(parts, files)
  parts.join("\n")
end

#cached_result(checksum) ⇒ Hash?

Retrieve a cached analysis result by checksum.

Parameters:

  • checksum (Object) —

    the checksum

Returns:

  • (Hash, nil) 


75
76
77
78
79
80
81
82
83
84
85
86
 
# File 'lib/rosett_ai/adopter/rule_adopter.rb', line 75

def cached_result(checksum)
  return nil unless cache_path.exist?

  data = RosettAi::YamlLoader.load_file(cache_path.to_s, permitted_classes: [Time, Date])
  return nil unless data.is_a?(Hash) && data['checksum'] == checksum
  return nil if cache_expired?(data)

  data['result']
rescue Psych::SyntaxError, Psych::DisallowedClass => e
  RosettAi.logger.warn("Corrupt adopt cache (#{e.message}), will re-analyze")
  nil
end

#content_checksum(files) ⇒ String

Compute a SHA-256 checksum of the given files.

Parameters:

  • files (Object) —

    the files

Returns:

  • (String) 


66
67
68
69
 
# File 'lib/rosett_ai/adopter/rule_adopter.rb', line 66

def content_checksum(files)
  content = files.sort.map { |f| File.read(f) }.join
  Digest::SHA256.hexdigest(content)
end

#discover_managed_files ⇒ Array<String>

Find all rai-managed rule files in the rules directory.

Returns:

  • (Array<String>) 


51
52
53
54
55
56
57
58
59
60
 
# File 'lib/rosett_ai/adopter/rule_adopter.rb', line 51

def discover_managed_files
  return [] unless rules_dir.exist?

  Dir.glob(rules_dir.join('*.md')).select do |file|
    first_line = File.open(file, &:readline)
    first_line.start_with?(GENERATED_MARKER)
  rescue EOFError
    false
  end.sort
end

#evaluate(local_only: false) ⇒ Object

Raises:



35
36
37
38
39
40
41
42
43
44
45
46
47
 
# File 'lib/rosett_ai/adopter/rule_adopter.rb', line 35

def evaluate(local_only: false)
  files = discover_managed_files
  raise RosettAi::AdoptError, 'No managed rule files found in rules directory' if files.empty?

  sensitive = sensitive_files
  api_files = filter_sensitive(files, sensitive)

  if local_only
    evaluate_local(files, sensitive)
  else
    evaluate_remote(files, api_files, sensitive)
  end
end

#filter_sensitive(files, sensitive_sources) ⇒ Array<String>

Exclude files marked as sensitive from the file list.

Parameters:

  • files (Object) —

    the files

  • sensitive_sources (Object) —

    the sensitive sources

Returns:

  • (Array<String>) 


169
170
171
172
173
174
175
176
177
 
# File 'lib/rosett_ai/adopter/rule_adopter.rb', line 169

def filter_sensitive(files, sensitive_sources)
  return files if sensitive_sources.empty?

  sensitive_names = sensitive_sources.map { |f| RosettAi::TextSanitizer.normalize_nfc(File.basename(f, '.yml')) }
  files.reject do |file|
    name = File.basename(file, '.md').sub(/\A[^-]+-/, '')
    sensitive_names.include?(name)
  end
end

#redact(content) ⇒ String

Apply redaction patterns to content before API submission.

Parameters:

  • content (Object) —

    the content

Returns:

  • (String) 


117
118
119
120
121
122
123
124
125
126
127
 
# File 'lib/rosett_ai/adopter/rule_adopter.rb', line 117

def redact(content)
  patterns = load_redaction_patterns
  result = content.dup
  patterns.each do |entry|
    regex = Regexp.new(entry['pattern'])
    result.gsub!(regex, entry['replacement'])
  rescue RegexpError => e
    RosettAi.logger.warn("Invalid redaction pattern '#{entry['pattern']}': #{e.message}")
  end
  result
end

#sensitive_files ⇒ Array<String>

List behaviour files marked as sensitive.

Returns:

  • (Array<String>) 


152
153
154
155
156
157
158
159
160
161
162
 
# File 'lib/rosett_ai/adopter/rule_adopter.rb', line 152

def sensitive_files
  behaviour_dir = RosettAi.root.join('conf', 'behaviour')
  return [] unless behaviour_dir.exist?

  Dir.glob(behaviour_dir.join('*.yml')).select do |file|
    data = RosettAi::YamlLoader.load_file(file)
    data.is_a?(Hash) && data['sensitive'] == true
  rescue StandardError
    false
  end
end

#write_cache(checksum, result) ⇒ void

This method returns an undefined value.

Persist analysis results to the cache file.

Parameters:

  • checksum (Object) —

    the checksum

  • result (Object) —

    the result



93
94
95
96
97
98
99
100
101
 
# File 'lib/rosett_ai/adopter/rule_adopter.rb', line 93

def write_cache(checksum, result)
  FileUtils.mkdir_p(cache_path.dirname)
  data = {
    'checksum' => checksum,
    'analyzed_at' => Time.now.utc.strftime('%Y-%m-%dT%H:%M:%SZ'),
    'result' => result
  }
  File.open(cache_path, 'w', 0o644) { |f| f.write(data.to_yaml) }
end