hyrum 0.1.0 → 0.2.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: b7177d332d4a6a3e7a9527a906e39d4f141af5444aedbb3acff837b07b93f619
-  data.tar.gz: 9b0d67445326a57b5aa4f1bc2926cf06455609f491c75c08e65889a8d019b481
+  metadata.gz: 51ba8be674113de6da2792d5d7489a08a03d3ef0e61976c19d2c90e51d9ad6d7
+  data.tar.gz: 52c4782d0863fab72e35c09e5514da71aa6186a5d5b64aa3223a2644b4da5142
 SHA512:
-  metadata.gz: aa204a19e6e0080792381c09904cc0fcf013b05490875b4f793f4135b3cec0d605d544261266a6b00f54be2a30ddfef4d3fd44e2f8be17060f49924fabe528b3
-  data.tar.gz: 17ceb22c5ea1b7a3837ce9195cc4b0f8a9b4276d502e3a9b4f4450aec61f5bcb780eaf135f2d9ac92717c060a81cf1423af751ef34de54faf532850c67b024f0
+  metadata.gz: 8d3aaed1301a484a1820a880d4178167710c684607f92f6ed5761ec8853baf7f46214c0c9c837f6db05aff2fc2882c3a048bfcc6298c396b09bb58f3a00c3ad4
+  data.tar.gz: '0160106939c1b07d0e8d8880d0f140c79aab70fbac80a2566aa22a24b8624e42ce99205570f17b9931e11ed1f2d9c58c5ad5b5a1ebf285f1f19e60d595ab002d'

data/CHANGELOG.md

@@ -5,6 +5,81 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [Unreleased]
+
+### Breaking Changes
+
+- **Dependency Migration**: Switched from `ruby-openai` gem to `ruby_llm` gem
+- **Environment Variable Changes**:
+  - `OPENAI_ACCESS_TOKEN` → `OPENAI_API_KEY` (renamed)
+  - `OLLAMA_URL` → `OLLAMA_API_BASE` (renamed, must include `/v1` suffix)
+- **Removed Features**:
+  - OpenAI organization ID configuration (can be added back if needed)
+
+### Added
+
+- Support for 10+ AI providers through ruby_llm:
+  - Anthropic Claude (anthropic)
+  - Google Gemini (gemini)
+  - Mistral AI (mistral)
+  - DeepSeek (deepseek)
+  - Perplexity (perplexity)
+  - OpenRouter (openrouter)
+  - Google Vertex AI (vertexai)
+  - AWS Bedrock (bedrock)
+  - GPUStack (gpustack)
+- Improved error messages with provider-specific configuration guidance
+- Structured JSON output via schemas for more reliable message parsing
+- Better model defaults for each provider
+
+### Changed
+
+- Renamed `OpenaiGenerator` to `AiGenerator` for implementation-agnostic naming
+- Simplified generator implementation using ruby_llm's unified API
+- Test suite now uses RubyLLM-level mocking instead of VCR HTTP cassettes
+- **Optimized default models for cost-efficiency**:
+  - Anthropic: `claude-sonnet-4` → `claude-haiku-20250514` (~10x cheaper, perfect for simple text tasks)
+  - Bedrock: Sonnet → Haiku version (same cost savings)
+  - For simple message generation, cheaper models provide excellent results
+- **Code Quality Improvements**:
+  - Extracted fake messages data to external JSON file (improved maintainability)
+  - Refactored API key environment variable mapping to use constant hash
+  - Reduced FakeGenerator from 298 to 36 lines
+  - Reduced AiGenerator cyclomatic complexity from 12 to 1
+  - All RuboCop warnings resolved
+
+### Removed
+
+- VCR test cassettes (replaced with direct RubyLLM mocking)
+- `OpenaiGenerator` class (replaced by `AiGenerator`)
+- **Dependencies**: `vcr` and `webmock` gems (no longer needed with RubyLLM-level mocking)
+
+### Migration Guide
+
+1. Update environment variables:
+   ```bash
+   # Old
+   export OPENAI_ACCESS_TOKEN=sk-...
+   export OLLAMA_URL=http://localhost:11434
+
+   # New
+   export OPENAI_API_KEY=sk-...
+   export OLLAMA_API_BASE=http://localhost:11434/v1
+   ```
+
+2. Add new provider API keys as needed:
+   ```bash
+   export ANTHROPIC_API_KEY=sk-ant-...
+   export GEMINI_API_KEY=...
+   ```
+
+3. Update your Gemfile and run `bundle install`:
+   ```ruby
+   gem 'hyrum'  # latest version
+   ```
+
+For detailed migration instructions, see the [README](README.md#configuration).
+
 ## [0.1.0] - 2024-12-16
 
 ### Fixed

data/README.md

@@ -26,11 +26,26 @@ no longer static, improving your api design.
 
 ## Example
 
+Generate coffee brewing error messages with OpenAI:
+
+```bash
+hyrum --service openai \
+      --message "The server refuses the attempt to brew coffee with a teapot" \
+      --key e418 \
+      --format ruby
+```
+
+Or use Anthropic's Claude:
+
 ```bash
-hyrum --service openai --key e418 --format ruby \
-      --message "The server refuses the attempt to brew coffee with a teapot"
+hyrum --service anthropic \
+      --message "The server refuses the attempt to brew coffee with a teapot" \
+      --key e418 \
+      --format ruby
 ```
 
+Output:
+
 ```ruby
 # frozen_string_literal: true
 
@@ -58,19 +73,83 @@ end
 ## Usage
 
 ```
-❯ hyrum --help # OR from the repo as `./exe/hyrum --help`
+❯ hyrum --help # OR from the repo as `./bin/hyrum --help`
 Usage: hyrum [options]
     -v, --[no-]verbose               Run verbosely
     -f, --format FORMAT              Output format. Supported formats are:
                                      ruby, javascript, python, java, text, json
-    -m, --message MESSAGE            Status message
-    -k, --key KEY                    Message key
-    -s, --service SERVICE            AI service: one of openai, ollama, fake
+                                     (default: text)
+    -m, --message MESSAGE            Status message (required unless fake)
+    -k, --key KEY                    Message key (default: status)
+    -n, --number NUMBER              Number of messages to generate (default: 5)
+    -s, --service SERVICE            AI service: one of openai, anthropic, gemini, ollama,
+                                     mistral, deepseek, perplexity, openrouter, vertexai,
+                                     bedrock, gpustack, fake (default: fake)
     -d, --model MODEL                AI model: must be a valid model for the selected service
+        --validate                   Enable quality validation (default: off)
+        --min-quality SCORE          Minimum quality score 0-100 (default: 70)
+        --strict                     Fail on quality issues instead of warning (default: false)
+        --show-scores                Include quality metrics in output (default: false)
     -h, --help                       Show this message
         --version                    Show version
 ```
 
+## Quality validation
+
+Hyrum can validate the quality of generated message variations to ensure they
+achieve Hyrum's Law goal: variations preserve the original message's meaning
+while using different wording.
+
+### Basic validation
+
+```bash
+hyrum --validate -s openai -m "Server error" -f ruby --show-scores
+```
+
+This includes quality metrics as comments in the output:
+
+```ruby
+# Quality Score: 82.5/100
+# - Semantic similarity: 94.0% (variations preserve meaning)
+# - Lexical diversity: 68.0% (variation in wording)
+
+# frozen_string_literal: true
+# ... rest of generated code
+```
+
+### Validation options
+
+- `--validate` - Enable quality validation (default: off)
+- `--min-quality SCORE` - Minimum acceptable quality score 0-100 (default: 70)
+- `--strict` - Exit with error if quality check fails (default: warning only)
+- `--show-scores` - Include quality metrics in output (default: false)
+
+### Strict mode for automated workflows
+
+Use strict mode to enforce quality in automated workflows:
+
+```bash
+hyrum --validate --strict --min-quality 75 -s openai -m "Error message" -f ruby
+```
+
+This exits with a non-zero status code if quality is below 75.
+
+### How it works
+
+The validator measures two key metrics:
+
+1. **Semantic Similarity** (≥85% required): Ensures each variation preserves the meaning of your original message
+   - Uses embedding models (OpenAI, Google, etc.) when available
+   - Falls back to word overlap heuristic if no embedding provider configured
+2. **Lexical Diversity** (30-70% required): Ensures variations use different words from each other
+
+The overall quality score is a weighted combination of both metrics.
+
+**Embedding Support**: Semantic similarity works best with embedding models.
+Configure any embedding provider (OpenAI, Google, etc.) and Hyrum will use it
+automatically. If no embedding provider is configured, validation still works
+using a word overlap heuristic.
+
 ## Installation
 Install the gem and add to the application's Gemfile by executing:
 
@@ -86,31 +165,49 @@ if you want to see output quickly, you can use the `-s fake` option to use a fak
 service provider that will generate stock responses.
 
 ```bash
-hyrum -s fake -f ruby -m "anything here"
+hyrum -s fake -f ruby -k "404" -n 5
 ```
 
 You don't even need to install the gem to use Hyrum, fake service provider or not.
 You can run the executable directly from a cloned repository.
 
 ```bash
-./exec/hyrum -s fake -f ruby -m "anything here"
+./bin/hyrum -s fake -f ruby -m "anything here"
 ```
 
-## Configruation
+## Configuration
 
 ### OpenAI (`--service openai`)
-Hyrum requires an OpenAI API token to access the language models. The API token should be
+Hyrum requires an OpenAI API key to access the language models. The API key should be
 set as an environment variable as shown below.
 
 ```bash
-export OPENAI_ACCESS_TOKEN=your_open_ai_token
+export OPENAI_API_KEY=your_openai_api_key
+```
+
+If you specify the `openai` service but no model, Hyrum will use `gpt-4o-mini`.
+
+### Anthropic (`--service anthropic`)
+To use Anthropic's Claude models, set your API key:
+
+```bash
+export ANTHROPIC_API_KEY=your_anthropic_api_key
 ```
 
-If you specify the `openai` service but no model, Hyrum will use the `gpt-o4-mini`.
+Default model: `claude-haiku-20250514`
+
+### Gemini (`--service gemini`)
+To use Google's Gemini models, set your API key:
+
+```bash
+export GEMINI_API_KEY=your_gemini_api_key
+```
+
+Default model: `gemini-2.0-flash-exp`
 
 ### Ollama (`--service ollama`)
 If you specify the `ollama` service, Hyrum will attempt to use the Ollama API
-running at `http://localhost:11434`. You can set the `OLLAMA_URL` environment
+running at `http://localhost:11434/v1`. You can set the `OLLAMA_API_BASE` environment
 variable to specify a different URL.
 
 Make sure your ollama server is running before using the `ollama` service.
@@ -122,9 +219,31 @@ ollama serve
 Use `ollama list` to see the available models. For more information on using
 ollama and downloading models, see the [ollama repository](http://ollama.com).
 
+Default model: `llama3`
+
+### Other providers
+
+Hyrum supports all providers available in [ruby_llm](https://github.com/crmne/ruby_llm),
+including:
+- **Mistral** (`--service mistral`) - Set `MISTRAL_API_KEY`
+- **DeepSeek** (`--service deepseek`) - Set `DEEPSEEK_API_KEY`
+- **Perplexity** (`--service perplexity`) - Set `PERPLEXITY_API_KEY`
+- **OpenRouter** (`--service openrouter`) - Set `OPENROUTER_API_KEY`
+- **Vertex AI** (`--service vertexai`) - Set `GOOGLE_CLOUD_PROJECT` and `GOOGLE_CLOUD_LOCATION`
+- **AWS Bedrock** (`--service bedrock`) - Uses AWS credential chain or set `AWS_ACCESS_KEY_ID`
+- **GPUStack** (`--service gpustack`) - Set `GPUSTACK_API_BASE` and `GPUSTACK_API_KEY`
+
+See the [ruby_llm configuration docs](https://ruby-llm.com/configuration) for detailed
+setup instructions for each provider.
+
 ## Supported formats and AI services
 
-See [Usage](#usage) for a list of supported formats and AI services.
+**Formats:** ruby, javascript, python, java, text, json
+
+**AI Services:** openai, anthropic, gemini, ollama, mistral, deepseek, perplexity,
+openrouter, vertexai, bedrock, gpustack, fake
+
+See [Configuration](#configuration) for setup details for each service.
 
 ## Compatibility
 

data/bin/hyrum

@@ -1,10 +1,9 @@
 #!/usr/bin/env ruby
 # frozen_string_literal: true
 
-$LOAD_PATH.unshift(File.dirname(__FILE__) + '/../lib')
+$LOAD_PATH.unshift("#{File.dirname(__FILE__)}/../lib")
 
 require 'bundler/setup'
 require 'hyrum'
 
-# Hyrum::CLI.start(ARGV)
 Hyrum.run(ARGV)

data/lib/hyrum/data/fake_messages.json

@@ -0,0 +1,282 @@
+{
+  "e400": [
+    "Bad Request: The server cannot process the request due to client error",
+    "Invalid syntax in the request parameters",
+    "The request could not be understood by the server",
+    "Missing required parameters in the request",
+    "Malformed request syntax"
+  ],
+  "e401": [
+    "Unauthorized: Authentication is required to access this resource",
+    "Invalid credentials provided",
+    "Access token has expired",
+    "Missing authentication token",
+    "You must log in to access this resource"
+  ],
+  "e402": [
+    "Payment Required: The requested resource requires payment",
+    "Subscription has expired",
+    "Please upgrade your account to access this feature",
+    "Payment verification failed",
+    "Resource access requires premium subscription"
+  ],
+  "e403": [
+    "Forbidden: You don't have permission to access this resource",
+    "Access denied due to insufficient privileges",
+    "Your IP address has been blocked",
+    "Account suspended",
+    "Resource access restricted to authorized users only"
+  ],
+  "e404": [
+    "Not Found: The requested resource could not be located",
+    "The page you're looking for doesn't exist",
+    "Resource has been moved or deleted",
+    "Invalid URL or endpoint",
+    "The requested item is no longer available"
+  ],
+  "e405": [
+    "Method Not Allowed: The requested HTTP method is not supported",
+    "This endpoint doesn't support the specified HTTP method",
+    "Invalid HTTP method for this resource",
+    "Please check the API documentation for supported methods",
+    "HTTP method not supported for this endpoint"
+  ],
+  "e406": [
+    "Not Acceptable: Cannot generate response matching Accept headers",
+    "Requested format is not available",
+    "Content type negotiation failed",
+    "Server cannot generate acceptable response",
+    "Unsupported media type requested"
+  ],
+  "e407": [
+    "Proxy Authentication Required: Authentication with proxy server needed",
+    "Please authenticate with the proxy first",
+    "Proxy credentials required",
+    "Missing proxy authentication",
+    "Cannot proceed without proxy authentication"
+  ],
+  "e408": [
+    "Request Timeout: The server timed out waiting for the request",
+    "Client took too long to send the complete request",
+    "Connection timed out while waiting for data",
+    "Request processing exceeded time limit",
+    "Please try submitting your request again"
+  ],
+  "e409": [
+    "Conflict: Request conflicts with current state of the server",
+    "Resource version conflict detected",
+    "Concurrent modification error",
+    "Data conflict with existing resource",
+    "Cannot process due to resource state conflict"
+  ],
+  "e410": [
+    "Gone: The requested resource is no longer available",
+    "Resource has been permanently removed",
+    "This version of the API has been deprecated",
+    "Content has been permanently deleted",
+    "Resource not available at this location"
+  ],
+  "e411": [
+    "Length Required: Content-Length header is required",
+    "Missing Content-Length header in request",
+    "Request must include content length",
+    "Cannot process request without content length",
+    "Please specify the content length"
+  ],
+  "e412": [
+    "Precondition Failed: Request preconditions failed",
+    "Resource state doesn't match expectations",
+    "Conditional request failed",
+    "Required conditions not met",
+    "Cannot proceed due to failed preconditions"
+  ],
+  "e413": [
+    "Payload Too Large: Request entity exceeds limits",
+    "File size is too large",
+    "Request data exceeds maximum allowed size",
+    "Please reduce the size of your request",
+    "Upload size limit exceeded"
+  ],
+  "e414": [
+    "URI Too Long: The requested URL exceeds server limits",
+    "URL length exceeds maximum allowed characters",
+    "Request URL is too long to process",
+    "Please shorten the URL and try again",
+    "URI length exceeds server configuration"
+  ],
+  "e415": [
+    "Unsupported Media Type: Request format not supported",
+    "Invalid content type in request",
+    "Server doesn't support this media format",
+    "Please check supported content types",
+    "Media type not accepted by the server"
+  ],
+  "e416": [
+    "Range Not Satisfiable: Cannot fulfill requested range",
+    "Requested range not available",
+    "Invalid range specified in request",
+    "Range header value not satisfiable",
+    "Cannot serve the requested content range"
+  ],
+  "e417": [
+    "Expectation Failed: Server cannot meet Expect header requirements",
+    "Expected condition could not be fulfilled",
+    "Server cannot satisfy expectations",
+    "Expect header requirements not met",
+    "Request expectations cannot be met"
+  ],
+  "e418": [
+    "I'm a teapot: Server refuses to brew coffee with a teapot",
+    "This server is a teapot, not a coffee maker",
+    "Coffee brewing request denied by teapot",
+    "Hyper Text Coffee Pot Control Protocol error",
+    "Cannot brew coffee using teapot protocol"
+  ],
+  "e421": [
+    "Misdirected Request: Server is not able to produce a response",
+    "Request was directed to wrong server",
+    "Cannot handle misdirected request",
+    "Invalid server configuration for request",
+    "Request cannot be processed by this server"
+  ],
+  "e422": [
+    "Unprocessable Entity: Request semantically incorrect",
+    "Validation failed for the request",
+    "Cannot process invalid request data",
+    "Semantic errors in request content",
+    "Request contains invalid parameters"
+  ],
+  "e423": [
+    "Locked: Resource is locked",
+    "Resource access blocked by lock",
+    "Cannot modify locked resource",
+    "Resource temporarily unavailable due to lock",
+    "Please wait for resource lock to clear"
+  ],
+  "e424": [
+    "Failed Dependency: Request failed due to previous request failure",
+    "Dependent request failed",
+    "Cannot proceed due to failed dependency",
+    "Previous request failure prevents completion",
+    "Dependency chain broken"
+  ],
+  "e425": [
+    "Too Early: Server unwilling to risk processing early request",
+    "Request arrived too early to process",
+    "Cannot handle premature request",
+    "Please retry request later",
+    "Early request rejected"
+  ],
+  "e426": [
+    "Upgrade Required: Client must upgrade protocol",
+    "Please upgrade to continue",
+    "Protocol upgrade necessary",
+    "Current protocol version not supported",
+    "Connection upgrade required"
+  ],
+  "e428": [
+    "Precondition Required: Request must be conditional",
+    "Missing required precondition",
+    "Please include precondition headers",
+    "Cannot process request without preconditions",
+    "Conditional request required"
+  ],
+  "e429": [
+    "Too Many Requests: Rate limit exceeded",
+    "Please slow down your requests",
+    "API rate limit reached",
+    "Too many requests in time window",
+    "Request quota exceeded"
+  ],
+  "e431": [
+    "Request Header Fields Too Large: Header fields too large",
+    "Headers exceed size limits",
+    "Request contains oversized headers",
+    "Please reduce header size",
+    "Header fields exceed server limits"
+  ],
+  "e451": [
+    "Unavailable For Legal Reasons: Content legally restricted",
+    "Access denied due to legal restrictions",
+    "Content blocked for legal reasons",
+    "Resource legally unavailable",
+    "Cannot serve content due to legal constraints"
+  ],
+  "e500": [
+    "Internal Server Error: Something went wrong on our end",
+    "Unexpected server error occurred",
+    "Server encountered an error processing request",
+    "Internal error in server configuration",
+    "System error, please try again later"
+  ],
+  "e501": [
+    "Not Implemented: Functionality not supported",
+    "Request method not supported by server",
+    "Feature not available on this server",
+    "Requested functionality not implemented",
+    "Server does not support this operation"
+  ],
+  "e502": [
+    "Bad Gateway: Invalid response from upstream server",
+    "Gateway received invalid response",
+    "Error communicating with upstream server",
+    "Invalid response from backend service",
+    "Gateway communication error"
+  ],
+  "e503": [
+    "Service Unavailable: Server temporarily unavailable",
+    "System under maintenance",
+    "Server is overloaded",
+    "Service temporarily offline",
+    "Please try again later"
+  ],
+  "e504": [
+    "Gateway Timeout: Upstream server timed out",
+    "Gateway connection timed out",
+    "Backend server not responding",
+    "Request timed out at gateway",
+    "Gateway failed to get timely response"
+  ],
+  "e505": [
+    "HTTP Version Not Supported: HTTP version not supported",
+    "Unsupported HTTP protocol version",
+    "Server doesn't support HTTP version",
+    "Please use a supported HTTP version",
+    "HTTP protocol version not compatible"
+  ],
+  "e506": [
+    "Variant Also Negotiates: Server configuration error",
+    "Circular reference in content negotiation",
+    "Internal configuration conflict",
+    "Content negotiation error",
+    "Server misconfiguration detected"
+  ],
+  "e507": [
+    "Insufficient Storage: Server out of storage",
+    "Not enough storage space available",
+    "Storage quota exceeded",
+    "Server storage capacity reached",
+    "Cannot process due to storage limits"
+  ],
+  "e508": [
+    "Loop Detected: Infinite loop detected in request",
+    "Request processing loop detected",
+    "Circular dependency found",
+    "Cannot complete due to infinite loop",
+    "Processing halted due to loop"
+  ],
+  "e510": [
+    "Not Extended: Further extensions needed",
+    "Required extension not supported",
+    "Cannot fulfill without extension",
+    "Missing required protocol extension",
+    "Extension support required"
+  ],
+  "e511": [
+    "Network Authentication Required: Network access requires authentication",
+    "Please authenticate with network",
+    "Network login required",
+    "Cannot access network without authentication",
+    "Network credentials required"
+  ]
+}

data/lib/hyrum/formats/formatter.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require 'erb'
+
 module Hyrum
   module Formats
     FORMATS = %i[ruby javascript python java text json].freeze
@@ -11,10 +13,14 @@ module Hyrum
         @options = options
       end
 
-      def format(messages)
+      def format(messages, validation_result = nil)
         template_file = File.join(__dir__, 'templates', "#{options[:format]}.erb")
         template = ERB.new(File.read(template_file), trim_mode: '-')
-        template.result_with_hash(messages: messages)
+        template.result_with_hash(
+          messages: messages,
+          validation_result: validation_result,
+          show_scores: options[:show_scores]
+        )
       end
     end
   end

data/lib/hyrum/formats/templates/java.erb

@@ -1,3 +1,12 @@
+<% if validation_result && show_scores -%>
+// Quality Score: <%= validation_result.score %>/100
+// - Semantic similarity: <%= validation_result.semantic_similarity %>% (variations preserve meaning)
+// - Lexical diversity: <%= validation_result.lexical_diversity %>% (variation in wording)
+<% validation_result.warnings.each do |warning| -%>
+// Warning: <%= warning %>
+<% end -%>
+//
+<% end -%>
 import java.util.HashMap;
 import java.util.List;
 import java.util.Random;

data/lib/hyrum/formats/templates/javascript.erb

@@ -1,3 +1,12 @@
+<% if validation_result && show_scores -%>
+// Quality Score: <%= validation_result.score %>/100
+// - Semantic similarity: <%= validation_result.semantic_similarity %>% (variations preserve meaning)
+// - Lexical diversity: <%= validation_result.lexical_diversity %>% (variation in wording)
+<% validation_result.warnings.each do |warning| -%>
+// Warning: <%= warning %>
+<% end -%>
+//
+<% end -%>
 const Messages = {
   MESSAGES: {
 <% messages.each do |key, values| -%>

data/lib/hyrum/formats/templates/json.erb

@@ -1 +1,14 @@
+<% if validation_result && show_scores -%>
+<%= JSON.pretty_generate({
+  quality: {
+    score: validation_result.score,
+    semantic_similarity: validation_result.semantic_similarity,
+    lexical_diversity: validation_result.lexical_diversity,
+    passed: validation_result.passed?,
+    warnings: validation_result.warnings
+  },
+  messages: messages
+}) %>
+<% else -%>
 <%= JSON.pretty_generate(messages) %>
+<% end -%>

data/lib/hyrum/formats/templates/python.erb

@@ -1,3 +1,12 @@
+<% if validation_result && show_scores -%>
+# Quality Score: <%= validation_result.score %>/100
+# - Semantic similarity: <%= validation_result.semantic_similarity %>% (variations preserve meaning)
+# - Lexical diversity: <%= validation_result.lexical_diversity %>% (variation in wording)
+<% validation_result.warnings.each do |warning| -%>
+# Warning: <%= warning %>
+<% end -%>
+#
+<% end -%>
 import random
 import sys