feat(test): introduce VCR-based recipe parsing harness

.gitignore

@@ -37,3 +37,7 @@
 !/app/assets/builds/.keep
 
 db/production*
+
+# Ignore VCR cassettes (recorded HTTP interactions)
+test/fixtures/vcr_cassettes/*
+!test/fixtures/vcr_cassettes/.gitkeep

Gemfile

@@ -24,6 +24,12 @@ group :development, :test do
   gem "mocha"
 end
 
+# Gems used exclusively in the test suite
+group :test do
+  gem "vcr", "~> 6.2"
+  gem "webmock", "~> 3.23"
+end
+
 group :development do
   gem "web-console"
 end

README.md

@@ -67,3 +67,8 @@ User.find_by!(email_address: "me@example.com").update!(admin: false)
 ## Authors
 
 - [@bradly](https://www.github.com/bradly)
+
+## Running the test suite
+
+See `TESTING.md` for details on recording and replaying VCR cassettes when
+testing the recipe parser.

TESTING.md

@@ -0,0 +1,70 @@
+# Testing the Recipe Parsing Logic
+
+This project relies on **VCR** to record and replay real HTTP responses in the
+test suite.  Using VCR keeps the tests **fast**, **repeatable**, and **offline
+friendly** while still exercising the full `RecipeExtractor` stack.
+
+## Directory layout
+
+- `test/fixtures/vcr_cassettes/` – YAML cassettes (recorded HTTP
+  interactions).  The entire directory is excluded from Git via `.gitignore` –
+  only a `.gitkeep` placeholder is tracked so the folder exists on fresh
+  clones.
+
+  **Why are cassettes ignored?**  Personal recipe URLs may expose private
+  browsing history or inadvertently redistribute copyrighted HTML.  Keeping
+  them local avoids licensing headaches while still giving every developer a
+  convenient harness.
+
+## Recording a new cassette
+
+1. **Pick a recipe URL** that you want to add as a regression test.
+2. Run the following rake task:
+
+   ```bash
+   # Uses the URL itself (parameter-ized) for the cassette name
+   bin/rails parsing:record_vcr[https://www.allrecipes.com/recipe/24074/alysias-basic-meat-lasagna]
+
+   # …or provide a custom cassette name
+   bin/rails parsing:record_vcr[https://www.bbcgoodfood.com/recipes/lemon-drizzle,citrus_cake]
+   ```
+
+The task will:
+
+- Boot Rails
+- Configure VCR in **record-all** mode
+- Fetch the URL through `RecipeExtractor`
+- Write the cassette to `test/fixtures/vcr_cassettes/<name>.yml`
+
+## Adding an assertion
+
+Open (or create) a test file under `test/parsing/` and wrap your extraction
+code in `VCR.use_cassette`:
+
+```ruby
+VCR.use_cassette('citrus_cake') do
+  extractor = RecipeExtractor.new('https://www.bbcgoodfood.com/recipes/lemon-drizzle')
+  data = extractor.data
+
+  assert_equal 'Lemon drizzle cake', data[:name]
+end
+```
+
+When the cassette is missing the test will be **skipped** (helpful for CI
+environments without network access).  Run the rake task locally to create it
+and the test will automatically start running.
+
+## Test run cheatsheet
+
+```bash
+# Run the full suite
+bin/rails test
+
+# Run only the parsing tests
+bin/rails test test/parsing
+
+# Re-record a single cassette
+bin/rails parsing:record_vcr[<url>,<cassette_name>]
+```
+
+Happy parsing! 🥣

lib/tasks/vcr.rake

@@ -0,0 +1,54 @@
+# frozen_string_literal: true
+
+# Rake tasks for working with VCR cassettes used by the recipe parsing tests.
+
+namespace :parsing do
+  desc <<~DESC
+    Record a VCR cassette for the given recipe URL.
+
+    Examples:
+      # Generate a cassette using an automatic, URL-based filename
+      bin/rails parsing:record_vcr[https://example.com/recipe]
+
+      # Specify a custom cassette name (without file extension)
+      bin/rails parsing:record_vcr[https://example.com/recipe,my_custom_name]
+  DESC
+  task :record_vcr, %i[url cassette_name] => :environment do |_t, args|
+    url = args[:url]
+
+    unless url.present?
+      warn 'ERROR: URL argument is required. Example: bin/rails parsing:record_vcr[https://example.com/recipe]'
+      exit 1
+    end
+
+    cassette_name = args[:cassette_name].presence || url.parameterize.presence || 'recipe'
+
+    cassette_dir = Rails.root.join('test', 'fixtures', 'vcr_cassettes')
+    FileUtils.mkdir_p(cassette_dir)
+
+    require 'vcr'
+    require 'webmock'
+
+    VCR.configure do |config|
+      config.cassette_library_dir = cassette_dir
+      config.hook_into :webmock
+      config.ignore_localhost = true
+      config.default_cassette_options = { record: :all }
+    end
+
+    puts "Recording VCR cassette '#{cassette_name}.yml' for #{url}..."
+
+    VCR.use_cassette(cassette_name) do
+      extractor = RecipeExtractor.new(url)
+      begin
+        extractor.data
+        puts 'OK: fetch complete.'
+      rescue StandardError => e
+        warn "ERROR: Failed to fetch recipe - #{e.class}: #{e.message}"
+        raise
+      end
+    end
+
+    puts "Cassette saved to #{cassette_dir.join(cassette_name + '.yml')}"
+  end
+end

test/fixtures/vcr_cassettes/.gitkeep

@@ -0,0 +1 @@
+

test/parsing/recipe_extractor_vcr_test.rb

@@ -0,0 +1,28 @@
+require 'test_helper'
+
+# Test harness that verifies RecipeExtractor can parse a real-world recipe
+# page using a VCR cassette.  The cassette file itself is *not* committed
+# to the repo; instead we skip the assertion unless a cassette has already
+# been recorded locally.
+
+class RecipeExtractorVCRTest < ActiveSupport::TestCase
+  CASSETTE_NAME = 'example_recipe'
+  TEST_URL      = 'https://www.allrecipes.com/recipe/24074/alysias-basic-meat-lasagna'
+
+  test 'extract recipe data via recorded HTTP interaction' do
+    cassette_file = Rails.root.join('test', 'fixtures', 'vcr_cassettes', "#{CASSETTE_NAME}.yml")
+
+    # Skip in CI or first-run scenarios where the cassette does not yet exist.
+    skip "VCR cassette not present (run `bin/rails parsing:record_vcr[#{TEST_URL}]` to create it)" unless File.exist?(cassette_file)
+
+    VCR.use_cassette(CASSETTE_NAME) do
+      extractor = RecipeExtractor.new(TEST_URL)
+      data = extractor.data
+
+      # Minimal smoke tests – adjust / extend as needed when new cassettes are added.
+      assert data[:name].present?, 'recipe name should be present'
+      assert data[:ingredients].present?, 'ingredients should be present'
+      assert data[:instruction_sections].present?, 'instruction sections should be present'
+    end
+  end
+end

test/test_helper.rb

@@ -3,6 +3,30 @@
 require "rails/test_help"
 require 'minitest/autorun'
 require 'mocha/minitest'
+# HTTP interaction recording / replay
+require 'webmock/minitest'
+require 'vcr'
+
+# Configure VCR to store cassettes under `test/fixtures/vcr_cassettes`
+VCR.configure do |config|
+  # Store cassettes in our dedicated fixtures directory
+  config.cassette_library_dir = Rails.root.join('test', 'fixtures', 'vcr_cassettes')
+
+  # Hook VCR into WebMock so that HTTP calls are intercepted
+  config.hook_into :webmock
+
+  # Allow localhost connections (capybara, rails server, etc.)
+  config.ignore_localhost = true
+
+  # Filter environment specific data (e.g. authorization headers)
+  config.filter_sensitive_data('<AUTH_TOKEN>') { ENV['AUTH_TOKEN'] }
+
+  # When a cassette is missing, record a new one in :new_episodes mode by default.
+  config.default_cassette_options = { record: :new_episodes }
+
+  # Enable more readable names for generated cassette files
+  config.configure_rspec_metadata! if config.respond_to?(:configure_rspec_metadata!)
+end
 
 module ActiveSupport
   class TestCase