Class: Html2rss::Config::Validator

Inherits:

Dry::Validation::Contract

Object
Dry::Validation::Contract
Html2rss::Config::Validator

show all

Defined in:: lib/html2rss/config/validator.rb

Overview

Validates the configuration hash using Dry::Validation. The configuration options adhere to the documented schema in README.md.

Defined Under Namespace

Classes: BotasaurusRequestExport

Constant Summary collapse

URI_REGEXP = URI format used for channel URL validation.

Url::URI_REGEXP

STYLESHEET_TYPES = Allowed stylesheet MIME types.

FeedBuilder::Rss::Stylesheet::TYPES

LANGUAGE_FORMAT_REGEX = Optional language/region format (en or en-US).

/\A[a-z]{2}(-[A-Z]{2})?\z/

BASE_STRATEGY_OPTIONS = Baseline strategy-plan enum (:auto plus concrete RequestService strategies).

Html2rss::FeedPipeline::StrategyPlan.accepted_names.freeze

DIRECTORY_TOPICS = Controlled vocabulary for catalog-only directory.topics (not RSS channel fields).

%w[
  sports energy tech science news entertainment jobs finance
  security travel environment consumer civic product research
].freeze

ChannelConfig = Contract for the top-level channel section.

Dry::Schema.Params do
  required(:url).filled(:string, format?: URI_REGEXP)
  optional(:title).maybe(:string)
  optional(:description).maybe(:string)
  optional(:language).maybe(:string, format?: LANGUAGE_FORMAT_REGEX)
  optional(:ttl).maybe(:integer, gt?: 0)
  optional(:time_zone).maybe(:string)
  optional(:author).maybe(:string)
  optional(:image).maybe(:string, format?: URI_REGEXP)
end

DirectoryConfig = Contract for catalog-only directory metadata (topics for feed directories).

Dry::Schema.Params do
  optional(:topics).value(:array, min_size?: 1).each(:string, included_in?: DIRECTORY_TOPICS)
end

StylesheetConfig = Contract for a stylesheet entry in stylesheets.

Dry::Schema.Params do
  required(:href).filled(:string)
  required(:type).filled(:string, included_in?: STYLESHEET_TYPES)
  optional(:media).maybe(:string)
end

BotasaurusRequestConfig = Contract for Botasaurus-specific request options (OpenAPI ScrapeRequest minus url).

Dry::Schema.Params do
  config.validate_keys = true

  botasaurus = Html2rss::RequestService::BotasaurusContract

  optional(:execution_mode).filled(:string, included_in?: botasaurus::EXECUTION_MODES)
  optional(:navigation_mode).filled(:string, included_in?: botasaurus::NAVIGATION_MODES)
  optional(:max_retries).filled(:integer, gteq?: 0, lteq?: botasaurus::MAX_RETRIES)
  optional(:wait_for_selector).maybe(:string)
  optional(:wait_timeout_seconds).filled(
    :integer, gteq?: botasaurus::MIN_WAIT_TIMEOUT_SECONDS, lteq?: botasaurus::MAX_WAIT_TIMEOUT_SECONDS
  )
  optional(:scroll).filled(:bool)
  optional(:block_images).filled(:bool)
  optional(:block_images_and_css).filled(:bool)
  optional(:block_trackers).filled(:bool)
  optional(:wait_for_complete_page_load).filled(:bool)
  optional(:headless).filled(:bool)
  optional(:proxy).filled(:string)
  optional(:user_agent).filled(:string)
  optional(:window_size).hash do
    botasaurus::WINDOW_SIZE_PROPERTIES.each do |key|
      required(key).filled(:integer, gt?: 0)
    end
  end
  optional(:lang).filled(:string)
  optional(:cookies).hash
  optional(:headers).hash
end

RequestConfig = Contract for the top-level request section.

Dry::Schema.Params do
  optional(:max_redirects).filled(:integer, gteq?: 0)
  optional(:max_requests).filled(:integer, gt?: 0)
  optional(:total_timeout_seconds).filled(:integer, gt?: 0)
  optional(:botasaurus).hash
  optional(:local_file_path).filled(:string)
end