sitedog_parser 0.3.1 → 0.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: da833e3fe74bc8bf12d8039a18c7bd0b4bd4e19bbe0cf8c78bfc9e68ae5cb8d4
-  data.tar.gz: 71f10cafaf8971e6cc2ddadd40847d8f2835c73cdaa530ec0eca2a0a5fd06605
+  metadata.gz: 753510fc5e4c0a2d565f29d0e5b894f653f83c23cb3354d27a631a74852ea421
+  data.tar.gz: 3a2ec09307309a917f4530a937b7443b061c2e671bc5e84a627e487ee7536a3c
 SHA512:
-  metadata.gz: bcaf3041c47b219f6b8d10b39cef222a2535e180f274ef1c8cd231d7d3f4400a50ed278e6de6014b53af42987c3b3e1979d842f70918a5b1ecaf3bd4b17a4257
-  data.tar.gz: 493b359dbfddb4f94b1eaa135f4b81e21029183f052d4777d15abaece5ac244c41e136a5fcd399ad4911d15bea415026308ae2ef4c2f0305f33c441ead4eada5
+  metadata.gz: dcb05fa986c51a3a69da4bb812042777e6c878d778215af6466d58aad90f643691e64399a98e17d403e79b5eba607287ed6129323626d69c0bcd8136056893d2
+  data.tar.gz: cf4e7360906081835272daeaa168c927d9d2a46787d4328afe22a1db03e82de9167a07280d6e41cdf81eeb26ff2de644494185a07a1a26124a523016ee821b4d

data/bin/sitedog_cli

@@ -4,6 +4,7 @@ require 'bundler/setup'
 require 'sitedog_parser'
 require 'optparse'
 require 'logger'
+require 'yaml'
 
 # Set default options
 options = {
@@ -132,10 +133,25 @@ end
 begin
   logger.debug "Processing file: #{file_path}"
 
-  # Convert YAML to JSON
+  # Load YAML to check raw data
+  raw_yaml = YAML.load_file(file_path)
+  if options[:debug]
+    logger.debug "Raw YAML data for debug:"
+    logger.debug raw_yaml.inspect
+    logger.debug ""
+  end
+
+  # Convert YAML to hash
   data = SitedogParser::Parser.to_hash(file_path, { logger: logger })
   logger.debug "Data converted to hash"
 
+  # Debug the parsed data
+  if options[:debug]
+    logger.debug "Parsed data structure:"
+    logger.debug data.inspect
+    logger.debug ""
+  end
+
   # Convert to JSON based on formatting options
   json_data = if options[:compact_children]
                 logger.debug "Generating JSON with compact inner objects"

data/lib/service.rb

@@ -1,11 +1,13 @@
-class Service < Data.define(:service, :url, :children, :image_url)
-  def initialize(service:, url: nil, children: [], image_url: nil)
+class Service < Data.define(:service, :url, :children, :image_url, :properties, :value)
+  def initialize(service:, url: nil, children: [], image_url: nil, properties: {}, value: nil)
     raise ArgumentError, "Service cannot be empty" if service.nil? || service.empty?
 
     service => String
     url => String if url
     children => Array if children
     image_url => String if image_url
+    properties => Hash if properties
+    # value может быть любого типа, поэтому не проверяем
 
     super
   end

data/lib/service_factory.rb

@@ -61,6 +61,10 @@ class ServiceFactory
     in Hash
       logger.debug "hash: #{data}"
 
+      # Check if all values are URL-like strings
+      all_url_like = data.values.all? { |v| v.is_a?(String) && UrlChecker.url_like?(v) }
+      logger.debug "All values are URL-like: #{all_url_like}, values: #{data.values.map { |v| "#{v.class}: #{v}" }.join(', ')}"
+
       # Protection from nil values in key fields
       if (data.key?(:service) || data.key?("service")) &&
          (data[:service].nil? || data["service"].nil?)
@@ -77,6 +81,8 @@ class ServiceFactory
           # Первый приоритет - поиск в словаре по URL
           child_dict_entry = dictionary.match(url_value)
 
+          logger.debug "Child for #{key}: service_name=#{service_name}, url=#{url_value}, dict_entry=#{child_dict_entry}"
+
           if child_dict_entry && child_dict_entry['name']
             # Если нашли запись в словаре по URL, используем её имя вместо ключа
             service_name = child_dict_entry['name']
@@ -103,10 +109,45 @@ class ServiceFactory
 
         # Create parent service with child elements
         if service_type && children.any?
+          logger.debug "Returning service for #{service_type} with #{children.size} children"
           return Service.new(service: service_type.to_s, children: children)
         elsif children.size == 1
-          # If only one service and no service_type, return it directly
+          # If only one service and no service_type, return it
+          logger.debug "Returning single child service (no service_type)"
           return children.first
+        else
+          logger.debug "Not returning a service for #{data.inspect}, service_type=#{service_type}, children.size=#{children.size}"
+        end
+      # 1.5 Check if hash contains at least some URL-like strings
+      elsif data.values.any? { |v| v.is_a?(String) && UrlChecker.url_like?(v) }
+        logger.debug "hash with some URL-like values: #{data.inspect}"
+
+        # Debug: Check each value for URL-like
+        data.each do |k, v|
+          if v.is_a?(String)
+            logger.debug "  Checking #{k}: #{v} - URL-like? #{UrlChecker.url_like?(v)}"
+          else
+            logger.debug "  Skipping non-string #{k}: #{v.class}"
+          end
+        end
+
+        # Сохраняем все значения в properties, сохраняя порядок
+        properties = {}
+        data.each do |key, value|
+          properties[key.to_s] = value
+          logger.debug "Added property for #{key}: #{value}"
+        end
+
+        # Create service with properties only
+        if !properties.empty?
+          service = Service.new(
+            service: service_type.to_s,
+            url: nil,
+            properties: properties,
+            children: []  # Пустой массив children
+          )
+          logger.debug "Returning service with #{properties.size} properties"
+          return service
         end
       end
 
@@ -225,19 +266,38 @@ class ServiceFactory
     in Array
       logger.debug "array: #{data}"
 
-      # Create services from array elements
-      children = data.map { |item| create(item, service_type, dictionary_path) }.compact
+      # Create services from all array elements for children
+      children = []
+      data.each_with_index do |item, index|
+        # Для URL-подобных строк используем стандартный механизм
+        if item.is_a?(String) && UrlChecker.url_like?(item)
+          child_service = create(item, service_type, dictionary_path, options)
+          children << child_service if child_service
+        else
+          # Для простых значений создаем сервис с value
+          child_service = Service.new(
+            service: service_type ? service_type.to_s : "value",
+            url: nil,
+            properties: {},
+            value: item  # Используем поле value
+          )
+          children << child_service
+          logger.debug "Created service with value for item #{index}: #{item.inspect}"
+        end
+      end
 
-      # If there are child services, create a parent service with them
-      if children.any? && service_type
-        return Service.new(service: service_type.to_s, children: children)
-      elsif children.size == 1
-        # If only one child service, return it
-        return children.first
+      # Return service with all items as children
+      if service_type
+        result = Service.new(
+          service: service_type.to_s,
+          url: nil,
+          children: children
+        )
+        logger.debug "Returning array service with #{children.size} children"
+        return result
       end
 
-      # If no child services or no name for parent service,
-      # return nil
+      # Fallback to nil if no service_type
       return nil
     else
       # Handle values that don't match any pattern

data/lib/sitedog_parser/version.rb

@@ -1,3 +1,3 @@
 module SitedogParser
-  VERSION = "0.3.1"
+  VERSION = "0.4.0"
 end

data/lib/sitedog_parser.rb

@@ -68,9 +68,16 @@ module SitedogParser
             # Для обычных полей создаем сервис
             service = ServiceFactory.create(data, service_type, dictionary_path, options)
 
+            # Debug output
+            if logger
+              logger.debug "ServiceFactory.create for #{service_type}: #{service.inspect}"
+            end
+
             if service
               services[service_type] ||= []
               services[service_type] << service
+            elsif logger
+              logger.debug "Service for #{service_type} is nil, field will be skipped"
             end
           end
         end
@@ -103,11 +110,49 @@ module SitedogParser
           if service_data.is_a?(Array) && service_data.first.is_a?(Service)
             # Преобразуем массив сервисов в массив хешей
             result[domain_key][service_type_key] = service_data.map do |service|
-              {
+              service_hash = {
                 'service' => service.service,
-                'url' => service.url,
-                'children' => service.children.map { |child| {'service' => child.service, 'url' => child.url} }
+                'url' => service.url
               }
+
+              # Добавляем image_url если он есть
+              if service.image_url
+                service_hash['image_url'] = service.image_url
+              end
+
+              # Добавляем children только если они есть
+              if service.children && !service.children.empty?
+                service_hash['children'] = service.children.map do |child|
+                  child_hash = {
+                    'service' => child.service,
+                    'url' => child.url
+                  }
+
+                  # Добавляем image_url для детей если он есть
+                  if child.image_url
+                    child_hash['image_url'] = child.image_url
+                  end
+
+                  # Добавляем properties для children если они есть
+                  if child.properties && !child.properties.empty?
+                    child_hash['properties'] = child.properties
+                  end
+
+                  # Добавляем value для children если оно есть
+                  if child.value
+                    child_hash['value'] = child.value
+                  end
+
+                  child_hash
+                end
+              end
+
+              # Добавляем properties, если они есть
+              if service.properties && !service.properties.empty?
+                service_hash['properties'] = service.properties
+              end
+
+              service_hash
             end
           else
             # Сохраняем простые поля как есть

data/lib/url_checker.rb

@@ -28,7 +28,7 @@ module UrlChecker
     end
 
     # Check for standard URLs
-    pattern = /^((?:https?|ftp|sftp|ftps|ssh|git|ws|wss):\/\/)?[a-zA-Z0-9][-a-zA-Z0-9.]+\.[a-zA-Z]{2,}(:[0-9]+)?(\/[-a-zA-Z0-9%_.~#+]*)*(\?[-a-zA-Z0-9%_&=.~#+]*)?(#[-a-zA-Z0-9%_&=.~#+\/]*)?$/
+    pattern = /^((?:https?|ftp|sftp|ftps|ssh|git|ws|wss):\/\/)?((?:[a-zA-Z0-9][-a-zA-Z0-9.]+\.[a-zA-Z]{2,})|(?:\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}))(:[0-9]+)?(\/[-a-zA-Z0-9%_.~#+]*)*(\?[-a-zA-Z0-9%_&=.~#+]*)?(#[-a-zA-Z0-9%_&=.~#+\/]*)?$/
 
     !!string.match(pattern)
   end
@@ -61,6 +61,11 @@ module UrlChecker
     # Remove protocol and www prefix if present
     domain = url.gsub(%r{^(?:https?://)?(?:www\.)?}, "")
 
+    # Check if it's an IP address
+    if domain.match?(/^\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}/)
+      return "IP Address"
+    end
+
     # Extract domain from URL by removing everything after first / or : or ? or #
     domain = domain.split(/[:\/?#]/).first
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: sitedog_parser
 version: !ruby/object:Gem::Version
-  version: 0.3.1
+  version: 0.4.0
 platform: ruby
 authors:
 - Ivan Nemytchenko
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2025-05-07 00:00:00.000000000 Z
+date: 2025-05-08 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler